Compare commits
4
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
 Sven Woop
|
86dbc8a437 | ||
|
Sven Woop
|
4c93274a31 | ||
|
Sven Woop
|
bd7ea15f50 | ||
|
Sven Woop
|
1df9591a8c |
@@ -1,7 +0,0 @@
|
||||
root = true
|
||||
|
||||
[*]
|
||||
end_of_line = lf
|
||||
insert_final_newline = true
|
||||
indent_style = space
|
||||
indent_size = 2
|
||||
@@ -11,4 +11,3 @@
|
||||
*.jpg binary
|
||||
*.png binary
|
||||
*.pdf binary
|
||||
*.rc text working-tree-encoding=UTF-16LE-BOM eol=CRLF
|
||||
|
||||
@@ -1,223 +0,0 @@
|
||||
## Copyright 2022 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
name: continuous
|
||||
|
||||
on: [push, workflow_dispatch]
|
||||
|
||||
jobs:
|
||||
|
||||
################################################################################
|
||||
################################################################################
|
||||
# CPU tests
|
||||
################################################################################
|
||||
################################################################################
|
||||
|
||||
########################################
|
||||
# Linux
|
||||
########################################
|
||||
|
||||
continuous-linux-x64-Debug-ICC2019_4-ISPC1_12_0-AVX512-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:27
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:ICC2019.4 ispc:ispc1.12.0 isa:AVX512 frequency_level:simd256 tasking:TBB2021.2.0 intensity:2 maxinstancelevelcount:4
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-linux-avx512vl-x64-Debug-ICC2019_4-ISPC1_12_0-AVX512-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:27
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:ICC2019.4 ispc:ispc1.12.0 isa:AVX512 frequency_level:simd256 tasking:TBB2021.2.0 intensity:2 maxinstancelevelcount:4
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-linux-avx512vl-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX512-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX512 frequency_level:simd128 tasking:TBB2021.2.0 GLFW:OFF intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-linux-x64-Debug-CLANG5_0_0-ISPC1_17_0-AVX2-INT-ADDRSANITIZER:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
runs-on: '[ "Linux", "docker", "build" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:CLANG5.0.0 ispc:ispc1.17.0 isa:AVX tasking:INT intensity:1 addrsanitizer:ON
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-ubuntu20_04-x64-Debug-GCC-ISPC1_12_0-AVX2-TBB-NAMESPACE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/ubuntu:20.04
|
||||
runs-on: '[ "Linux", "docker", "build" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.12.0 isa:AVX2 frequency_level:simd256 tasking:TBB intensity:1 api_namespace:myembree
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-linux-avx512vl-x64-Debug-ICX2022_1_2-ISPC1_17_0-AVX512-INT:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/ubuntu:20.04
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug threads:4 compiler:ICX2022.1.2 ispc:ispc1.17.0 isa:AVX512 frequency_level:simd256 tasking:INT intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
#continuous-ubuntu17_10-x64-Debug-GCC-ISPC1_12_0-AVX2-TBB-NAMESPACE:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
# with:
|
||||
# image: embree/ubuntu:17.10
|
||||
# runs-on: '[ "Linux", "docker" ]'
|
||||
# cmd: |
|
||||
# python scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.12.0 isa:AVX2 tasking:TBB intensity:1 api_namespace:myembree
|
||||
# python scripts/test.py build
|
||||
# python scripts/test.py test
|
||||
|
||||
continuous-centos7_4-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
runs-on: '[ "Linux", "docker", "build" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 frequency_level:simd128 tasking:TBB intensity:1
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
continuous-fedora32-x64-Debug-GCC-ISPC-AVX512-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:32
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isa:AVX512 tasking:TBB COMPACT_POLYS:ON MIN_WIDTH:ON intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
|
||||
########################################
|
||||
# Windows
|
||||
########################################
|
||||
continuous-windows-x64-RelWithDebInfo-V140-ISPC1_9_2-SSE2-TBB2020_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
runs-on: '[ "Windows", "NAS", "build" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 threads:4 build:RelWithDebInfo compiler:V140 ispc:ispc1.9.2 isa:SSE2 tasking:TBB2020.2 intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
# continuous-windows-Win32-RelWithDebInfoDebug-V140-ISPC1_12_0-SSE2-TBB2020_2:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "Windows", "NAS", "build" ]'
|
||||
# cmd: |
|
||||
# python scripts/test.py configure platform:Win32 build:RelWithDebInfo compiler:V140 ispc:ispc1.12.0 isa:SSE2 tasking:TBB2020.2 intensity:2
|
||||
# python scripts/test.py build
|
||||
# python scripts/test.py test
|
||||
|
||||
continuous-windows-x64-RelWithDebInfo-V141-ISPC1_17_0-SSE2-TBB2020_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
runs-on: '[ "Windows", "NAS", "build" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V141 ispc:ispc1.17.0 isa:SSE2 tasking:TBB2020.2 intensity:2 maxinstancelevelcount:8
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
# continuous-windows-x64-RelWithDebInfo-ICX2022_0_0-ISPC1_16_1-AVX512-INT:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "Windows", "NAS", "build", "avx512" ]'
|
||||
# cmd: |
|
||||
# python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:ICX2022.0.0 ispc:ispc1.16.1 isa:AVX512 tasking:INT threads:16 intensity:2
|
||||
# python scripts/test.py build
|
||||
# python scripts/test.py test
|
||||
|
||||
|
||||
########################################
|
||||
# MacOS
|
||||
########################################
|
||||
|
||||
continuous-macosx-ARM-Debug-CLANG-ISPC1_17_0-NEON-TBB_HOMEBREW:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "arm" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:arm64 build:Debug compiler:CLANG ispc:ispc1.17.0 isa:NEON tasking:TBB_HOMEBREW intensity:2
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
continuous-macosx-ARM-Debug-CLANG-ISPC1_17_0-NEON-AVX2-TBB_HOMEBREW:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "arm" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:arm64 build:Debug compiler:CLANG ispc:ispc1.17.0 isa:NEON2X tasking:TBB_HOMEBREW intensity:2
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# continuous-macosx-x64-Debug-ICC2020_1-ISPC1_14_1-SSE2-TBB2021_1_1:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "macOS", "embree" ]'
|
||||
# cmd: |
|
||||
# scripts/test.py configure platform:x64 build:Debug compiler:ICC2020.1 ispc:ispc1.14.1 isa:SSE2 tasking:TBB2021.1.1 intensity:2
|
||||
# scripts/test.py build
|
||||
# scripts/test.py test
|
||||
|
||||
continuous-macosx-x64-Debug-CLANG-ISPC1_17_0-SSE2-TBB2020_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "embree" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG ispc:ispc1.17.0 isa:SSE2 tasking:TBB2020.2 intensity:2
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
|
||||
continuous-linux-x64-Debug-GCC-ISPC1_17_0-SSE2-TBB-KLOCWORK:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_static_analysis.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
runs-on: '[ "Linux", "docker", "build" ]'
|
||||
cmd: |
|
||||
export CI_PROJECT_DIR=$GITHUB_WORKSPACE
|
||||
export CI_JOB_ID=$GITHUB_RUN_ID
|
||||
python scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:SSE2 tasking:TBB klocwork:ON intensity:0
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
scripts/klocwork_gen_report.sh
|
||||
scripts/store_files.sh $GITHUB_REPOSITORY $GITHUB_RUN_ID klocwork "klocwork/report.log"
|
||||
|
||||
@@ -1,645 +0,0 @@
|
||||
## Copyright 2022 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
name: nightly
|
||||
|
||||
on:
|
||||
schedule:
|
||||
- cron: '30 06 * * *'
|
||||
workflow_dispatch:
|
||||
|
||||
jobs:
|
||||
|
||||
################################################################################
|
||||
################################################################################
|
||||
# CPU tests
|
||||
################################################################################
|
||||
################################################################################
|
||||
|
||||
########################################
|
||||
# Linux
|
||||
########################################
|
||||
|
||||
# CLANG compilation and testing of different ISAs
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_12_0-SSE2-TBB-NAMESPACE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.12.0 isa:SSE2 tasking:TBB2019.9 intensity:4 api_namespace:myembree maxinstancelevelcount:8
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG5_0_0-ISPC1_15_0-AVX2-INT-ADDRSANITIZER:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG5.0.0 ispc:ispc1.15.0 isa:AVX tasking:INT intensity:2 addrsanitizer:ON
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_16_1-AVX2-TBB2021_2_0-STATIC:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.16.1 isa:AVX2 frequency_level:simd128 tasking:TBB2021.2.0 intensity:4 STATIC_LIB:ON
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-avx512vl-x64-Debug-CLANG4_0_0-ISPC1_16_1-AVX512-TBB2020_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.16.1 isa:AVX512 frequency_level:simd256 tasking:TBB2020.2 intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# GCC compilation and testing of different ISAs
|
||||
|
||||
nightly-linux-x64-Debug-GCC-ISPC1_16_1-SSE2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.16.1 isa:SSE2 tasking:TBB2021.2.0 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-GCC-ISPC1_16_1-AVX-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.16.1 isa:AVX tasking:TBB2019.9 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# ICC compilation and testing of different ISAs
|
||||
|
||||
nightly-linux-x64-Debug-ICC2016_3-ISPC1_16_1-AVX-TBB2019_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2016.3 ispc:ispc1.16.1 isa:AVX tasking:TBB2019.2 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-avx512vl-x64-Debug-ICC2017_1-ISPC1_16_1-AVX512-TBB2019_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2017.1 ispc:ispc1.16.1 isa:AVX512 frequency_level:simd256 tasking:TBB2019.2 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-avx512vl-x64-Debug-ICC2018_0-ISPC1_16_1-AVX512-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2018.0 ispc:ispc1.16.1 isa:AVX512 tasking:TBB2019.9 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-ICC2019_4-ISPC1_16_1-AVX2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2019.4 ispc:ispc1.16.1 isa:AVX2 tasking:TBB2021.2.0 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-ICC2020_1-ISPC1_17_0-AVX2-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2020.1 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# Testing on different Linux distributions
|
||||
|
||||
nightly-ubuntu20_04-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/ubuntu:20.04
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-ubuntu20_04-x64-Release-GCC-ISPC1_14_1-AVX2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/ubuntu:20.04
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Release compiler:GCC ispc:ispc1.14.1 isa:AVX2 tasking:TBB2021.2.0 intensity:2
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-ubuntu16_04-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/ubuntu:16.04
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 tasking:TBB MIN_WIDTH:ON intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-centos7_4-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 tasking:TBB COMPACT_POLYS:ON intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
#nightly-centos7_0-x64-Debug-GCC-ISPC1_17_0-AVX2-TBB:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
# with:
|
||||
# image: embree/centos:7.0
|
||||
# cmd: |
|
||||
# scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc1.17.0 isa:AVX2 tasking:TBB intensity:3"]
|
||||
# scripts/test.py build
|
||||
# scripts/test.py test
|
||||
|
||||
nightly-fedora32-x64-Debug-GCC-ISPC-AVX512-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:32
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isas:AVX512 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-fedora28-x64-Debug-GCC-ISPC-AVX512-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:28
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isas:SSE2-SSE42-AVX-AVX2-AVX512 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-fedora27-x64-Debug-GCC-ISPC-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:27
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isa:AVX2 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-fedora26-x64-Debug-GCC-ISPC-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:26
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isa:AVX2 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-fedora25-x64-Debug-GCC-ISPC-AVX2-TBB:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:GCC ispc:ispc isa:AVX2 tasking:TBB intensity:3
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# Compilation test of disabled features
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-AVX2-TBB2019_9-NO-ISPC:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 isa:AVX2 tasking:TBB2019.9 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-NO-TUTORIALS:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 TUTORIALS:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-BACKFACECULLING:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 BACKFACE_CULLING:ON intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-IGNORE-INVALID-RAYS:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 IGNORE_INVALID_RAYS:ON intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-NO-FILTER-FUNCTION:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 FILTER_FUNCTION:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-RAYMASKS:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 RAY_MASK:ON intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-NO-PACKETS:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 RAY_PACKETS:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-STATCOUNTER:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 STAT_COUNTER:ON intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# Compilation test of individual ISAs
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-ISAS-SSE2-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isas:SSE2 tasking:TBB2019.9 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-ISAS-SSE42-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isas:SSE42 tasking:TBB2021.2.0 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-ISAS-AVX-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isas:AVX tasking:TBB2019.9 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-ISAS-AVX2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isas:AVX2 tasking:TBB2021.2.0 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-ISAS-AVX512-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isas:AVX512 tasking:TBB2019.9 intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# Compilation test of individual primitive types enabled
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-TRI:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 TRI:ON QUAD:OFF GRID:OFF CURVE:OFF SUBDIV:OFF USERGEOM:OFF INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-QUAD:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 TRI:OFF QUAD:ON GRID:OFF CURVE:OFF SUBDIV:OFF USERGEOM:OFF INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-GRID:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 TRI:OFF QUAD:OFF GRID:ON CURVE:OFF SUBDIV:OFF USERGEOM:OFF INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-CURVE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 TRI:ON QUAD:OFF GRID:OFF CURVE:ON SUBDIV:OFF USERGEOM:OFF INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-SUBDIV:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 TRI:ON QUAD:OFF GRID:OFF CURVE:OFF SUBDIV:ON USERGEOM:OFF INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2019_9-USERGEOM:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2019.9 TRI:ON QUAD:OFF GRID:OFF CURVE:OFF SUBDIV:OFF USERGEOM:ON INSTANCE:OFF intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-linux-x64-Debug-CLANG4_0_0-ISPC1_17_0-AVX2-TBB2021_2_0-INSTANCE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/fedora:25
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG4.0.0 ispc:ispc1.17.0 isa:AVX2 tasking:TBB2021.2.0 TRI:ON QUAD:OFF GRID:OFF CURVE:OFF SUBDIV:OFF USERGEOM:OFF INSTANCE:ON intensity:0
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
|
||||
########################################
|
||||
# Windows
|
||||
########################################
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-V140-ISPC1_9_2-SSE2-INT:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V140 ispc:ispc1.9.2 isa:SSE2 tasking:INT intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-V140-ISPC1_9_2-AVX-TBB2019_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V140 ispc:ispc1.9.2 isa:AVX tasking:TBB2019.2 intensity:4
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-Win32-RelWithDebInfo-V140-ISPC1_12_0-SSE2-TBB2020_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:Win32 build:RelWithDebInfo compiler:V140 ispc:ispc1.12.0 isa:SSE2 tasking:TBB2020.2 intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-V140-ISPC1_12_0-AVX2-PPL:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V140 ispc:ispc1.12.0 isa:AVX2 frequency_level:simd256 tasking:PPL intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-V141-ISPC1_13_0-AVX2-TBB2019_9-STATIC:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V141 ispc:ispc1.13.0 isa:AVX2 frequency_level:simd128 tasking:TBB2019.9 intensity:2 STATIC_LIB:ON
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-V142-ISPC1_13_0-AVX2-TBB2020_2-STATIC:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:V142 ispc:ispc1.13.0 isa:AVX2 tasking:TBB2020.2 intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-LLVM-CLANG-ISPC1_13_0-AVX2-TBB2019_9-NAMESPACE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:LLVM_CLANG ispc:ispc1.13.0 isa:AVX2 tasking:TBB2019.9 intensity:4 api_namespace:myembree
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-ICC17-ISPC1_13_0-AVX2-TBB2019_2:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:ICC17-VC14 ispc:ispc1.13.0 isa:AVX2 frequency_level:simd128 tasking:TBB2019.2 intensity:4
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
# nightly-windows-x64-RelWithDebInfo-ICC18-ISPC1_13_0-AVX512-TBB2020_2:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "Windows", "NAS", "build", "avx512" ]'
|
||||
# cmd: |
|
||||
# python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:ICC18-VC14 ispc:ispc1.13.0 isa:AVX512 tasking:TBB2020.2 intensity:2
|
||||
# python scripts/test.py build
|
||||
# python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-ICC19-ISPC1_13_0-AVX512-TBB2019_9:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
runs-on: '[ "Windows", "NAS", "build", "avx512" ]'
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:RelWithDebInfo compiler:ICC19-VC14 ispc:ispc1.13.0 isa:AVX512 tasking:TBB2019.9 intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
nightly-windows-x64-RelWithDebInfo-ICC17-ISPC1_14_1-AVX2-TBB2021_2_0:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Release compiler:V142 ispc:ispc1.14.1 isa:AVX2 tasking:TBB2021.2.0 intensity:2
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
|
||||
########################################
|
||||
# MacOS
|
||||
########################################
|
||||
|
||||
# CLANG compilation and testing of different ISAs
|
||||
|
||||
nightly-macosx-x64-Debug-CLANG-ISPC1_12_0-SSE2-TBB2020_2-NAMESPACE:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "embree" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG ispc:ispc1.12.0 isa:SSE2 tasking:TBB2020.2 intensity:4 api_namespace:myembree maxinstancelevelcount:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-macosx-x64-Debug-CLANG-ISPC1_13_0-AVX-INT:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "embree" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG ispc:ispc1.13.0 isa:AVX tasking:INT intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
nightly-macosx-x64-Debug-CLANG-ISPC1_13_0-SSE2-TBB2019_9-STATIC:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "embree" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:CLANG ispc:ispc1.13.0 isa:SSE2 tasking:TBB2019.9 intensity:4 STATIC_LIB:ON
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
# ICC compilation and testing of different ISAs
|
||||
|
||||
# nightly-macosx-x64-Debug-ICC15-ISPC1_13_0-AVX-INT:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "macOS", "embree" ]'
|
||||
# cmd: |
|
||||
# scripts/test.py configure platform:x64 build:Debug compiler:ICC15 ispc:ispc1.13.0 isa:AVX tasking:INT intensity:4
|
||||
# scripts/test.py build
|
||||
# scripts/test.py test
|
||||
|
||||
# nightly-macosx-x64-Debug-ICC2019_4-ISPC1_13_0-AVX2-TBB2020_2:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "macOS", "avx2", "embree" ]'
|
||||
# cmd: |
|
||||
# scripts/test.py configure platform:x64 build:Debug compiler:ICC2019.4 ispc:ispc1.13.0 isa:AVX2 frequency_level:simd256 tasking:TBB2020.2 intensity:4
|
||||
# scripts/test.py build
|
||||
# scripts/test.py test
|
||||
|
||||
# nightly-macosx-x64-Debug-ICC2020_1-ISPC1_13_0-AVX2-TBB2019_9:
|
||||
# secrets: inherit
|
||||
# uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
# with:
|
||||
# runs-on: '[ "macOS", "avx2", "embree" ]'
|
||||
# cmd: |
|
||||
# scripts/test.py configure platform:x64 build:Debug compiler:ICC2020.1 ispc:ispc1.13.0 isa:AVX2 frequency_level:simd128 tasking:TBB2019.9 intensity:4
|
||||
# scripts/test.py build
|
||||
# scripts/test.py test
|
||||
|
||||
nightly-macosx-x64-Debug-ICC2021_1_1-ISPC1_13_0-AVX2-TBB2021_1_1:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "avx2", "embree" ]'
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Debug compiler:ICC2021.1.1 ispc:ispc1.13.0 isa:AVX2 frequency_level:simd128 tasking:TBB2021.1.1 intensity:4
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
@@ -1,85 +0,0 @@
|
||||
## Copyright 2022 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
name: release
|
||||
|
||||
on:
|
||||
schedule:
|
||||
- cron: '30 04 * * *'
|
||||
workflow_dispatch:
|
||||
|
||||
jobs:
|
||||
|
||||
#####################################################################
|
||||
# Release Builds
|
||||
####################################################################
|
||||
|
||||
release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_windows.yml@main
|
||||
with:
|
||||
runs-on: '[ "Windows", "NAS", "build", "avx512" ]'
|
||||
artifact-out: release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-path: ./build/*.zip
|
||||
cmd: |
|
||||
python scripts/test.py configure platform:x64 build:Release compiler:ICC19-VC141 ispc:ispc1.17.0 isas:SSE2-SSE42-AVX-AVX2-AVX512 tasking:TBB2021.2.0 intensity:4 package:ZIP
|
||||
python scripts/test.py build
|
||||
python scripts/test.py test
|
||||
|
||||
release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_macos.yml@main
|
||||
with:
|
||||
runs-on: '[ "macOS", "embree", "sign", "avx2" ]'
|
||||
artifact-out: release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-path: ./build/*.zip
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Release compiler:ICC2021.1.1 ispc:ispc1.17.0 isas:SSE2-SSE42-AVX-AVX2 tasking:TBB2021.2.0 intensity:4 package:ZIP
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP:
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/embree_docker.yml@main
|
||||
with:
|
||||
image: embree/centos:7.4
|
||||
runs-on: '[ "Linux", "docker", "avx512" ]'
|
||||
artifact-out: release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-path: ./build/*.tar.gz
|
||||
cmd: |
|
||||
scripts/test.py configure platform:x64 build:Release compiler:ICC2020.1 ispc:ispc1.17.0 isas:SSE2-SSE42-AVX-AVX2-AVX512 tasking:TBB2021.2.0 intensity:4 package:ZIP
|
||||
scripts/test.py build
|
||||
scripts/test.py test
|
||||
|
||||
|
||||
#####################################################################
|
||||
# Release binaries scans
|
||||
#####################################################################
|
||||
binary-analysis:
|
||||
needs:
|
||||
- release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
- release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
- release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/binary_analysis.yml@main
|
||||
with:
|
||||
project: embree
|
||||
artifact-in-linux: release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-in-macos: release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-in-windows: release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
path: "build/*.zip build/*.gz"
|
||||
|
||||
antivirus-scan:
|
||||
needs:
|
||||
- release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
- release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
- release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
secrets: inherit
|
||||
uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/antivirus_scan.yml@main
|
||||
with:
|
||||
project: embree
|
||||
artifact-in-linux: release-windows-x64-Release-ICC19-VC141-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-in-macos: release-macosx-x64-Release-ICC2021_1_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
artifact-in-windows: release-linux-x64-Release-ICC2020_1-ISPC1_17_0-TBB2021_2_0-PACKAGE-ZIP
|
||||
path: "build/*.zip build/*.gz"
|
||||
|
||||
+4
-1
@@ -7,5 +7,8 @@ x64
|
||||
.DS_Store
|
||||
kernels/config.h
|
||||
kernels/hash.h
|
||||
include/embree3/rtcore_config.h
|
||||
build
|
||||
models
|
||||
tbb
|
||||
doc/intelstyle
|
||||
|
||||
|
||||
+416
-632
File diff suppressed because it is too large
Load Diff
+14
-215
@@ -1,207 +1,6 @@
|
||||
Version History
|
||||
---------------
|
||||
|
||||
### Embree 3.13.5
|
||||
- Fixed bug in bounding flat Catmull Rom curves of subdivision level 4.
|
||||
- Improved self intersection avoidance for
|
||||
RTC_GEOMETRY_TYPE_DISC_POINT geometry type. Intersections are
|
||||
skipped if the ray origin lies inside the sphere defined by the
|
||||
point primitive. Self intersection avoidance can get disabled at compile time
|
||||
using the EMBREE_DISC_POINT_SELF_INTERSECTION_AVOIDANCE cmake option.
|
||||
- Fixed spatial splitting for non-planar quads.
|
||||
|
||||
### Embree 3.13.4
|
||||
- Using 8-wide BVH and double pumped NEON instructions on Apple M1 gives 8% performance boost.
|
||||
- Fixed binning related crash in SAH BVH builder.
|
||||
- Added EMBREE_TBB_COMPONENT cmake option to define the component/library name of Intel® TBB (default: tbb).
|
||||
- Embree supports now Intel® oneAPI DPC++/C++ Compiler 2022.0.0
|
||||
|
||||
### Embree 3.13.3
|
||||
- Invalid multi segment motion blurred normal oriented curves are properly excluded from BVH build.
|
||||
- Fixing issue with normal oriented curve construction when center curve curvature is very large.
|
||||
Due to this change normal oriented curve shape changes slightly.
|
||||
- Fixed crash caused by disabling a geometry and then detaching it from the scene.
|
||||
- Bugfix in emulated ray packet intersection when EMBREE_RAY_PACKETS is turned off.
|
||||
- Bugfix for linear quaternion interpolation fallback.
|
||||
- Fixed issues with spaces in path to Embree build folder.
|
||||
- Some fixes to compile Embree in SSE mode using WebAssembly.
|
||||
- Bugfix for occlusion rays with grids and ray packets.
|
||||
- We do no longer provide installers for Windows and macOS, please use the ZIP files instead.
|
||||
- Upgrading to Intel® ISPC 1.17.0 for release build.
|
||||
- Upgrading to Intel® oneTBB 2021.5.0 for release build.
|
||||
|
||||
### Embree 3.13.2
|
||||
- Avoiding spatial split positions that are slightly out of geometry bounds.
|
||||
- Introduced rtcGetGeometryThreadSafe function, which is a thread safe version of rtcGetGeometry.
|
||||
- Using more accurate rcp implementation.
|
||||
- Bugfix to rare corner case of high quality BVH builder.
|
||||
|
||||
### Embree 3.13.1
|
||||
- Added support for Intel® ISPC ARM target.
|
||||
- Releases upgrade to Intel® TBB 2021.3.0 and Intel® ISPC 1.16.1
|
||||
|
||||
### Embree 3.13.0
|
||||
- Added support for Apple M1 CPUs.
|
||||
- RTC_SUBDIVISION_MODE_NO_BOUNDARY now works properly for non-manifold edges.
|
||||
- CMake target 'uninstall' is not defined if it already exists.
|
||||
- Embree no longer reads the .embree3 config files, thus all configuration has
|
||||
to get passed through the config string to rtcNewDevice.
|
||||
- Releases upgrade to Intel® TBB 2021.2.0 and Intel® ISPC 1.15.0
|
||||
- Intel® TBB dll is automatically copied into build folder after build on windows.
|
||||
|
||||
### Embree 3.12.2
|
||||
- Fixed wrong uv and Ng for grid intersector in robust mode for AVX.
|
||||
- Removed optimizations for Knights Landing.
|
||||
- Upgrading release builds to use Intel® oneTBB 2021.1.1
|
||||
|
||||
### Embree 3.12.1
|
||||
|
||||
- Changed default frequency level to SIMD128 for Skylake, Cannon Lake, Comet Lake and Tiger Lake CPUs.
|
||||
This change typically improves performance for renderers that just use SSE by maintaining higher
|
||||
CPU frequencies. In case your renderer is AVX optimized you can get higher ray tracing performance
|
||||
by configuring the frequency level to simd256 through passing frequency_level=simd256 to rtcNewDevice.
|
||||
|
||||
### Embree 3.12.0
|
||||
|
||||
- Added linear cone curve geometry support. In this mode a real geometric surface for curves
|
||||
with linear basis is rendered using capped cones. They are discontinuous at edge boundaries.
|
||||
- Enabled fast two level builder for instances when low quality build is requested.
|
||||
- Bugfix for BVH build when geometries got disabled.
|
||||
- Added EMBREE_BACKFACE_CULLING_CURVES cmake option. This allows for a cheaper round
|
||||
linear curve intersection when correct internal tracking and back hits are not required.
|
||||
The new cmake option defaults to OFF.
|
||||
- User geometries with invalid bounds with lower>upper in some dimension will be ignored.
|
||||
- Increased robustness for grid interpolation code and fixed returned out of range u/v
|
||||
coordinates for grid primitive.
|
||||
- Fixed handling of motion blur time range for sphere, discs, and oriented disc geometries.
|
||||
- Fixed missing model data in releases.
|
||||
- Ensure compatibility to newer versions of Intel® oneTBB.
|
||||
- Motion blur BVH nodes no longer store NaN values.
|
||||
|
||||
### Embree 3.11.0
|
||||
|
||||
- Round linear curves now automatically check for the existence of left and right
|
||||
connected segments if the flags buffer is empty. Left segments exist if the
|
||||
segment(id-1) + 1 == segment(id) and similarly for right segments.
|
||||
- Implemented the min-width feature for curves and points, which allows to increase the
|
||||
radius in a distance dependent way, such that the curve or points thickness is n pixels wide.
|
||||
- Round linear curves are closed now also at their start.
|
||||
- Embree no longer supports Visual Studio 2013 starting with this release.
|
||||
- Bugfix in subdivision tessellation level assignment for non-quad base primitives
|
||||
- Small meshes are directly added to top level build phase of two-level builder to reduce memory consumption.
|
||||
- Enabled fast two level builder for user geometries when low quality build is requested.
|
||||
|
||||
### Embree 3.10.0
|
||||
|
||||
- Added EMBREE_COMPACT_POLYS CMake option which enables double indexed triangle and quad
|
||||
leaves to reduce memory consumption in compact mode by an additional 40% at about
|
||||
15% performance impact. This new mode is disabled by default.
|
||||
- Compile fix for Intel® oneTBB 2021.1-beta05
|
||||
- Releases upgrade to Intel® TBB 2020.2
|
||||
- Compile fix for Intel® ISPC v1.13.0
|
||||
- Adding RPATH to libembree.so in releases
|
||||
- Increased required CMake version to 3.1.0
|
||||
- Made instID member for array of pointers ray stream layout optional again.
|
||||
|
||||
### Embree 3.9.0
|
||||
|
||||
- Added round linear curve geometry support. In this mode a real geometric surface for curves
|
||||
with linear basis is rendered using capped cones with spherical filling between
|
||||
the curve segments.
|
||||
- Added rtcGetSceneDevice API function, that returns the device a scene got created in.
|
||||
- Improved performance of round curve rendering by up to 1.8x.
|
||||
- Bugfix to sphere intersection filter invocation for back hit.
|
||||
- Fixed wrong assertion that triggered for invalid curves which anyway get filtered out.
|
||||
- RelWithDebInfo mode no longer enables assertions.
|
||||
- Fixed an issue in FindTBB.cmake that caused compile error with Debug build under Linux.
|
||||
- Embree releases no longer provide RPMs for Linux. Please use the RPMs coming with the package
|
||||
manager of your Linux distribution.
|
||||
|
||||
### Embree 3.8.0
|
||||
|
||||
- Added collision detection support for user geometries (see rtcCollide API function)
|
||||
- Passing geomID to user geometry callbacks.
|
||||
- Bugfix in AVX512VL codepath for rtcIntersect1
|
||||
- For sphere geometries the intersection filter gets now invoked for
|
||||
front and back hit.
|
||||
- Fixed some bugs for quaternion motion blur.
|
||||
- RTCIntersectContext always non-const in Embree API
|
||||
- Made RTCHit aligned to 16 bytes in Embree API
|
||||
|
||||
### New Features in Embree 3.7.0
|
||||
- Added quaternion motion blur for correct interpolation of rotational transformations.
|
||||
- Fixed wrong bounding calculations when a motion blurred instance did
|
||||
instantiate a motion blurred scene.
|
||||
- In robust mode the depth test consistently uses tnear <= t <= tfar now in order
|
||||
to robustly continue traversal at a previous hit point
|
||||
in a way that guarantees reaching all hits, even hits at the same place.
|
||||
- Fixed depth test in robust mode to be precise at tnear and tfar.
|
||||
- Added next_hit tutorial to demonstrate robustly collecting all hits
|
||||
along a ray using multiple ray queries.
|
||||
- Implemented robust mode for curves. This has a small performance impact but
|
||||
fixes bounding problems with flat curves.
|
||||
- Improved quality of motion blur BVH by using linear bounds during binning.
|
||||
- Implemented issue with motion blur builder where number of time segments
|
||||
for SAH heuristic were counted wrong due to some numerical issues.
|
||||
- Fixed an accuracy issue with rendering very short fat curves.
|
||||
- rtcCommitScene can now get called during rendering from multiple threads
|
||||
to lazily build geometry. When Intel® TBB is used this causes a much lower overhead
|
||||
than using rtcJoinCommitScene.
|
||||
- Geometries can now get attached to multiple scenes at the same time, which
|
||||
simplifies mapping general scene graphs to API.
|
||||
- Updated to Intel® TBB 2019.9 for release builds.
|
||||
- Fixed a bug in the BVH builder for Grid geometries.
|
||||
- Added macOS Catalina support to Embree releases.
|
||||
|
||||
### New Features in Embree 3.6.1
|
||||
- Restored binary compatibility between Embree 3.6 and 3.5 when single-level instancing is used.
|
||||
- Fixed bug in subgrid intersector
|
||||
- Removed point query alignment in Intel® ISPC header
|
||||
|
||||
### New Features in Embree 3.6
|
||||
- Added Catmull-Rom curve types.
|
||||
- Added support for multi-level instancing.
|
||||
- Added support for point queries.
|
||||
- Fixed a bug preventing normal oriented curves being used unless timesteps were
|
||||
specified.
|
||||
- Fixed bug in external BVH builder when configured for dynamic build.
|
||||
- Added support for new config flag "user_threads=N" to device initialization
|
||||
which sets the number of threads used by Intel® TBB but created by the user.
|
||||
- Fixed automatic vertex buffer padding when using rtcSetNewGeometry API function.
|
||||
|
||||
### New Features in Embree 3.5.2
|
||||
- Added EMBREE_API_NAMESPACE cmake option that allows to put all Embree API functions
|
||||
inside a user defined namespace.
|
||||
- Added EMBREE_LIBRARY_NAME cmake option that allows to rename the Embree library.
|
||||
- When Embree is compiled as static library, EMBREE_STATIC_LIB has no longer to get
|
||||
defined before including the Embree API headers.
|
||||
- Added CPU frequency_level device configuration to allow an application to specify the
|
||||
frequency level it wants to run on. This forces Embree to not use optimizations that
|
||||
may reduce the CPU frequency below that level. By default Embree is configured to the
|
||||
the AVX-heavy frequency level, thus if the application uses solely non-AVX code, configuring
|
||||
the Embree device with "frequency_level=simd128" may give better performance.
|
||||
- Fixed a bug in the spatial split builder which caused it to fail
|
||||
for scenes with more than 2^24 geometries.
|
||||
|
||||
### New Features in Embree 3.5.1
|
||||
- Fixed ray/sphere intersector to work also for non-normalized rays.
|
||||
- Fixed self intersection avoidance for ray oriented discs when
|
||||
non-normalized rays were used.
|
||||
- Increased maximal face valence for subdiv patch to 64 and reduced stack size
|
||||
requirement for subdiv patch evaluation.
|
||||
|
||||
### New Features in Embree 3.5.0
|
||||
- Changed normal oriented curve definition to fix waving artefacts.
|
||||
- Fixed bounding issue for normal oriented motion blurred curves.
|
||||
- Fixed performance issue with motion blurred point geometry.
|
||||
- Fixed generation of documentation with new pandoc versions.
|
||||
|
||||
### New Features in Embree 3.4.0
|
||||
- Added point primitives (spheres, ray-oriented discs, normal-oriented discs).
|
||||
- Fixed crash triggered by scenes with only invalid primitives.
|
||||
- Improved robustness of quad/grid-based intersectors.
|
||||
- Upgraded to Intel® TBB 2019.2 for release builds.
|
||||
|
||||
### New Features in Embree 3.3.0
|
||||
- Added support for motion blur time range per geometry. This way geometries
|
||||
can appear and disappear during the camera shutter and time steps do not have
|
||||
@@ -358,7 +157,7 @@ Version History
|
||||
by 5-15%.
|
||||
- Fixed tbb_debug.lib linking error under Windows.
|
||||
- Fast coherent ray stream and packet code paths now also work in robust mode.
|
||||
- Using less aggressive prefetching for large BVH nodes which
|
||||
- Using less agressive prefetching for large BVH nodes which
|
||||
results in 1-2% higher ray tracing performance.
|
||||
- Precompiled binaries have stack-protector enabled, except for
|
||||
traversal kernels. BVH builders can be slightly slower due to this
|
||||
@@ -369,7 +168,7 @@ Version History
|
||||
fixed, and one can enable only AVX2 and still get best
|
||||
performance by using an 8-wide BVH.
|
||||
- Fixed rtcOccluded1 and rtcOccluded1Ex API functions which were
|
||||
broken in Intel® ISPC.
|
||||
broken in ISPC.
|
||||
- Providing MSI installer for Windows.
|
||||
|
||||
### New Features in Embree 2.16.5
|
||||
@@ -395,7 +194,7 @@ Version History
|
||||
cracks when using displacement mapping but reduces performance
|
||||
at irregular vertices.
|
||||
- Fixed a bug where subdivision geometry was not properly updated
|
||||
when modifying only the tessellation rate and vertex array.
|
||||
when modifying only the tesselation rate and vertex array.
|
||||
|
||||
### New Features in Embree 2.16.2
|
||||
- Fixed bug that caused NULL intersection context in intersection
|
||||
@@ -558,14 +357,14 @@ Version History
|
||||
If you use Embree v2.11.0 please upgrade to Embree v2.12.0.
|
||||
- Reduced memory consumption for dynamic scenes containing small
|
||||
meshes.
|
||||
- Added support to start and affinitize Intel® TBB worker threads by passing
|
||||
- Added support to start and affinitize TBB worker threads by passing
|
||||
"`start_threads=1,set_affinity=1`" to `rtcNewDevice`. These settings
|
||||
are recommended on systems with a high thread count.
|
||||
- `rtcInterpolate2` can now be called within a displacement shader.
|
||||
- Added initial support for Microsoft's Parallel Pattern Library (PPL)
|
||||
as tasking system alternative (for optimal performance Intel® TBB is
|
||||
as tasking system alternative (for optimal performance TBB is
|
||||
highly recommended).
|
||||
- Updated to Intel® TBB 2017 which is released under the Apache v2.0 license.
|
||||
- Updated to TBB 2017 which is released under the Apache v2.0 license.
|
||||
- Dropped support for Visual Studio 2012 Win32 compiler. Visual Studio
|
||||
2012 x64 is still supported.
|
||||
|
||||
@@ -629,14 +428,14 @@ Version History
|
||||
- Added support for quad geometry (replaces triangle-pairs feature).
|
||||
- Added support for linear motion blur of user geometries.
|
||||
- Improved performance through AVX-512 optimizations.
|
||||
- Improved performance of lazy scene build (when using Intel® TBB 4.4 update
|
||||
- Improved performance of lazy scene build (when using TBB 4.4 update
|
||||
2).
|
||||
- Improved performance through huge page support under linux.
|
||||
|
||||
### New Features in Embree 2.7.1
|
||||
|
||||
- Internal tasking system supports cancellation of build operations.
|
||||
- Intel® ISPC mode for robust and compact scenes got significantly faster
|
||||
- ISPC mode for robust and compact scenes got significantly faster
|
||||
(implemented hybrid traversal for bvh4.triangle4v and
|
||||
bvh4.triangle4i).
|
||||
- Hair rendering got faster as we fixed some issues with the SAH
|
||||
@@ -661,7 +460,7 @@ Version History
|
||||
- Added device concept to Embree to allow different components of an
|
||||
application to use Embree without interfering with each other.
|
||||
- Fixed memory leak in twolevel builder used for dynamic scenes.
|
||||
- Fixed bug in tessellation cache that caused crashes for subdivision
|
||||
- Fixed bug in tesselation cache that caused crashes for subdivision
|
||||
surfaces.
|
||||
- Fixed bug in internal task scheduler that caused deadlocks when
|
||||
using `rtcCommitThread`.
|
||||
@@ -735,10 +534,10 @@ Version History
|
||||
progress and cancel long build operations
|
||||
- BVH builders can be used to build user defined hierarchies inside
|
||||
the application (see tutorial [BVH Builder])
|
||||
- Switched to Intel® TBB as default tasking system on Xeon to get even faster
|
||||
- Switched to TBB as default tasking system on Xeon to get even faster
|
||||
hierarchy build times and better integration for applications that
|
||||
also use Intel® TBB
|
||||
- `rtcCommit` can get called from multiple Intel® TBB threads to join the
|
||||
also use TBB
|
||||
- `rtcCommit` can get called from multiple TBB threads to join the
|
||||
hierarchy build operations
|
||||
|
||||
### New Features in Embree 2.4
|
||||
@@ -826,7 +625,7 @@ Version History
|
||||
- Support for the Intel® Xeon Phi™ coprocessor platform
|
||||
- Support for high-performance "packet" kernels on SSE, AVX, and Xeon
|
||||
Phi
|
||||
- Integration with the Intel® Implicit SPMD Program Compiler (Intel® ISPC)
|
||||
- Integration with the Intel® SPMD Program Compiler (ISPC)
|
||||
- Instantiation and fast BVH reconstruction
|
||||
- Example photo-realistic rendering engine for both C++ and Intel® ISPC
|
||||
- Example photo-realistic rendering engine for both C++ and ISPC
|
||||
|
||||
|
||||
+128
-306
@@ -1,10 +1,22 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
CMAKE_MINIMUM_REQUIRED(VERSION 3.1.0)
|
||||
SET(EMBREE_VERSION_MAJOR 3)
|
||||
SET(EMBREE_VERSION_MINOR 13)
|
||||
SET(EMBREE_VERSION_PATCH 5)
|
||||
SET(EMBREE_VERSION_MINOR 3)
|
||||
SET(EMBREE_VERSION_PATCH 0)
|
||||
SET(EMBREE_VERSION_NOTE "")
|
||||
|
||||
SET(EMBREE_VERSION ${EMBREE_VERSION_MAJOR}.${EMBREE_VERSION_MINOR}.${EMBREE_VERSION_PATCH})
|
||||
@@ -13,32 +25,29 @@ SET(CPACK_RPM_PACKAGE_RELEASE 1)
|
||||
|
||||
PROJECT(embree${EMBREE_VERSION_MAJOR})
|
||||
|
||||
# We use our own strip tool on macOS to sign during install. This is required as CMake modifies RPATH of the binary during install.
|
||||
IF (APPLE AND EMBREE_SIGN_FILE)
|
||||
SET(EMBREE_STRIP ${CMAKE_STRIP})
|
||||
SET(CMAKE_STRIP "${PROJECT_BINARY_DIR}/post_install_target.sh")
|
||||
CONFIGURE_FILE(scripts/post_install_target.sh.in "${PROJECT_BINARY_DIR}/post_install_target.sh" @ONLY)
|
||||
ENDIF()
|
||||
CMAKE_MINIMUM_REQUIRED(VERSION 2.8.11)
|
||||
|
||||
MACRO (SIGN_TARGET target)
|
||||
IF (EMBREE_SIGN_FILE)
|
||||
IF (WIN32)
|
||||
# on Windows we sign and do not strip as debug symbols not included in binary
|
||||
# on Windows we only strip as debug symbols not included in binary
|
||||
ADD_CUSTOM_COMMAND(TARGET ${target} POST_BUILD
|
||||
COMMAND ${EMBREE_SIGN_FILE} $<TARGET_FILE:${target}>)
|
||||
COMMAND ${EMBREE_SIGN_FILE} -vv -c embree $<TARGET_FILE:${target}>)
|
||||
ELSEIF (APPLE)
|
||||
# on MacOSX we strip and sign here for testing purposes but also during install, as CMake modifies binary during install
|
||||
# on MacOSX we strip and sign
|
||||
ADD_CUSTOM_COMMAND(TARGET ${target} POST_BUILD
|
||||
COMMAND ${EMBREE_STRIP} -x $<TARGET_FILE:${target}>
|
||||
COMMAND ${EMBREE_SIGN_FILE} -o runtime -e "${CMAKE_SOURCE_DIR}/common/cmake/embree.entitlements" $<TARGET_FILE:${target}>)
|
||||
COMMAND ${CMAKE_STRIP} -x $<TARGET_FILE:${target}>
|
||||
COMMAND ${EMBREE_SIGN_FILE} -vv -c embree $<TARGET_FILE:${target}>)
|
||||
ELSE()
|
||||
# on Linux signing of binaries is not supported and stripping is done during install
|
||||
# on Linux just strip as signing not supported
|
||||
ADD_CUSTOM_COMMAND(TARGET ${target} POST_BUILD
|
||||
COMMAND ${CMAKE_STRIP} -x $<TARGET_FILE:${target}>)
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
# find git version
|
||||
IF(EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
|
||||
IF(EXISTS ${CMAKE_CURRENT_SOURCE_DIR}/.git)
|
||||
FIND_PACKAGE(Git)
|
||||
IF(GIT_FOUND)
|
||||
EXECUTE_PROCESS(
|
||||
@@ -52,6 +61,15 @@ IF(EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/rtcore_version.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/include/embree3/rtcore_version.h"
|
||||
)
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/hash.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/hash.h"
|
||||
)
|
||||
|
||||
IF(COMMAND cmake_policy)
|
||||
if (POLICY CMP0003)
|
||||
cmake_policy(SET CMP0003 NEW)
|
||||
@@ -59,15 +77,6 @@ IF(COMMAND cmake_policy)
|
||||
if (POLICY CMP0042)
|
||||
cmake_policy(SET CMP0042 NEW)
|
||||
endif()
|
||||
if(POLICY CMP0072)
|
||||
cmake_policy(SET CMP0072 NEW)
|
||||
endif()
|
||||
if(POLICY CMP0022)
|
||||
cmake_policy(SET CMP0022 NEW)
|
||||
endif()
|
||||
if(POLICY CMP0074)
|
||||
cmake_policy(SET CMP0074 NEW)
|
||||
endif()
|
||||
ENDIF(COMMAND cmake_policy)
|
||||
|
||||
MARK_AS_ADVANCED(CMAKE_BACKWARDS_COMPATIBILITY)
|
||||
@@ -79,18 +88,9 @@ MARK_AS_ADVANCED(CMAKE_OSX_DEPLOYMENT_TARGET)
|
||||
MARK_AS_ADVANCED(CMAKE_OSX_SYSROOT)
|
||||
MARK_AS_ADVANCED(CLEAR CMAKE_CXX_COMPILER)
|
||||
|
||||
SET(CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/common/cmake" ${CMAKE_MODULE_PATH})
|
||||
SET(CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/common/cmake ${CMAKE_MODULE_PATH})
|
||||
|
||||
IF (BUILD_TESTING)
|
||||
INCLUDE(test)
|
||||
ENDIF()
|
||||
|
||||
set(BUILD_DOC OFF CACHE INTERNAL "build documentation (internal only)")
|
||||
IF (BUILD_DOC)
|
||||
ADD_SUBDIRECTORY(doc)
|
||||
ENDIF()
|
||||
|
||||
OPTION(EMBREE_TUTORIALS "Enable to build Embree tutorials" ON)
|
||||
INCLUDE(test)
|
||||
|
||||
##############################################################
|
||||
# Embree configuration
|
||||
@@ -112,22 +112,10 @@ ELSE()
|
||||
ENDIF()
|
||||
|
||||
OPTION(EMBREE_ISPC_SUPPORT "Build Embree with support for ISPC applications." ON)
|
||||
IF (EMSCRIPTEN)
|
||||
SET(EMBREE_ISPC_SUPPORT OFF CACHE BOOL "Build Embree with support for ISPC applications." FORCE)
|
||||
ENDIF()
|
||||
|
||||
SET(EMBREE_API_NAMESPACE "" CACHE STRING "C++ namespace to put API symbols into.")
|
||||
SET(EMBREE_LIBRARY_NAME "embree${EMBREE_VERSION_MAJOR}" CACHE STRING "Name of the embree library file (default is embree${EMBREE_VERSION_MAJOR})")
|
||||
|
||||
IF (EMBREE_API_NAMESPACE AND EMBREE_ISPC_SUPPORT)
|
||||
MESSAGE(FATAL_ERROR "You cannot enable ISPC when EMBREE_API_NAMESPACE is used.")
|
||||
ENDIF()
|
||||
OPTION(EMBREE_TUTORIALS "Enable to build Embree tutorials" ON)
|
||||
|
||||
##############################################################################
|
||||
# Configurations (add configurations also to common/cmake/embree-config.cmake)
|
||||
|
||||
OPTION(EMBREE_ZIP_MODE "Create Embree ZIP package" OFF)
|
||||
MARK_AS_ADVANCED(EMBREE_ZIP_MODE)
|
||||
# Configurations (add configurations also to common/cmake/embree-config-default.cmake)
|
||||
|
||||
OPTION(EMBREE_INSTALL_DEPENDENCIES "Install Embree dependencies in binary packages and install" OFF)
|
||||
MARK_AS_ADVANCED(EMBREE_INSTALL_DEPENDENCIES)
|
||||
@@ -139,7 +127,6 @@ OPTION(EMBREE_RAY_MASK "Enables ray mask support.")
|
||||
OPTION(EMBREE_BACKFACE_CULLING "Enables backface culling.")
|
||||
OPTION(EMBREE_FILTER_FUNCTION "Enables filter functions." ON)
|
||||
OPTION(EMBREE_IGNORE_INVALID_RAYS "Ignores invalid rays." OFF) # FIXME: enable by default?
|
||||
OPTION(EMBREE_COMPACT_POLYS "Enables double indexed poly layout." OFF)
|
||||
|
||||
OPTION(EMBREE_GEOMETRY_TRIANGLE "Enables support for triangle geometries." ON)
|
||||
OPTION(EMBREE_GEOMETRY_QUAD "Enables support for quad geometries." ON)
|
||||
@@ -148,30 +135,11 @@ OPTION(EMBREE_GEOMETRY_SUBDIVISION "Enables support for subdiv geometries." ON)
|
||||
OPTION(EMBREE_GEOMETRY_USER "Enables support for user geometries." ON)
|
||||
OPTION(EMBREE_GEOMETRY_INSTANCE "Enables support for instances." ON)
|
||||
OPTION(EMBREE_GEOMETRY_GRID "Enables support for grid geometries." ON)
|
||||
OPTION(EMBREE_GEOMETRY_POINT "Enables support for point geometries." ON)
|
||||
OPTION(EMBREE_RAY_PACKETS "Enabled support for ray packets." ON)
|
||||
|
||||
SET(EMBREE_MAX_INSTANCE_LEVEL_COUNT 1 CACHE STRING "Maximum number of instance levels.")
|
||||
SET(EMBREE_CURVE_SELF_INTERSECTION_AVOIDANCE_FACTOR 2.0 CACHE STRING "Self intersection avoidance factor for flat curves. Specify floating point value in range 0 to inf.")
|
||||
OPTION(EMBREE_DISC_POINT_SELF_INTERSECTION_AVOIDANCE "Enables self intersection avoidance for ray oriented discs." ON)
|
||||
OPTION(EMBREE_MIN_WIDTH "Enables min-width feature to enlarge curve and point thickness to pixel width." OFF)
|
||||
|
||||
##############################################################
|
||||
# Platform detection and defaults
|
||||
##############################################################
|
||||
|
||||
# detect ARM compilation
|
||||
IF (APPLE AND CMAKE_SYSTEM_NAME STREQUAL "Darwin" AND (CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64" OR CMAKE_OSX_ARCHITECTURES MATCHES "arm64"))
|
||||
MESSAGE(STATUS "Building for Apple silicon")
|
||||
SET(EMBREE_ARM ON)
|
||||
ELSEIF(CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64")
|
||||
MESSAGE(STATUS "Building for AArch64")
|
||||
SET(EMBREE_ARM ON)
|
||||
ENDIF()
|
||||
|
||||
SET(EMBREE_TASKING_SYSTEM "TBB" CACHE STRING "Selects tasking system")
|
||||
SET(EMBREE_TBB_COMPONENT "tbb" CACHE STRING "The TBB component/library name.")
|
||||
|
||||
IF (WIN32)
|
||||
SET_PROPERTY(CACHE EMBREE_TASKING_SYSTEM PROPERTY STRINGS TBB INTERNAL PPL)
|
||||
ELSE()
|
||||
@@ -198,6 +166,11 @@ ELSE()
|
||||
LIST(APPEND ISPC_DEFINITIONS -DTASKING_INTERNAL)
|
||||
ENDIF()
|
||||
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/config.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/config.h"
|
||||
)
|
||||
|
||||
##############################################################
|
||||
# Compiler
|
||||
##############################################################
|
||||
@@ -211,32 +184,16 @@ IF (WIN32)
|
||||
ENDIF()
|
||||
SET_PROPERTY(GLOBAL PROPERTY USE_FOLDERS ON)
|
||||
|
||||
#message ("CMAKE_GENERATOR_TOOLSET: ${CMAKE_GENERATOR_TOOLSET}")
|
||||
#message ("CMAKE_CXX_COMPILER_ID: ${CMAKE_CXX_COMPILER_ID}")
|
||||
#message ("CMAKE_CXX_COMPILER: ${CMAKE_CXX_COMPILER}")
|
||||
|
||||
IF (${CMAKE_CXX_COMPILER} MATCHES ".*dpcpp")
|
||||
message(FATAL_ERROR "DPCPP compiler not supported on Windows yet. Use ICX instead.")
|
||||
ENDIF()
|
||||
|
||||
IF (${CMAKE_CXX_COMPILER_ID} MATCHES "IntelLLVM" OR
|
||||
(${CMAKE_CXX_COMPILER_ID} MATCHES "Clang" AND ${CMAKE_CXX_COMPILER} MATCHES ".*icx"))
|
||||
MESSAGE("-- DPCPP detected")
|
||||
INCLUDE(dpcpp)
|
||||
ELSEIF(${CMAKE_GENERATOR_TOOLSET} MATCHES "^LLVM")
|
||||
MESSAGE("-- CLANG detected")
|
||||
IF (${CMAKE_GENERATOR_TOOLSET} MATCHES "^LLVM" )
|
||||
MESSAGE("CLANG detected")
|
||||
INCLUDE(clang)
|
||||
ELSEIF(${CMAKE_GENERATOR_TOOLSET} MATCHES "^Intel")
|
||||
MESSAGE("-- Intel Compiler detected")
|
||||
MESSAGE("Intel Compiler detected")
|
||||
INCLUDE (intel)
|
||||
ELSEIF(${CMAKE_CXX_COMPILER_FRONTEND_VARIANT} MATCHES "GNU")
|
||||
MESSAGE("-- GNU-like Compiler detected")
|
||||
INCLUDE(gnu)
|
||||
ELSE()
|
||||
IF (EMBREE_ISA_AVX512)
|
||||
IF (EMBREE_ISA_AVX512KNL OR EMBREE_ISA_AVX512SKX)
|
||||
MESSAGE(FATAL_ERROR "Microsoft Visual C++ Compiler does not support AVX512. Please use Intel Compiler or Clang.")
|
||||
ENDIF()
|
||||
MESSAGE("-- MSVC detected")
|
||||
INCLUDE (msvc)
|
||||
ENDIF()
|
||||
|
||||
@@ -244,19 +201,8 @@ ELSE (WIN32)
|
||||
IF(CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
INCLUDE (crayprgenv)
|
||||
ELSE()
|
||||
GET_FILENAME_COMPONENT(CXX_COMPILER_NAME ${CMAKE_CXX_COMPILER} NAME)
|
||||
|
||||
STRING(TOLOWER "${CMAKE_CXX_COMPILER_ID}" _LOWER_CXX_COMPILER_ID)
|
||||
STRING(REPLACE "appleclang" "clang" _LOWER_CXX_COMPILER_ID ${_LOWER_CXX_COMPILER_ID})
|
||||
STRING(REPLACE "intelllvm" "dpcpp" _LOWER_CXX_COMPILER_ID ${_LOWER_CXX_COMPILER_ID})
|
||||
IF(${_LOWER_CXX_COMPILER_ID} MATCHES "clang" AND ${CXX_COMPILER_NAME} MATCHES "icpx")
|
||||
STRING(REPLACE "clang" "dpcpp" _LOWER_CXX_COMPILER_ID ${_LOWER_CXX_COMPILER_ID})
|
||||
ENDIF()
|
||||
IF(${_LOWER_CXX_COMPILER_ID} MATCHES "clang" AND ${CXX_COMPILER_NAME} MATCHES "dpcpp")
|
||||
STRING(REPLACE "clang" "dpcpp" _LOWER_CXX_COMPILER_ID ${_LOWER_CXX_COMPILER_ID})
|
||||
ENDIF()
|
||||
STRING(TOUPPER "${_LOWER_CXX_COMPILER_ID}" _UPPER_CXX_COMPILER_ID)
|
||||
MESSAGE("-- ${_UPPER_CXX_COMPILER_ID} detected")
|
||||
INCLUDE(${_LOWER_CXX_COMPILER_ID} OPTIONAL RESULT_VARIABLE COMPILER_FOUND)
|
||||
IF (NOT COMPILER_FOUND)
|
||||
MESSAGE(FATAL_ERROR "Unsupported compiler: " ${CMAKE_CXX_COMPILER_ID})
|
||||
@@ -277,92 +223,57 @@ ENDIF (WIN32)
|
||||
# ISA configuration
|
||||
##############################################################
|
||||
|
||||
# just for compatibility with old naming
|
||||
IF(DEFINED EMBREE_ISA_AVX512SKX)
|
||||
UNSET(EMBREE_ISA_AVX512 CACHE)
|
||||
SET(EMBREE_ISA_AVX512 ${EMBREE_ISA_AVX512SKX} CACHE BOOL "")
|
||||
ENDIF()
|
||||
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_ARM_NEON "${CMAKE_BINARY_DIR}" "${PROJECT_SOURCE_DIR}/common/cmake/check_arm_neon.cpp")
|
||||
IF (COMPILER_SUPPORTS_ARM_NEON)
|
||||
SET(EMBREE_ARM ON)
|
||||
ENDIF()
|
||||
|
||||
IF (CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
IF(CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
SET(EMBREE_MAX_ISA "DEFAULT" CACHE STRING "Selects highest ISA to support.")
|
||||
ELSEIF (EMSCRIPTEN)
|
||||
SET(EMBREE_MAX_ISA "SSE2" CACHE STRING "Selects highest ISA to support.")
|
||||
ELSE()
|
||||
SET(EMBREE_MAX_ISA "NONE" CACHE STRING "Selects highest ISA to support.")
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
SET_PROPERTY(CACHE EMBREE_MAX_ISA PROPERTY STRINGS NONE NEON NEON2X)
|
||||
ELSE()
|
||||
SET_PROPERTY(CACHE EMBREE_MAX_ISA PROPERTY STRINGS NONE SSE2 SSE4.2 AVX AVX2 AVX512 DEFAULT)
|
||||
ENDIF()
|
||||
SET_PROPERTY(CACHE EMBREE_MAX_ISA PROPERTY STRINGS NONE SSE2 SSE4.2 AVX AVX2 AVX512KNL AVX512SKX DEFAULT)
|
||||
|
||||
IF (EMBREE_MAX_ISA STREQUAL "NONE")
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX ${CMAKE_BINARY_DIR} ${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp COMPILE_DEFINITIONS ${FLAGS_AVX})
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX2 ${CMAKE_BINARY_DIR} ${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp COMPILE_DEFINITIONS ${FLAGS_AVX2})
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX512KNL ${CMAKE_BINARY_DIR} ${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp COMPILE_DEFINITIONS ${FLAGS_AVX512KNL})
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX512SKX ${CMAKE_BINARY_DIR} ${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp COMPILE_DEFINITIONS ${FLAGS_AVX512SKX})
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
IF (APPLE)
|
||||
OPTION(EMBREE_ISA_NEON "Enables NEON ISA." OFF)
|
||||
OPTION(EMBREE_ISA_NEON2X "Enables NEON ISA double pumped." ON)
|
||||
ELSE()
|
||||
OPTION(EMBREE_ISA_NEON "Enables NEON ISA." ON)
|
||||
OPTION(EMBREE_ISA_NEON2X "Enables NEON ISA double pumped." OFF)
|
||||
ENDIF()
|
||||
OPTION(EMBREE_ISA_SSE2 "Enables SSE2 ISA." ON)
|
||||
OPTION(EMBREE_ISA_SSE42 "Enables SSE4.2 ISA." ON)
|
||||
OPTION(EMBREE_ISA_AVX "Enables AVX ISA." ${COMPILER_SUPPORTS_AVX})
|
||||
OPTION(EMBREE_ISA_AVX2 "Enables AVX2 ISA." ${COMPILER_SUPPORTS_AVX2})
|
||||
IF (WIN32 OR APPLE)
|
||||
OPTION(EMBREE_ISA_AVX512KNL "Enables AVX512 ISA for Knights Landing." OFF)
|
||||
OPTION(EMBREE_ISA_AVX512SKX "Enables AVX512 ISA for Skylake." OFF)
|
||||
ELSE()
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX "${CMAKE_BINARY_DIR}" "${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp" COMPILE_DEFINITIONS ${FLAGS_AVX})
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX2 "${CMAKE_BINARY_DIR}" "${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp" COMPILE_DEFINITIONS ${FLAGS_AVX2})
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_AVX512 "${CMAKE_BINARY_DIR}" "${PROJECT_SOURCE_DIR}/common/cmake/check_isa.cpp" COMPILE_DEFINITIONS ${FLAGS_AVX512})
|
||||
|
||||
OPTION(EMBREE_ISA_SSE2 "Enables SSE2 ISA." ON)
|
||||
OPTION(EMBREE_ISA_SSE42 "Enables SSE4.2 ISA." ON)
|
||||
OPTION(EMBREE_ISA_AVX "Enables AVX ISA." ${COMPILER_SUPPORTS_AVX})
|
||||
OPTION(EMBREE_ISA_AVX2 "Enables AVX2 ISA." ${COMPILER_SUPPORTS_AVX2})
|
||||
IF (WIN32 OR APPLE)
|
||||
OPTION(EMBREE_ISA_AVX512 "Enables AVX512 ISA." OFF)
|
||||
ELSE()
|
||||
OPTION(EMBREE_ISA_AVX512 "Enables AVX512 ISA." ${COMPILER_SUPPORTS_AVX512})
|
||||
ENDIF()
|
||||
OPTION(EMBREE_ISA_AVX512KNL "Enables AVX512 ISA for Knights Landing." OFF) # compilation on GCC 8 broken
|
||||
OPTION(EMBREE_ISA_AVX512SKX "Enables AVX512 ISA for Skylake." ${COMPILER_SUPPORTS_AVX512SKX})
|
||||
ENDIF()
|
||||
|
||||
ELSEIF (EMBREE_MAX_ISA STREQUAL "DEFAULT")
|
||||
UNSET(EMBREE_ISA_NEON CACHE)
|
||||
UNSET(EMBREE_ISA_NEON2X CACHE)
|
||||
UNSET(EMBREE_ISA_SSE2 CACHE)
|
||||
UNSET(EMBREE_ISA_SSE42 CACHE)
|
||||
UNSET(EMBREE_ISA_AVX CACHE)
|
||||
UNSET(EMBREE_ISA_AVX2 CACHE)
|
||||
UNSET(EMBREE_ISA_AVX512 CACHE)
|
||||
SET(EMBREE_ISA_NEON OFF)
|
||||
SET(EMBREE_ISA_NEON2X OFF)
|
||||
UNSET(EMBREE_ISA_AVX512KNL CACHE)
|
||||
UNSET(EMBREE_ISA_AVX512SKX CACHE)
|
||||
SET(EMBREE_ISA_SSE2 OFF)
|
||||
SET(EMBREE_ISA_SSE42 OFF)
|
||||
SET(EMBREE_ISA_AVX OFF)
|
||||
SET(EMBREE_ISA_AVX2 OFF)
|
||||
SET(EMBREE_ISA_AVX512 OFF)
|
||||
SET(EMBREE_ISA_AVX512KNL OFF)
|
||||
SET(EMBREE_ISA_AVX512SKX OFF)
|
||||
MESSAGE(STATUS "Detecting default ISA...")
|
||||
INCLUDE(check_isa_default)
|
||||
CHECK_ISA_DEFAULT(EMBREE_ISA_DEFAULT)
|
||||
MESSAGE(STATUS "Detected default ISA: ${EMBREE_ISA_DEFAULT}")
|
||||
SET(EMBREE_ISA_${EMBREE_ISA_DEFAULT} ON)
|
||||
|
||||
ELSE()
|
||||
UNSET(EMBREE_ISA_NEON CACHE)
|
||||
UNSET(EMBREE_ISA_NEON2X CACHE)
|
||||
UNSET(EMBREE_ISA_SSE2 CACHE)
|
||||
UNSET(EMBREE_ISA_SSE42 CACHE)
|
||||
UNSET(EMBREE_ISA_AVX CACHE)
|
||||
UNSET(EMBREE_ISA_AVX2 CACHE)
|
||||
UNSET(EMBREE_ISA_AVX512 CACHE)
|
||||
|
||||
IF(EMBREE_MAX_ISA STREQUAL "NEON")
|
||||
SET(ISA 1)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "NEON2X")
|
||||
SET(ISA 2)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "SSE2")
|
||||
UNSET(EMBREE_ISA_AVX512KNL CACHE)
|
||||
UNSET(EMBREE_ISA_AVX512SKX CACHE)
|
||||
|
||||
IF(EMBREE_MAX_ISA STREQUAL "SSE2")
|
||||
SET(ISA 1)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "SSE4.2")
|
||||
SET(ISA 2)
|
||||
@@ -370,87 +281,45 @@ ELSE()
|
||||
SET(ISA 3)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "AVX2")
|
||||
SET(ISA 4)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "AVX512")
|
||||
SET(ISA 5)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "AVX512SKX") # just for compatibility
|
||||
SET(ISA 5)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "AVX512KNL")
|
||||
SET(ISA 5)
|
||||
ELSEIF(EMBREE_MAX_ISA STREQUAL "AVX512SKX")
|
||||
SET(ISA 6)
|
||||
ELSE()
|
||||
MESSAGE(FATAL_ERROR "Unsupported ISA specified: " ${EMBREE_MAX_ISA})
|
||||
ENDIF()
|
||||
|
||||
SET(EMBREE_ISA_NEON OFF)
|
||||
SET(EMBREE_ISA_NEON2X OFF)
|
||||
SET(EMBREE_ISA_SSE2 OFF)
|
||||
SET(EMBREE_ISA_SSE42 OFF)
|
||||
SET(EMBREE_ISA_AVX OFF)
|
||||
SET(EMBREE_ISA_AVX2 OFF)
|
||||
SET(EMBREE_ISA_AVX512 OFF)
|
||||
SET(EMBREE_ISA_AVX512KNL OFF)
|
||||
SET(EMBREE_ISA_AVX512SKX OFF)
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
IF (ISA GREATER 0)
|
||||
SET(EMBREE_ISA_NEON ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 1)
|
||||
SET(EMBREE_ISA_NEON2X ON)
|
||||
ENDIF ()
|
||||
ELSE()
|
||||
IF (ISA GREATER 0)
|
||||
SET(EMBREE_ISA_SSE2 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 1)
|
||||
SET(EMBREE_ISA_SSE42 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 2)
|
||||
SET(EMBREE_ISA_AVX ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 3)
|
||||
SET(EMBREE_ISA_AVX2 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 4)
|
||||
SET(EMBREE_ISA_AVX512 ON)
|
||||
ENDIF ()
|
||||
ENDIF()
|
||||
IF (ISA GREATER 0)
|
||||
SET(EMBREE_ISA_SSE2 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 1)
|
||||
SET(EMBREE_ISA_SSE42 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 2)
|
||||
SET(EMBREE_ISA_AVX ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 3)
|
||||
SET(EMBREE_ISA_AVX2 ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 4)
|
||||
SET(EMBREE_ISA_AVX512KNL ON)
|
||||
ENDIF ()
|
||||
IF (ISA GREATER 5)
|
||||
SET(EMBREE_ISA_AVX512SKX ON)
|
||||
ENDIF ()
|
||||
ENDIF()
|
||||
|
||||
IF(CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
SET(EMBREE_ISA_SSE42 OFF)
|
||||
ENDIF()
|
||||
|
||||
IF (APPLE AND EMBREE_STATIC_LIB)
|
||||
|
||||
# count number of set ISAs
|
||||
SET(NUMISA 0)
|
||||
IF (EMBREE_ISA_NEON)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_NEON2X)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_SSE2)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_SSE42)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_AVX)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_AVX2)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_AVX512)
|
||||
MATH(EXPR NUMISA "${NUMISA}+1")
|
||||
ENDIF()
|
||||
|
||||
IF (NUMISA GREATER 1)
|
||||
IF (${CMAKE_CXX_COMPILER_ID} MATCHES "Clang")
|
||||
IF (${CMAKE_CXX_COMPILER_VERSION} VERSION_GREATER "9.0.0" OR ${CMAKE_CXX_COMPILER_VERSION} VERSION_EQUAL "9.0.0")
|
||||
MESSAGE(FATAL_ERROR "Using Embree as static library is not supported with AppleClang >= 9.0 when multiple ISAs are selected. Please either build a shared library or enable only one ISA.")
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
##############################################################
|
||||
# ISA configuration continued
|
||||
##############################################################
|
||||
@@ -459,36 +328,15 @@ SET(SSE2 0)
|
||||
SET(SSE42 1)
|
||||
SET(AVX 2)
|
||||
SET(AVX2 3)
|
||||
SET(AVX512 4)
|
||||
SET(AVX512KNL 4)
|
||||
SET(AVX512SKX 5)
|
||||
|
||||
UNSET(FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST -1)
|
||||
SET(ISA_LOWEST_AVX 2)
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
IF (EMBREE_ISA_NEON2X)
|
||||
LIST(APPEND ISPC_TARGETS "neon-i32x8")
|
||||
ELSEIF (EMBREE_ISA_NEON)
|
||||
LIST(APPEND ISPC_TARGETS "neon-i32x4")
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_NEON)
|
||||
SET(EMBREE_ISA_SSE2 ON)
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_NEON2X)
|
||||
SET(EMBREE_ISA_SSE2 OFF)
|
||||
SET(EMBREE_ISA_SSE42 OFF)
|
||||
SET(EMBREE_ISA_AVX OFF)
|
||||
SET(EMBREE_ISA_AVX2 ON)
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_SSE2)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_SSE2)
|
||||
IF (NOT EMBREE_ARM)
|
||||
LIST(APPEND ISPC_TARGETS "sse2")
|
||||
ENDIF()
|
||||
LIST(APPEND ISPC_TARGETS "sse2")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${SSE2})
|
||||
SET(FLAGS_LOWEST ${FLAGS_SSE2})
|
||||
@@ -497,9 +345,7 @@ ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_SSE42)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_SSE42)
|
||||
IF (NOT EMBREE_ARM)
|
||||
LIST(APPEND ISPC_TARGETS "sse4")
|
||||
ENDIF()
|
||||
LIST(APPEND ISPC_TARGETS "sse4")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${SSE42})
|
||||
SET(FLAGS_LOWEST ${FLAGS_SSE42})
|
||||
@@ -508,9 +354,7 @@ ENDIF ()
|
||||
|
||||
IF (EMBREE_ISA_AVX)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_AVX)
|
||||
IF (NOT EMBREE_ARM)
|
||||
LIST(APPEND ISPC_TARGETS "avx")
|
||||
ENDIF()
|
||||
LIST(APPEND ISPC_TARGETS "avx")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${AVX})
|
||||
SET(ISA_LOWEST_AVX ${AVX})
|
||||
@@ -520,9 +364,7 @@ ENDIF ()
|
||||
|
||||
IF (EMBREE_ISA_AVX2)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_AVX2)
|
||||
IF (NOT EMBREE_ARM)
|
||||
LIST(APPEND ISPC_TARGETS "avx2")
|
||||
ENDIF()
|
||||
LIST(APPEND ISPC_TARGETS "avx2")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${AVX2})
|
||||
SET(ISA_LOWEST_AVX ${AVX2})
|
||||
@@ -530,21 +372,24 @@ IF (EMBREE_ISA_AVX2)
|
||||
ENDIF()
|
||||
ENDIF ()
|
||||
|
||||
IF (EMBREE_ISA_AVX512)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_AVX512)
|
||||
IF (NOT EMBREE_ARM)
|
||||
LIST(APPEND ISPC_TARGETS "avx512skx-i32x16")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_AVX512KNL)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_AVX512KNL)
|
||||
LIST(APPEND ISPC_TARGETS "avx512knl-i32x16")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${AVX512})
|
||||
SET(ISA_LOWEST_AVX ${AVX512})
|
||||
SET(FLAGS_LOWEST ${FLAGS_AVX512})
|
||||
SET(ISA_LOWEST ${AVX512KNL})
|
||||
SET(FLAGS_LOWEST ${FLAGS_AVX512KNL})
|
||||
ENDIF()
|
||||
ENDIF ()
|
||||
|
||||
IF (ISA_LOWEST EQUAL -1)
|
||||
MESSAGE(FATAL_ERROR "You have to enable at least one ISA!")
|
||||
ENDIF()
|
||||
IF (EMBREE_ISA_AVX512SKX)
|
||||
ADD_DEFINITIONS(-DEMBREE_TARGET_AVX512SKX)
|
||||
LIST(APPEND ISPC_TARGETS "avx512skx-i32x16")
|
||||
IF(NOT FLAGS_LOWEST)
|
||||
SET(ISA_LOWEST ${AVX512SKX})
|
||||
SET(ISA_LOWEST_AVX ${AVX512SKX})
|
||||
SET(FLAGS_LOWEST ${FLAGS_AVX512SKX})
|
||||
ENDIF()
|
||||
ENDIF ()
|
||||
|
||||
INCLUDE (ispc)
|
||||
|
||||
@@ -554,32 +399,12 @@ INCLUDE (ispc)
|
||||
include(package)
|
||||
|
||||
##############################################################
|
||||
# Create Config files
|
||||
# Search paths
|
||||
##############################################################
|
||||
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/config.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/config.h"
|
||||
)
|
||||
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/rtcore_config.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/include/embree3/rtcore_config.h"
|
||||
)
|
||||
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/hash.h.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/hash.h"
|
||||
)
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/export.linux.map.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/export.linux.map"
|
||||
)
|
||||
CONFIGURE_FILE(
|
||||
"${PROJECT_SOURCE_DIR}/kernels/export.macosx.map.in"
|
||||
"${PROJECT_SOURCE_DIR}/kernels/export.macosx.map"
|
||||
)
|
||||
|
||||
IF (TASKING_TBB)
|
||||
FIND_PACKAGE(TBB REQUIRED)
|
||||
INCLUDE_DIRECTORIES(${TBB_INCLUDE_DIRS})
|
||||
ENDIF()
|
||||
|
||||
##############################################################
|
||||
# Output paths
|
||||
@@ -602,20 +427,17 @@ ENDIF()
|
||||
# Uninstall
|
||||
##############################################################
|
||||
|
||||
IF (NOT TARGET uninstall)
|
||||
configure_file(
|
||||
"${CMAKE_CURRENT_SOURCE_DIR}/common/cmake/uninstall.cmake.in"
|
||||
"${CMAKE_CURRENT_BINARY_DIR}/uninstall.cmake"
|
||||
IMMEDIATE @ONLY)
|
||||
configure_file(
|
||||
"${CMAKE_CURRENT_SOURCE_DIR}/common/cmake/uninstall.cmake.in"
|
||||
"${CMAKE_CURRENT_BINARY_DIR}/uninstall.cmake"
|
||||
IMMEDIATE @ONLY)
|
||||
|
||||
add_custom_target(uninstall
|
||||
COMMAND ${CMAKE_COMMAND} -P "${CMAKE_CURRENT_BINARY_DIR}/uninstall.cmake")
|
||||
add_custom_target(uninstall
|
||||
COMMAND ${CMAKE_COMMAND} -P ${CMAKE_CURRENT_BINARY_DIR}/uninstall.cmake)
|
||||
|
||||
SET_PROPERTY(TARGET uninstall PROPERTY FOLDER CMakePredefinedTargets)
|
||||
ENDIF()
|
||||
SET_PROPERTY(TARGET uninstall PROPERTY FOLDER CMakePredefinedTargets)
|
||||
|
||||
##############################################################
|
||||
# Has to be last
|
||||
##############################################################
|
||||
|
||||
INCLUDE(CPack)
|
||||
|
||||
+31
-3
@@ -1,5 +1,33 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
set(CTEST_PROJECT_NAME "Embree")
|
||||
set(TEST_MODELS_HASH 866df160e2cc1074d68677e1e6e4135f8c324995)
|
||||
set(CTEST_NIGHTLY_START_TIME "22:00:00 UTC")
|
||||
set(TEST_MODELS_HASH ba1bd24186ccdf939d2e749a514f81718f3a2b20)
|
||||
|
||||
IF (NOT CTEST_DROP_SITE)
|
||||
set(CTEST_DROP_METHOD "http")
|
||||
set(CTEST_DROP_SITE "cdash")
|
||||
set(CTEST_DROP_LOCATION "/CDash/submit.php?project=Embree")
|
||||
set(CTEST_DROP_SITE_CDASH TRUE)
|
||||
endif()
|
||||
|
||||
list (APPEND CTEST_CUSTOM_WARNING_EXCEPTION "warning #1478") # deprecated function used
|
||||
#list (APPEND CTEST_CUSTOM_WARNING_EXCEPTION "warning #10237") # -lcilkrts linked in dynamically, static library not available")
|
||||
list (APPEND CTEST_CUSTOM_WARNING_EXCEPTION "-Wextern-initializer")
|
||||
|
||||
SET(CTEST_CUSTOM_MAXIMUM_PASSED_TEST_OUTPUT_SIZE 200000)
|
||||
SET(CTEST_CUSTOM_MAXIMUM_FAILED_TEST_OUTPUT_SIZE 800000)
|
||||
|
||||
-14
@@ -1,14 +0,0 @@
|
||||
Security Policy
|
||||
===============
|
||||
|
||||
Intel is committed to rapidly addressing security vulnerabilities
|
||||
affecting our customers and providing clear guidance on the solution,
|
||||
impact, severity and mitigation.
|
||||
|
||||
Reporting a Vulnerability
|
||||
-------------------------
|
||||
|
||||
Please [report any security vulnerabilities][guidelines] in this project
|
||||
utilizing the [guidelines here][guidelines].
|
||||
|
||||
[guidelines]: https://www.intel.com/content/www/us/en/security-center/vulnerability-handling-guidelines.html "Vulnerability Handling Guidelines"
|
||||
+17
-3
@@ -1,8 +1,22 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
ADD_SUBDIRECTORY(sys)
|
||||
ADD_SUBDIRECTORY(math)
|
||||
ADD_SUBDIRECTORY(simd)
|
||||
ADD_SUBDIRECTORY(lexers)
|
||||
ADD_SUBDIRECTORY(tasking)
|
||||
ADD_SUBDIRECTORY(tasking)
|
||||
ADD_SUBDIRECTORY(algorithms)
|
||||
|
||||
@@ -0,0 +1,31 @@
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
ADD_LIBRARY(algorithms OBJECT
|
||||
parallel_for.cpp
|
||||
parallel_reduce.cpp
|
||||
parallel_prefix_sum.cpp
|
||||
parallel_for_for.cpp
|
||||
parallel_for_for_prefix_sum.cpp
|
||||
parallel_partition.cpp
|
||||
parallel_sort.cpp
|
||||
parallel_set.cpp
|
||||
parallel_map.cpp
|
||||
parallel_filter.cpp
|
||||
)
|
||||
|
||||
SET_PROPERTY(TARGET algorithms PROPERTY FOLDER common)
|
||||
SET_PROPERTY(TARGET algorithms APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
@@ -1,55 +0,0 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#pragma once
|
||||
|
||||
#include <functional>
|
||||
#include "parallel_reduce.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
|
||||
template<typename Index, class UnaryPredicate>
|
||||
__forceinline bool parallel_any_of (Index first, Index last, UnaryPredicate pred)
|
||||
{
|
||||
bool ret = false;
|
||||
|
||||
#if defined(TASKING_TBB)
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
tbb::parallel_for(tbb::blocked_range<size_t>{first, last}, [&ret,pred,&context](const tbb::blocked_range<size_t>& r) {
|
||||
if (context.is_group_execution_cancelled()) return;
|
||||
for (size_t i = r.begin(); i != r.end(); ++i) {
|
||||
if (pred(i)) {
|
||||
ret = true;
|
||||
context.cancel_group_execution();
|
||||
}
|
||||
}
|
||||
});
|
||||
#else
|
||||
tbb::parallel_for(tbb::blocked_range<size_t>{first, last}, [&ret,pred](const tbb::blocked_range<size_t>& r) {
|
||||
if (tbb::task::self().is_cancelled()) return;
|
||||
for (size_t i = r.begin(); i != r.end(); ++i) {
|
||||
if (pred(i)) {
|
||||
ret = true;
|
||||
tbb::task::self().cancel_group_execution();
|
||||
}
|
||||
}
|
||||
});
|
||||
#endif
|
||||
#else
|
||||
ret = parallel_reduce (first, last, false, [pred](const range<size_t>& r)->bool {
|
||||
bool localret = false;
|
||||
for (auto i=r.begin(); i<r.end(); ++i) {
|
||||
localret |= pred(i);
|
||||
}
|
||||
return localret;
|
||||
},
|
||||
std::bit_or<bool>()
|
||||
);
|
||||
#endif
|
||||
|
||||
return ret;
|
||||
}
|
||||
|
||||
} // end namespace
|
||||
@@ -0,0 +1,69 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_filter.h"
|
||||
#include "../sys/regression.h"
|
||||
#include <map>
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_filter_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_filter_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
auto pred = [&]( uint32_t v ) { return (v & 0x3) == 0; };
|
||||
|
||||
for (size_t N=10; N<1000000; N=size_t(2.1*N))
|
||||
{
|
||||
size_t N0 = rand() % N;
|
||||
|
||||
/* initialize array with random numbers */
|
||||
std::vector<uint32_t> src(N);
|
||||
std::map<uint32_t,int> m;
|
||||
for (size_t i=0; i<N; i++) src[i] = rand();
|
||||
|
||||
/* count elements up */
|
||||
for (size_t i=N0; i<N; i++)
|
||||
if (pred(src[i]))
|
||||
m[src[i]] = 0;
|
||||
for (size_t i=N0; i<N; i++)
|
||||
if (pred(src[i]))
|
||||
m[src[i]]++;
|
||||
|
||||
/* filter array */
|
||||
//size_t M = sequential_filter(src.data(),N0,N,pred);
|
||||
size_t M = parallel_filter(src.data(),N0,N,size_t(1024),pred);
|
||||
|
||||
/* check if filtered data is correct */
|
||||
for (size_t i=N0; i<M; i++) {
|
||||
passed &= pred(src[i]);
|
||||
m[src[i]]--;
|
||||
}
|
||||
for (size_t i=N0; i<M; i++)
|
||||
passed &= (m[src[i]] == 0);
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_filter_regression_test parallel_filter_regression("parallel_filter_regression");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -0,0 +1,61 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_for.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_for_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_for_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
const size_t M = 10;
|
||||
for (size_t N=10; N<10000000; N=size_t(2.1*N))
|
||||
{
|
||||
/* sequentially calculate sum of squares */
|
||||
size_t sum0 = 0;
|
||||
for (size_t i=0; i<N; i++) {
|
||||
sum0 += i*i;
|
||||
}
|
||||
|
||||
/* parallel calculation of sum of squares */
|
||||
for (size_t m=0; m<M; m++)
|
||||
{
|
||||
std::atomic<size_t> sum1(0);
|
||||
parallel_for( size_t(0), size_t(N), size_t(1024), [&](const range<size_t>& r)
|
||||
{
|
||||
size_t s = 0;
|
||||
for (size_t i=r.begin(); i<r.end(); i++)
|
||||
s += i*i;
|
||||
sum1 += s;
|
||||
});
|
||||
passed = sum0 == sum1;
|
||||
}
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_for_regression_test parallel_for_regression("parallel_for_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -23,21 +36,13 @@ namespace embree
|
||||
if (!TaskScheduler::wait())
|
||||
throw std::runtime_error("task cancelled");
|
||||
}
|
||||
|
||||
#elif defined(TASKING_TBB)
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},context);
|
||||
if (context.is_group_execution_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#else
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
});
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#endif
|
||||
|
||||
#elif defined(TASKING_PPL)
|
||||
concurrency::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
@@ -59,20 +64,11 @@ namespace embree
|
||||
throw std::runtime_error("task cancelled");
|
||||
|
||||
#elif defined(TASKING_TBB)
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
tbb::parallel_for(tbb::blocked_range<Index>(first,last,minStepSize),[&](const tbb::blocked_range<Index>& r) {
|
||||
func(range<Index>(r.begin(),r.end()));
|
||||
},context);
|
||||
if (context.is_group_execution_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#else
|
||||
tbb::parallel_for(tbb::blocked_range<Index>(first,last,minStepSize),[&](const tbb::blocked_range<Index>& r) {
|
||||
tbb::parallel_for(tbb::blocked_range<Index>(first,last,minStepSize),[&](const tbb::blocked_range<Index>& r) {
|
||||
func(range<Index>(r.begin(),r.end()));
|
||||
});
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#endif
|
||||
|
||||
#elif defined(TASKING_PPL)
|
||||
concurrency::parallel_for(first, last, Index(1) /*minStepSize*/, [&](Index i) {
|
||||
@@ -97,20 +93,11 @@ namespace embree
|
||||
template<typename Index, typename Func>
|
||||
__forceinline void parallel_for_static( const Index N, const Func& func)
|
||||
{
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},tbb::simple_partitioner(),context);
|
||||
if (context.is_group_execution_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#else
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},tbb::simple_partitioner());
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#endif
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},tbb::simple_partitioner());
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
}
|
||||
|
||||
typedef tbb::affinity_partitioner affinity_partitioner;
|
||||
@@ -118,20 +105,11 @@ namespace embree
|
||||
template<typename Index, typename Func>
|
||||
__forceinline void parallel_for_affinity( const Index N, const Func& func, tbb::affinity_partitioner& ap)
|
||||
{
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},ap,context);
|
||||
if (context.is_group_execution_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#else
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},ap);
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
#endif
|
||||
tbb::parallel_for(Index(0),N,Index(1),[&](Index i) {
|
||||
func(i);
|
||||
},ap);
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
}
|
||||
|
||||
#else
|
||||
|
||||
@@ -0,0 +1,76 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_for_for.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_for_for_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_for_for_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
/* create vector with random numbers */
|
||||
size_t sum0 = 0;
|
||||
size_t K = 0;
|
||||
const size_t M = 1000;
|
||||
std::vector<std::vector<size_t>* > array2(M);
|
||||
for (size_t i=0; i<M; i++) {
|
||||
const size_t N = rand() % 1024;
|
||||
K+=N;
|
||||
array2[i] = new std::vector<size_t>(N);
|
||||
for (size_t j=0; j<N; j++)
|
||||
sum0 += (*array2[i])[j] = rand();
|
||||
}
|
||||
|
||||
/* array to test global index */
|
||||
std::vector<atomic<size_t>> verify_k(K);
|
||||
for (size_t i=0; i<K; i++) verify_k[i].store(0);
|
||||
|
||||
/* add all numbers using parallel_for_for */
|
||||
std::atomic<size_t> sum1(0);
|
||||
parallel_for_for( array2, size_t(1), [&](std::vector<size_t>* v, const range<size_t>& r, size_t k) -> size_t
|
||||
{
|
||||
size_t s = 0;
|
||||
for (size_t i=r.begin(); i<r.end(); i++) {
|
||||
s += (*v)[i];
|
||||
verify_k[k++]++;
|
||||
}
|
||||
sum1 += s;
|
||||
return sum1;
|
||||
});
|
||||
passed &= (sum0 == sum1);
|
||||
|
||||
/* check global index */
|
||||
for (size_t i=0; i<K; i++)
|
||||
passed &= (verify_k[i] == 1);
|
||||
|
||||
/* delete vectors again */
|
||||
for (size_t i=0; i<array2.size(); i++)
|
||||
delete array2[i];
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_for_for_regression_test parallel_for_for_regression("parallel_for_for_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -30,20 +43,15 @@ namespace embree
|
||||
template<typename ArrayArray>
|
||||
__forceinline ParallelForForState (ArrayArray& array2, const size_t minStepSize) {
|
||||
init(array2,minStepSize);
|
||||
}
|
||||
|
||||
template<typename SizeFunc>
|
||||
__forceinline ParallelForForState (const size_t numArrays, const SizeFunc& getSize, const size_t minStepSize) {
|
||||
init(numArrays,getSize,minStepSize);
|
||||
}
|
||||
|
||||
template<typename SizeFunc>
|
||||
__forceinline void init ( const size_t numArrays, const SizeFunc& getSize, const size_t minStepSize )
|
||||
template<typename ArrayArray>
|
||||
__forceinline void init ( ArrayArray& array2, const size_t minStepSize )
|
||||
{
|
||||
/* first calculate total number of elements */
|
||||
size_t N = 0;
|
||||
for (size_t i=0; i<numArrays; i++) {
|
||||
N += getSize(i);
|
||||
for (size_t i=0; i<array2.size(); i++) {
|
||||
N += array2[i] ? array2[i]->size() : 0;
|
||||
}
|
||||
this->N = N;
|
||||
|
||||
@@ -59,8 +67,8 @@ namespace embree
|
||||
size_t k0 = (++taskIndex)*N/taskCount;
|
||||
for (size_t i=0, k=0; taskIndex < taskCount; i++)
|
||||
{
|
||||
assert(i<numArrays);
|
||||
size_t j=0, M = getSize(i);
|
||||
assert(i<array2.size());
|
||||
size_t j=0, M = array2[i] ? array2[i]->size() : 0;
|
||||
while (j<M && k+M-j >= k0 && taskIndex < taskCount) {
|
||||
assert(taskIndex<taskCount);
|
||||
i0[taskIndex] = i;
|
||||
@@ -72,12 +80,6 @@ namespace embree
|
||||
}
|
||||
}
|
||||
|
||||
template<typename ArrayArray>
|
||||
__forceinline void init ( ArrayArray& array2, const size_t minStepSize )
|
||||
{
|
||||
init(array2.size(),[&](size_t i) { return array2[i] ? array2[i]->size() : 0; },minStepSize);
|
||||
}
|
||||
|
||||
__forceinline size_t size() const {
|
||||
return N;
|
||||
}
|
||||
|
||||
@@ -0,0 +1,98 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_for_for_prefix_sum.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_for_for_prefix_sum_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_for_for_prefix_sum_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
/* create vector with random numbers */
|
||||
const size_t M = 10;
|
||||
std::vector<atomic<size_t>> flattened;
|
||||
typedef std::vector<std::vector<size_t>* > ArrayArray;
|
||||
ArrayArray array2(M);
|
||||
size_t K = 0;
|
||||
for (size_t i=0; i<M; i++) {
|
||||
const size_t N = rand() % 10;
|
||||
K += N;
|
||||
array2[i] = new std::vector<size_t>(N);
|
||||
for (size_t j=0; j<N; j++)
|
||||
(*array2[i])[j] = rand() % 10;
|
||||
}
|
||||
|
||||
/* array to test global index */
|
||||
std::vector<atomic<size_t>> verify_k(K);
|
||||
for (size_t i=0; i<K; i++) verify_k[i].store(0);
|
||||
|
||||
ParallelForForPrefixSumState<size_t> state(array2,size_t(1));
|
||||
|
||||
/* dry run only counts */
|
||||
size_t S = parallel_for_for_prefix_sum0( state, array2, size_t(0), [&](std::vector<size_t>* v, const range<size_t>& r, size_t k) -> size_t
|
||||
{
|
||||
size_t s = 0;
|
||||
for (size_t i=r.begin(); i<r.end(); i++) {
|
||||
s += (*v)[i];
|
||||
verify_k[k++]++;
|
||||
}
|
||||
return s;
|
||||
}, [](size_t v0, size_t v1) { return v0+v1; });
|
||||
|
||||
/* create properly sized output array */
|
||||
flattened.resize(S);
|
||||
for (auto& a : flattened) a.store(0);
|
||||
|
||||
/* now we actually fill the flattened array */
|
||||
parallel_for_for_prefix_sum1( state, array2, size_t(0), [&](std::vector<size_t>* v, const range<size_t>& r, size_t k, const size_t base) -> size_t
|
||||
{
|
||||
size_t s = 0;
|
||||
for (size_t i=r.begin(); i<r.end(); i++) {
|
||||
for (size_t j=0; j<(*v)[i]; j++) {
|
||||
flattened[base+s+j]++;
|
||||
}
|
||||
s += (*v)[i];
|
||||
verify_k[k++]++;
|
||||
}
|
||||
return s;
|
||||
}, [](size_t v0, size_t v1) { return v0+v1; });
|
||||
|
||||
/* check global index */
|
||||
for (size_t i=0; i<K; i++)
|
||||
passed &= (verify_k[i] == 2);
|
||||
|
||||
/* check if each element was assigned exactly once */
|
||||
for (size_t i=0; i<flattened.size(); i++)
|
||||
passed &= (flattened[i] == 1);
|
||||
|
||||
/* delete arrays again */
|
||||
for (size_t i=0; i<array2.size(); i++)
|
||||
delete array2[i];
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_for_for_prefix_sum_regression_test parallel_for_for_prefix_sum_regression("parallel_for_for_prefix_sum_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -17,56 +30,12 @@ namespace embree
|
||||
__forceinline ParallelForForPrefixSumState (ArrayArray& array2, const size_t minStepSize)
|
||||
: ParallelForForState(array2,minStepSize) {}
|
||||
|
||||
template<typename SizeFunc>
|
||||
__forceinline ParallelForForPrefixSumState (size_t numArrays, const SizeFunc& getSize, const size_t minStepSize)
|
||||
: ParallelForForState(numArrays,getSize,minStepSize) {}
|
||||
|
||||
ParallelPrefixSumState<Value> prefix_state;
|
||||
};
|
||||
|
||||
template<typename SizeFunc, typename Index, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum0_( ParallelForForPrefixSumState<Value>& state, Index minStepSize,
|
||||
const SizeFunc& getSize, const Value& identity, const Func& func, const Reduction& reduction)
|
||||
{
|
||||
/* calculate number of tasks to use */
|
||||
const size_t taskCount = state.taskCount;
|
||||
|
||||
/* perform parallel prefix sum */
|
||||
parallel_for(taskCount, [&](const size_t taskIndex)
|
||||
{
|
||||
const size_t k0 = (taskIndex+0)*state.size()/taskCount;
|
||||
const size_t k1 = (taskIndex+1)*state.size()/taskCount;
|
||||
size_t i0 = state.i0[taskIndex];
|
||||
size_t j0 = state.j0[taskIndex];
|
||||
|
||||
/* iterate over arrays */
|
||||
size_t k=k0;
|
||||
Value N=identity;
|
||||
for (size_t i=i0; k<k1; i++) {
|
||||
const size_t size = getSize(i);
|
||||
const size_t r0 = j0, r1 = min(size,r0+k1-k);
|
||||
if (r1 > r0) N = reduction(N, func((Index)i,range<Index>((Index)r0,(Index)r1),(Index)k));
|
||||
k+=r1-r0; j0 = 0;
|
||||
}
|
||||
state.prefix_state.counts[taskIndex] = N;
|
||||
});
|
||||
|
||||
/* calculate prefix sum */
|
||||
Value sum=identity;
|
||||
for (size_t i=0; i<taskCount; i++)
|
||||
{
|
||||
const Value c = state.prefix_state.counts[i];
|
||||
state.prefix_state.sums[i] = sum;
|
||||
sum=reduction(sum,c);
|
||||
}
|
||||
|
||||
return sum;
|
||||
}
|
||||
|
||||
template<typename SizeFunc, typename Index, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum1_( ParallelForForPrefixSumState<Value>& state, Index minStepSize,
|
||||
const SizeFunc& getSize,
|
||||
const Value& identity, const Func& func, const Reduction& reduction)
|
||||
template<typename ArrayArray, typename Index, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum0( ParallelForForPrefixSumState<Value>& state, ArrayArray& array2, Index minStepSize,
|
||||
const Value& identity, const Func& func, const Reduction& reduction)
|
||||
{
|
||||
/* calculate number of tasks to use */
|
||||
const size_t taskCount = state.taskCount;
|
||||
@@ -82,9 +51,9 @@ namespace embree
|
||||
size_t k=k0;
|
||||
Value N=identity;
|
||||
for (size_t i=i0; k<k1; i++) {
|
||||
const size_t size = getSize(i);
|
||||
const size_t size = array2[i] ? array2[i]->size() : 0;
|
||||
const size_t r0 = j0, r1 = min(size,r0+k1-k);
|
||||
if (r1 > r0) N = reduction(N, func((Index)i,range<Index>((Index)r0,(Index)r1),(Index)k,reduction(state.prefix_state.sums[taskIndex],N)));
|
||||
if (r1 > r0) N = reduction(N, func(array2[i],range<size_t>(r0,r1),k));
|
||||
k+=r1-r0; j0 = 0;
|
||||
}
|
||||
state.prefix_state.counts[taskIndex] = N;
|
||||
@@ -103,28 +72,42 @@ namespace embree
|
||||
}
|
||||
|
||||
template<typename ArrayArray, typename Index, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum0( ParallelForForPrefixSumState<Value>& state,
|
||||
ArrayArray& array2, Index minStepSize,
|
||||
const Value& identity, const Func& func, const Reduction& reduction)
|
||||
__forceinline Value parallel_for_for_prefix_sum1( ParallelForForPrefixSumState<Value>& state, ArrayArray& array2, Index minStepSize,
|
||||
const Value& identity, const Func& func, const Reduction& reduction)
|
||||
{
|
||||
return parallel_for_for_prefix_sum0_(state,minStepSize,
|
||||
[&](Index i) { return array2[i] ? array2[i]->size() : 0; },
|
||||
identity,
|
||||
[&](Index i, const range<Index>& r, Index k) { return func(array2[i], r, k, i); },
|
||||
reduction);
|
||||
}
|
||||
/* calculate number of tasks to use */
|
||||
const size_t taskCount = state.taskCount;
|
||||
/* perform parallel prefix sum */
|
||||
parallel_for(taskCount, [&](const size_t taskIndex)
|
||||
{
|
||||
const size_t k0 = (taskIndex+0)*state.size()/taskCount;
|
||||
const size_t k1 = (taskIndex+1)*state.size()/taskCount;
|
||||
size_t i0 = state.i0[taskIndex];
|
||||
size_t j0 = state.j0[taskIndex];
|
||||
|
||||
template<typename ArrayArray, typename Index, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum1( ParallelForForPrefixSumState<Value>& state,
|
||||
ArrayArray& array2, Index minStepSize,
|
||||
const Value& identity, const Func& func, const Reduction& reduction)
|
||||
{
|
||||
return parallel_for_for_prefix_sum1_(state,minStepSize,
|
||||
[&](Index i) { return array2[i] ? array2[i]->size() : 0; },
|
||||
identity,
|
||||
[&](Index i, const range<Index>& r, Index k, const Value& base) { return func(array2[i], r, k, i, base); },
|
||||
reduction);
|
||||
}
|
||||
/* iterate over arrays */
|
||||
size_t k=k0;
|
||||
Value N=identity;
|
||||
for (size_t i=i0; k<k1; i++) {
|
||||
const size_t size = array2[i] ? array2[i]->size() : 0;
|
||||
const size_t r0 = j0, r1 = min(size,r0+k1-k);
|
||||
if (r1 > r0) N = reduction(N, func(array2[i],range<size_t>(r0,r1),k,reduction(state.prefix_state.sums[taskIndex],N)));
|
||||
k+=r1-r0; j0 = 0;
|
||||
}
|
||||
state.prefix_state.counts[taskIndex] = N;
|
||||
});
|
||||
|
||||
/* calculate prefix sum */
|
||||
Value sum=identity;
|
||||
for (size_t i=0; i<taskCount; i++)
|
||||
{
|
||||
const Value c = state.prefix_state.counts[i];
|
||||
state.prefix_state.sums[i] = sum;
|
||||
sum=reduction(sum,c);
|
||||
}
|
||||
|
||||
return sum;
|
||||
}
|
||||
|
||||
template<typename ArrayArray, typename Value, typename Func, typename Reduction>
|
||||
__forceinline Value parallel_for_for_prefix_sum0( ParallelForForPrefixSumState<Value>& state, ArrayArray& array2,
|
||||
|
||||
@@ -0,0 +1,60 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_map.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_map_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_map_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
/* create key/value vectors with random numbers */
|
||||
const size_t N = 10000;
|
||||
std::vector<uint32_t> keys(N);
|
||||
std::vector<uint32_t> vals(N);
|
||||
for (size_t i=0; i<N; i++) keys[i] = 2*unsigned(i)*647382649;
|
||||
for (size_t i=0; i<N; i++) std::swap(keys[i],keys[rand()%N]);
|
||||
for (size_t i=0; i<N; i++) vals[i] = 2*rand();
|
||||
|
||||
/* create map */
|
||||
parallel_map<uint32_t,uint32_t> map;
|
||||
map.init(keys,vals);
|
||||
|
||||
/* check that all keys are properly mapped */
|
||||
for (size_t i=0; i<N; i++) {
|
||||
const uint32_t* val = map.lookup(keys[i]);
|
||||
passed &= val && (*val == vals[i]);
|
||||
}
|
||||
|
||||
/* check that these keys are not in the map */
|
||||
for (size_t i=0; i<N; i++) {
|
||||
passed &= !map.lookup(keys[i]+1);
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_map_regression_test parallel_map_regression("parallel_map_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -0,0 +1,66 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_partition.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_partition_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_partition_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
for (size_t i=0; i<100; i++)
|
||||
{
|
||||
/* create random permutation */
|
||||
size_t N = std::rand() % 1000000;
|
||||
std::vector<unsigned> array(N);
|
||||
for (unsigned i=0; i<N; i++) array[i] = i;
|
||||
for (auto& v : array) std::swap(v,array[std::rand()%array.size()]);
|
||||
size_t split = std::rand() % (N+1);
|
||||
|
||||
/* perform parallel partitioning */
|
||||
size_t left_sum = 0, right_sum = 0;
|
||||
size_t mid = parallel_partitioning(array.data(),0,array.size(),0,left_sum,right_sum,
|
||||
[&] ( size_t i ) { return i < split; },
|
||||
[] ( size_t& sum, unsigned v) { sum += v; },
|
||||
[] ( size_t& sum, size_t v) { sum += v; },
|
||||
128);
|
||||
|
||||
/*serial_partitioning(array.data(),0,array.size(),left_sum,right_sum,
|
||||
[&] ( size_t i ) { return i < split; },
|
||||
[] ( size_t& left_sum, int v) { left_sum += v; });*/
|
||||
|
||||
/* verify result */
|
||||
passed &= mid == split;
|
||||
passed &= left_sum == split*(split-1)/2;
|
||||
passed &= right_sum == N*(N-1)/2-left_sum;
|
||||
for (size_t i=0; i<split; i++) passed &= array[i] < split;
|
||||
for (size_t i=split; i<N; i++) passed &= array[i] >= split;
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_partition_regression_test parallel_partition_regression("parallel_partition_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -262,22 +275,4 @@ namespace embree
|
||||
return begin+p->partition(leftReduction,rightReduction);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
template<typename T, typename IsLeft>
|
||||
inline size_t parallel_partitioning(T* array,
|
||||
const size_t begin,
|
||||
const size_t end,
|
||||
const IsLeft& is_left,
|
||||
size_t BLOCK_SIZE = 128)
|
||||
{
|
||||
size_t leftReduction = 0;
|
||||
size_t rightReduction = 0;
|
||||
return parallel_partitioning(
|
||||
array,begin,end,0,leftReduction,rightReduction,is_left,
|
||||
[] (size_t& t,const T& ref) { },
|
||||
[] (size_t& t0,size_t& t1) { },
|
||||
BLOCK_SIZE);
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
@@ -0,0 +1,61 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_prefix_sum.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_prefix_sum_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_prefix_sum_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
const size_t M = 10;
|
||||
|
||||
for (size_t N=10; N<10000000; N=size_t(2.1*N))
|
||||
{
|
||||
/* initialize array with random numbers */
|
||||
uint32_t sum0 = 0;
|
||||
std::vector<uint32_t> src(N);
|
||||
for (size_t i=0; i<N; i++) {
|
||||
sum0 += src[i] = rand();
|
||||
}
|
||||
|
||||
/* calculate parallel prefix sum */
|
||||
std::vector<uint32_t> dst(N);
|
||||
for (auto& v : dst) v = 0;
|
||||
|
||||
for (size_t i=0; i<M; i++) {
|
||||
uint32_t sum1 = parallel_prefix_sum(src,dst,N,0,std::plus<uint32_t>());
|
||||
passed &= (sum0 == sum1);
|
||||
}
|
||||
|
||||
/* check if prefix sum is correct */
|
||||
for (size_t i=0, sum=0; i<N; sum+=src[i++])
|
||||
passed &= ((uint32_t)sum == dst[i]);
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_prefix_sum_regression_test parallel_prefix_sum_regression("parallel_prefix_sum_regression");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -0,0 +1,62 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_reduce.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_reduce_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_reduce_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
const size_t M = 10;
|
||||
for (size_t N=10; N<10000000; N=size_t(2.1*N))
|
||||
{
|
||||
/* sequentially calculate sum of squares */
|
||||
size_t sum0 = 0;
|
||||
for (size_t i=0; i<N; i++) {
|
||||
sum0 += i*i;
|
||||
}
|
||||
|
||||
/* parallel calculation of sum of squares */
|
||||
for (size_t m=0; m<M; m++)
|
||||
{
|
||||
size_t sum1 = parallel_reduce( size_t(0), size_t(N), size_t(1024), size_t(0), [&](const range<size_t>& r) -> size_t
|
||||
{
|
||||
size_t s = 0;
|
||||
for (size_t i=r.begin(); i<r.end(); i++)
|
||||
s += i*i;
|
||||
return s;
|
||||
},
|
||||
[](const size_t v0, const size_t v1) {
|
||||
return v0+v1;
|
||||
});
|
||||
passed = sum0 == sum1;
|
||||
}
|
||||
}
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_reduce_regression_test parallel_reduce_regression("parallel_reduce_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -26,8 +39,8 @@ namespace embree
|
||||
const Index threadCount = (Index) TaskScheduler::threadCount();
|
||||
taskCount = min(taskCount,threadCount,maxTasks);
|
||||
|
||||
/* parallel invocation of all tasks */
|
||||
dynamic_large_stack_array(Value,values,taskCount,8192); // consumes at most 8192 bytes on the stack
|
||||
/* parallel invokation of all tasks */
|
||||
dynamic_large_stack_array(Value,values,taskCount,4096); // consumes at most 4096 bytes on the stack
|
||||
parallel_for(taskCount, [&](const Index taskIndex) {
|
||||
const Index k0 = first+(taskIndex+0)*(last-first)/taskCount;
|
||||
const Index k1 = first+(taskIndex+1)*(last-first)/taskCount;
|
||||
@@ -53,22 +66,12 @@ namespace embree
|
||||
return parallel_reduce_internal(taskCount,first,last,minStepSize,identity,func,reduction);
|
||||
|
||||
#elif defined(TASKING_TBB)
|
||||
#if TBB_INTERFACE_VERSION >= 12002
|
||||
tbb::task_group_context context;
|
||||
const Value v = tbb::parallel_reduce(tbb::blocked_range<Index>(first,last,minStepSize),identity,
|
||||
[&](const tbb::blocked_range<Index>& r, const Value& start) { return reduction(start,func(range<Index>(r.begin(),r.end()))); },
|
||||
reduction,context);
|
||||
if (context.is_group_execution_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
return v;
|
||||
#else
|
||||
const Value v = tbb::parallel_reduce(tbb::blocked_range<Index>(first,last,minStepSize),identity,
|
||||
[&](const tbb::blocked_range<Index>& r, const Value& start) { return reduction(start,func(range<Index>(r.begin(),r.end()))); },
|
||||
reduction);
|
||||
if (tbb::task::self().is_cancelled())
|
||||
throw std::runtime_error("task cancelled");
|
||||
return v;
|
||||
#endif
|
||||
#else // TASKING_PPL
|
||||
struct AlignedValue
|
||||
{
|
||||
|
||||
@@ -0,0 +1,56 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_set.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
struct parallel_set_regression_test : public RegressionTest
|
||||
{
|
||||
parallel_set_regression_test(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
|
||||
/* create vector with random numbers */
|
||||
const size_t N = 10000;
|
||||
std::vector<uint32_t> unsorted(N);
|
||||
for (size_t i=0; i<N; i++) unsorted[i] = 2*rand();
|
||||
|
||||
/* created set from numbers */
|
||||
parallel_set<uint32_t> sorted;
|
||||
sorted.init(unsorted);
|
||||
|
||||
/* check that all elements are in the set */
|
||||
for (size_t i=0; i<N; i++) {
|
||||
passed &= sorted.lookup(unsorted[i]);
|
||||
}
|
||||
|
||||
/* check that these elements are not in the set */
|
||||
for (size_t i=0; i<N; i++) {
|
||||
passed &= !sorted.lookup(unsorted[i]+1);
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
parallel_set_regression_test parallel_set_regression("parallel_set_regression_test");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -0,0 +1,63 @@
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "parallel_sort.h"
|
||||
#include "../sys/regression.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
template<typename Key>
|
||||
struct RadixSortRegressionTest : public RegressionTest
|
||||
{
|
||||
RadixSortRegressionTest(const char* name) : RegressionTest(name) {
|
||||
registerRegressionTest(this);
|
||||
}
|
||||
|
||||
bool run ()
|
||||
{
|
||||
bool passed = true;
|
||||
const size_t M = 10;
|
||||
|
||||
for (size_t N=10; N<1000000; N=size_t(2.1*N))
|
||||
{
|
||||
std::vector<Key> src(N); memset(src.data(),0,N*sizeof(Key));
|
||||
std::vector<Key> tmp(N); memset(tmp.data(),0,N*sizeof(Key));
|
||||
for (size_t i=0; i<N; i++) src[i] = uint64_t(rand())*uint64_t(rand());
|
||||
|
||||
/* calculate checksum */
|
||||
Key sum0 = 0; for (size_t i=0; i<N; i++) sum0 += src[i];
|
||||
|
||||
/* sort numbers */
|
||||
for (size_t i=0; i<M; i++) {
|
||||
radix_sort<Key>(src.data(),tmp.data(),N);
|
||||
}
|
||||
|
||||
/* calculate checksum */
|
||||
Key sum1 = 0; for (size_t i=0; i<N; i++) sum1 += src[i];
|
||||
if (sum0 != sum1) passed = false;
|
||||
|
||||
/* check if numbers are sorted */
|
||||
for (size_t i=1; i<N; i++)
|
||||
passed &= src[i-1] <= src[i];
|
||||
}
|
||||
|
||||
return passed;
|
||||
}
|
||||
};
|
||||
|
||||
RadixSortRegressionTest<uint32_t> test_u32("RadixSortRegressionTestU32");
|
||||
RadixSortRegressionTest<uint64_t> test_u64("RadixSortRegressionTestU64");
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -320,7 +333,7 @@ namespace embree
|
||||
#pragma nounroll
|
||||
#endif
|
||||
for (size_t i=startID; i<endID; i++) {
|
||||
#if defined(__64BIT__)
|
||||
#if defined(__X86_64__)
|
||||
const size_t index = ((size_t)(Key)src[i] >> (size_t)shift) & (size_t)mask;
|
||||
#else
|
||||
const Key index = ((Key)src[i] >> shift) & mask;
|
||||
@@ -382,7 +395,7 @@ namespace embree
|
||||
#endif
|
||||
for (size_t i=startID; i<endID; i++) {
|
||||
const Ty elt = src[i];
|
||||
#if defined(__64BIT__)
|
||||
#if defined(__X86_64__)
|
||||
const size_t index = ((size_t)(Key)src[i] >> (size_t)shift) & (size_t)mask;
|
||||
#else
|
||||
const size_t index = ((Key)src[i] >> shift) & mask;
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
IF (NOT OPENIMAGEIO_ROOT)
|
||||
SET(OPENIMAGEIO_ROOT $ENV{OPENIMAGEIO_ROOT})
|
||||
@@ -36,7 +49,7 @@ FIND_PATH(OPENIMAGEIO_ROOT include/OpenImageIO/imageio.h
|
||||
DOC "Root of OpenImageIO installation"
|
||||
HINTS ${OPENIMAGEIO_ROOT}
|
||||
PATHS
|
||||
"${PROJECT_SOURCE_DIR}/oiio"
|
||||
${PROJECT_SOURCE_DIR}/oiio
|
||||
/usr/local
|
||||
/usr
|
||||
/
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
FIND_PATH( PNG_INCLUDE_DIR NAMES png.h )
|
||||
FIND_LIBRARY( PNG_LIBRARIES NAMES png )
|
||||
|
||||
+141
-464
@@ -1,482 +1,159 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
#===============================================================================
|
||||
# This script will attempt to find TBB and set up a TBB target.
|
||||
#
|
||||
# The user may specify a version and lists of required and optional components:
|
||||
#
|
||||
# find_package(TBB 2017.0 EXACT REQUIRED
|
||||
# tbb tbbmalloc
|
||||
# OPTIONAL_COMPONENTS tbbmalloc_proxy
|
||||
# QUIET)
|
||||
#
|
||||
# If this target exists already, the script will attempt to re-use it, but fail
|
||||
# if version or components do not match the user-specified requirements.
|
||||
#
|
||||
# If all the required component targets (e.g. TBB::tbb) exist, the script will
|
||||
# attempt to create a target TBB and link existing component targets to it.
|
||||
# It will fail if the component target version does not match the user-specified
|
||||
# requirements.
|
||||
#
|
||||
# The user may specify the following variables to help the search process:
|
||||
# - TBB_ROOT
|
||||
# - TBB_INCLUDE_DIR
|
||||
#
|
||||
# After the script has run successfully, there is a target TBB, as well as
|
||||
# component targets TBB::<COMPONENT>, e.g. TBB::tbbmalloc.
|
||||
#
|
||||
# The targets will attempt to link to release versions of TBB in release mode,
|
||||
# and debug versions in debug mode.
|
||||
#
|
||||
# In addition to the targets, the script defines:
|
||||
#
|
||||
# TBB_FOUND
|
||||
# TBB_INCLUDE_DIRS
|
||||
#
|
||||
#===============================================================================
|
||||
IF (NOT TBB_ROOT)
|
||||
SET(TBB_ROOT $ENV{TBB_ROOT})
|
||||
ENDIF()
|
||||
IF (NOT TBB_ROOT)
|
||||
SET(TBB_ROOT $ENV{TBBROOT})
|
||||
ENDIF()
|
||||
|
||||
# We use INTERFACE libraries, which are only supported in 3.x
|
||||
cmake_minimum_required(VERSION 3.1)
|
||||
IF (WIN32)
|
||||
# workaround for parentheses in variable name / CMP0053
|
||||
SET(PROGRAMFILESx86 "PROGRAMFILES(x86)")
|
||||
SET(PROGRAMFILES32 "$ENV{${PROGRAMFILESx86}}")
|
||||
IF (NOT PROGRAMFILES32)
|
||||
SET(PROGRAMFILES32 "$ENV{PROGRAMFILES}")
|
||||
ENDIF()
|
||||
IF (NOT PROGRAMFILES32)
|
||||
SET(PROGRAMFILES32 "C:/Program Files (x86)")
|
||||
ENDIF()
|
||||
FIND_PATH(EMBREE_TBB_ROOT include/tbb/tbb.h
|
||||
DOC "Root of TBB installation"
|
||||
PATHS ${PROJECT_SOURCE_DIR}/tbb
|
||||
NO_DEFAULT_PATH
|
||||
)
|
||||
FIND_PATH(EMBREE_TBB_ROOT include/tbb/tbb.h
|
||||
HINTS ${TBB_ROOT}
|
||||
PATHS
|
||||
${PROJECT_SOURCE_DIR}/../tbb
|
||||
"${PROGRAMFILES32}/IntelSWTools/compilers_and_libraries/windows/tbb"
|
||||
"${PROGRAMFILES32}/Intel/Composer XE/tbb"
|
||||
"${PROGRAMFILES32}/Intel/compilers_and_libraries/windows/tbb"
|
||||
)
|
||||
|
||||
# These two are used to automatically find the root and include directories.
|
||||
set(_TBB_INCLUDE_SUBDIR "include")
|
||||
set(_TBB_HEADER "tbb/tbb.h")
|
||||
IF (CMAKE_SIZEOF_VOID_P EQUAL 8)
|
||||
SET(TBB_ARCH intel64)
|
||||
ELSE()
|
||||
SET(TBB_ARCH ia32)
|
||||
ENDIF()
|
||||
|
||||
# Initialize cache variable; but use existing non-cache variable as the default,
|
||||
# and fall back to the environment variable.
|
||||
if (NOT TBB_ROOT)
|
||||
set(TBB_ROOT "$ENV{TBB_ROOT}")
|
||||
endif()
|
||||
# precompiled TBB packages currently only support VC12 and VC14
|
||||
IF (MSVC12)
|
||||
SET(TBB_VCVER vc12)
|
||||
ELSEIF(MSVC14)
|
||||
SET(TBB_VCVER vc14)
|
||||
ENDIF()
|
||||
|
||||
set(TBB_ROOT "${TBB_ROOT}" CACHE PATH "The root path of TBB.")
|
||||
SET(TBB_LIBDIR ${EMBREE_TBB_ROOT}/lib/${TBB_ARCH}/${TBB_VCVER})
|
||||
SET(TBB_BINDIR ${EMBREE_TBB_ROOT}/bin/${TBB_ARCH}/${TBB_VCVER})
|
||||
|
||||
#===============================================================================
|
||||
# Error messages that respect the user's wishes about peace and quiet.
|
||||
#===============================================================================
|
||||
IF (EMBREE_TBB_ROOT STREQUAL "")
|
||||
FIND_PATH(TBB_INCLUDE_DIR tbb/task_scheduler_init.h)
|
||||
FIND_LIBRARY(TBB_LIBRARY tbb)
|
||||
FIND_LIBRARY(TBB_LIBRARY_MALLOC tbbmalloc)
|
||||
ELSE()
|
||||
SET(TBB_INCLUDE_DIR TBB_INCLUDE_DIR-NOTFOUND)
|
||||
SET(TBB_LIBRARY TBB_LIBRARY-NOTFOUND)
|
||||
SET(TBB_LIBRARY_MALLOC TBB_LIBRARY_MALLOC-NOTFOUND)
|
||||
FIND_PATH(TBB_INCLUDE_DIR tbb/task_scheduler_init.h PATHS ${EMBREE_TBB_ROOT}/include NO_DEFAULT_PATH)
|
||||
FIND_LIBRARY(TBB_LIBRARY tbb PATHS ${TBB_LIBDIR} ${EMBREE_TBB_ROOT}/lib NO_DEFAULT_PATH)
|
||||
FIND_LIBRARY(TBB_LIBRARY_MALLOC tbbmalloc PATHS ${TBB_LIBDIR} ${EMBREE_TBB_ROOT}/lib NO_DEFAULT_PATH)
|
||||
ENDIF()
|
||||
|
||||
function(rk_tbb_status)
|
||||
if (NOT TBB_FIND_QUIETLY)
|
||||
message(STATUS "${ARGV}")
|
||||
endif()
|
||||
endfunction()
|
||||
ELSE ()
|
||||
|
||||
function(rk_tbb_warning)
|
||||
if (NOT TBB_FIND_QUIETLY)
|
||||
message(WARNING "${ARGV}")
|
||||
endif()
|
||||
endfunction()
|
||||
FIND_PATH(EMBREE_TBB_ROOT include/tbb/tbb.h
|
||||
DOC "Root of TBB installation"
|
||||
PATHS ${PROJECT_SOURCE_DIR}/tbb
|
||||
NO_DEFAULT_PATH
|
||||
)
|
||||
FIND_PATH(EMBREE_TBB_ROOT include/tbb/tbb.h
|
||||
DOC "Root of TBB installation"
|
||||
HINTS ${TBB_ROOT}
|
||||
PATHS
|
||||
${PROJECT_SOURCE_DIR}/tbb
|
||||
/opt/intel/composerxe/tbb
|
||||
/opt/intel/compilers_and_libraries/tbb
|
||||
/opt/intel/tbb
|
||||
)
|
||||
|
||||
macro(rk_tbb_error)
|
||||
if (TBB_FIND_REQUIRED)
|
||||
message(FATAL_ERROR "${ARGV}")
|
||||
else()
|
||||
rk_tbb_warning("${ARGV}")
|
||||
endif()
|
||||
return()
|
||||
endmacro()
|
||||
|
||||
#===============================================================================
|
||||
# Extract a list of required and optional components.
|
||||
#===============================================================================
|
||||
|
||||
macro(rk_tbb_list_components)
|
||||
# cmake provides the TBB_FIND_COMPONENTS and
|
||||
# TBB_FIND_REQUIRED_<C> variables based on the invocation
|
||||
# of find_package.
|
||||
if (TBB_FIND_COMPONENTS STREQUAL "")
|
||||
set(_REQUIRED_COMPONENTS "tbb")
|
||||
set(_OPTIONAL_COMPONENTS "tbbmalloc"
|
||||
"tbbmalloc_proxy"
|
||||
"tbbbind"
|
||||
"tbbpreview")
|
||||
else()
|
||||
set(_REQUIRED_COMPONENTS "")
|
||||
set(_OPTIONAL_COMPONENTS "")
|
||||
foreach (C IN LISTS TBB_FIND_COMPONENTS)
|
||||
if (${TBB_FIND_REQUIRED_${C}})
|
||||
list(APPEND _REQUIRED_COMPONENTS ${C})
|
||||
else()
|
||||
list(APPEND _OPTIONAL_COMPONENTS ${C})
|
||||
endif()
|
||||
endforeach()
|
||||
endif()
|
||||
|
||||
rk_tbb_status("Looking for TBB components ${_REQUIRED_COMPONENTS}"
|
||||
" (${_OPTIONAL_COMPONENTS})")
|
||||
endmacro()
|
||||
|
||||
#===============================================================================
|
||||
# List components that are available, and check if any REQUIRED components
|
||||
# are missing.
|
||||
#===============================================================================
|
||||
|
||||
macro(rk_tbb_check_components)
|
||||
set(_TBB_MISSING_COMPONENTS "")
|
||||
set(_TBB_AVAILABLE_COMPONENTS "")
|
||||
|
||||
foreach (C IN LISTS _REQUIRED_COMPONENTS)
|
||||
if (TARGET TBB::${C})
|
||||
list(APPEND _TBB_AVAILABLE_COMPONENTS ${C})
|
||||
IF (EMBREE_TBB_ROOT STREQUAL "")
|
||||
FIND_PATH(TBB_INCLUDE_DIR tbb/task_scheduler_init.h)
|
||||
FIND_LIBRARY(TBB_LIBRARY tbb)
|
||||
FIND_LIBRARY(TBB_LIBRARY_MALLOC tbbmalloc)
|
||||
|
||||
ELSEIF (EXISTS ${EMBREE_TBB_ROOT}/cmake/TBBBuild.cmake AND EXISTS ${EMBREE_TBB_ROOT}/src/tbb/tbb_version.h)
|
||||
OPTION(EMBREE_TBB_STATIC_LIB "Build TBB as a static library (building TBB as a static library is NOT recommended)")
|
||||
if (EMBREE_TBB_STATIC_LIB)
|
||||
include(${EMBREE_TBB_ROOT}/cmake/TBBBuild.cmake)
|
||||
tbb_build(TBB_ROOT ${EMBREE_TBB_ROOT} CONFIG_DIR TBB_DIR MAKE_ARGS extra_inc=big_iron.inc)
|
||||
SET(TBB_INCLUDE_DIR ${EMBREE_TBB_ROOT}/include)
|
||||
SET(TBB_LIBRARY ${PROJECT_BINARY_DIR}/tbb_cmake_build/tbb_cmake_build_subdir_release/libtbb.a)
|
||||
SET(TBB_LIBRARY_MALLOC ${PROJECT_BINARY_DIR}/tbb_cmake_build/tbb_cmake_build_subdir_release/libtbbmalloc.a)
|
||||
else()
|
||||
list(APPEND _TBB_MISSING_COMPONENTS ${C})
|
||||
include(${EMBREE_TBB_ROOT}/cmake/TBBBuild.cmake)
|
||||
tbb_build(TBB_ROOT ${EMBREE_TBB_ROOT} CONFIG_DIR TBB_DIR)
|
||||
SET(TBB_INCLUDE_DIR ${EMBREE_TBB_ROOT}/include)
|
||||
SET(TBB_LIBRARY ${PROJECT_BINARY_DIR}/tbb_cmake_build/tbb_cmake_build_subdir_release/libtbb.so.2)
|
||||
SET(TBB_LIBRARY_MALLOC ${PROJECT_BINARY_DIR}/tbb_cmake_build/tbb_cmake_build_subdir_release/libtbbmalloc.so.2)
|
||||
endif()
|
||||
endforeach()
|
||||
|
||||
ELSE()
|
||||
SET(TBB_INCLUDE_DIR TBB_INCLUDE_DIR-NOTFOUND)
|
||||
SET(TBB_LIBRARY TBB_LIBRARY-NOTFOUND)
|
||||
SET(TBB_LIBRARY_MALLOC TBB_LIBRARY_MALLOC-NOTFOUND)
|
||||
IF (APPLE)
|
||||
FIND_PATH(TBB_INCLUDE_DIR tbb/task_scheduler_init.h PATHS ${EMBREE_TBB_ROOT}/include NO_DEFAULT_PATH)
|
||||
FIND_LIBRARY(TBB_LIBRARY tbb PATHS ${EMBREE_TBB_ROOT}/lib NO_DEFAULT_PATH)
|
||||
FIND_LIBRARY(TBB_LIBRARY_MALLOC tbbmalloc PATHS ${EMBREE_TBB_ROOT}/lib NO_DEFAULT_PATH)
|
||||
ELSE()
|
||||
FIND_PATH(TBB_INCLUDE_DIR tbb/task_scheduler_init.h PATHS ${EMBREE_TBB_ROOT}/include NO_DEFAULT_PATH)
|
||||
SET(TBB_HINTS HINTS ${EMBREE_TBB_ROOT}/lib/intel64/gcc4.4 ${EMBREE_TBB_ROOT}/lib ${EMBREE_TBB_ROOT}/lib64 PATHS /usr/libx86_64-linux-gnu/)
|
||||
FIND_LIBRARY(TBB_LIBRARY tbb ${TBB_HINTS})
|
||||
FIND_LIBRARY(TBB_LIBRARY_MALLOC tbbmalloc ${TBB_HINTS})
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
foreach (C IN LISTS _OPTIONAL_COMPONENTS)
|
||||
if (TARGET TBB::${C})
|
||||
list(APPEND _TBB_AVAILABLE_COMPONENTS ${C})
|
||||
endif()
|
||||
endforeach()
|
||||
endmacro()
|
||||
ENDIF()
|
||||
|
||||
#===============================================================================
|
||||
# Check the version of the TBB root we found.
|
||||
#===============================================================================
|
||||
INCLUDE(FindPackageHandleStandardArgs)
|
||||
FIND_PACKAGE_HANDLE_STANDARD_ARGS(TBB DEFAULT_MSG TBB_INCLUDE_DIR TBB_LIBRARY TBB_LIBRARY_MALLOC)
|
||||
|
||||
macro(rk_tbb_check_version)
|
||||
# Extract the version we found in our root.
|
||||
if(EXISTS "${TBB_INCLUDE_DIR}/oneapi/tbb/version.h")
|
||||
set(_TBB_VERSION_HEADER "oneapi/tbb/version.h")
|
||||
elseif(EXISTS "${TBB_INCLUDE_DIR}/tbb/tbb_stddef.h")
|
||||
set(_TBB_VERSION_HEADER "tbb/tbb_stddef.h")
|
||||
elseif(EXISTS "${TBB_INCLUDE_DIR}/tbb/version.h")
|
||||
set(_TBB_VERSION_HEADER "tbb/version.h")
|
||||
else()
|
||||
rk_tbb_error("Missing TBB version information. Could not find"
|
||||
"tbb/tbb_stddef.h or tbb/version.h in ${TBB_INCLUDE_DIR}")
|
||||
endif()
|
||||
file(READ "${TBB_INCLUDE_DIR}/${_TBB_VERSION_HEADER}" VERSION_HEADER_CONTENT)
|
||||
string(REGEX MATCH "#define TBB_VERSION_MAJOR ([0-9]+)" DUMMY "${VERSION_HEADER_CONTENT}")
|
||||
set(TBB_VERSION_MAJOR ${CMAKE_MATCH_1})
|
||||
string(REGEX MATCH "#define TBB_VERSION_MINOR ([0-9]+)" DUMMY "${VERSION_HEADER_CONTENT}")
|
||||
set(TBB_VERSION_MINOR ${CMAKE_MATCH_1})
|
||||
set(TBB_VERSION "${TBB_VERSION_MAJOR}.${TBB_VERSION_MINOR}")
|
||||
set(TBB_VERSION_STRING "${TBB_VERSION}")
|
||||
IF (TBB_FOUND)
|
||||
SET(TBB_INCLUDE_DIRS ${TBB_INCLUDE_DIR})
|
||||
SET(TBB_LIBRARIES ${TBB_LIBRARY} ${TBB_LIBRARY_MALLOC})
|
||||
ENDIF()
|
||||
|
||||
# If the user provided information about required versions, check them!
|
||||
if (TBB_FIND_VERSION)
|
||||
if (${TBB_FIND_VERSION_EXACT} AND NOT
|
||||
TBB_VERSION VERSION_EQUAL ${TBB_FIND_VERSION})
|
||||
rk_tbb_error("Requested exact TBB version ${TBB_FIND_VERSION},"
|
||||
" but found ${TBB_VERSION}")
|
||||
elseif(TBB_VERSION VERSION_LESS ${TBB_FIND_VERSION})
|
||||
rk_tbb_error("Requested minimum TBB version ${TBB_FIND_VERSION},"
|
||||
" but found ${TBB_VERSION}")
|
||||
endif()
|
||||
endif()
|
||||
MARK_AS_ADVANCED(TBB_INCLUDE_DIR)
|
||||
MARK_AS_ADVANCED(TBB_LIBRARY)
|
||||
MARK_AS_ADVANCED(TBB_LIBRARY_MALLOC)
|
||||
|
||||
rk_tbb_status("Found TBB version ${TBB_VERSION} at ${TBB_ROOT}")
|
||||
endmacro()
|
||||
##############################################################
|
||||
# Install TBB
|
||||
##############################################################
|
||||
|
||||
#===============================================================================
|
||||
# Reuse existing targets.
|
||||
# NOTE: This must be a macro, as we rely on return() to exit this script.
|
||||
#===============================================================================
|
||||
IF (EMBREE_INSTALL_DEPENDENCIES)
|
||||
IF (WIN32)
|
||||
INSTALL(PROGRAMS ${TBB_BINDIR}/tbb.dll ${TBB_BINDIR}/tbbmalloc.dll DESTINATION ${CMAKE_INSTALL_BINDIR} COMPONENT examples)
|
||||
INSTALL(PROGRAMS ${TBB_LIBDIR}/tbb.lib ${TBB_LIBDIR}/tbbmalloc.lib DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT lib)
|
||||
ELSEIF (APPLE)
|
||||
INSTALL(PROGRAMS ${EMBREE_TBB_ROOT}/lib/libtbb.dylib ${EMBREE_TBB_ROOT}/lib/libtbbmalloc.dylib DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT lib)
|
||||
ELSE()
|
||||
INSTALL(PROGRAMS ${EMBREE_TBB_ROOT}/lib/intel64/gcc4.4/libtbb.so.2 ${EMBREE_TBB_ROOT}/lib/intel64/gcc4.4/libtbbmalloc.so.2 DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT lib)
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
macro(rk_tbb_reuse_existing_target_components)
|
||||
rk_tbb_check_components()
|
||||
|
||||
if (_TBB_MISSING_COMPONENTS STREQUAL "")
|
||||
rk_tbb_status("Found existing TBB component targets: ${_TBB_AVAILABLE_COMPONENTS}")
|
||||
|
||||
# Get TBB_INCLUDE_DIR if not already set to check for the version of the
|
||||
# existing component targets (making the assumption that they all have
|
||||
# the same version)
|
||||
if (NOT TBB_INCLUDE_DIR)
|
||||
list(GET _TBB_AVAILABLE_COMPONENTS 0 first_target)
|
||||
get_target_property(TBB_INCLUDE_DIR TBB::${first_target} INTERFACE_INCLUDE_DIRECTORIES)
|
||||
foreach(TGT IN LISTS _TBB_AVAILABLE_COMPONENTS)
|
||||
get_target_property(_TGT_INCLUDE_DIR TBB::${TGT} INTERFACE_INCLUDE_DIRECTORIES)
|
||||
if (NOT _TGT_INCLUDE_DIR STREQUAL "${TBB_INCLUDE_DIR}")
|
||||
rk_tbb_error("Existing TBB component targets have inconsistent include directories.")
|
||||
endif()
|
||||
endforeach()
|
||||
endif()
|
||||
|
||||
find_path(TBB_INCLUDE_DIR
|
||||
NAMES "${_TBB_HEADER}"
|
||||
PATHS "${TBB_INCLUDE_DIRS}")
|
||||
|
||||
# Extract TBB_ROOT from the include path so that rk_tbb_check_version
|
||||
# prints the correct tbb location
|
||||
string(REPLACE "/${_TBB_INCLUDE_SUBDIR}" "" TBB_ROOT "${TBB_INCLUDE_DIR}")
|
||||
rk_tbb_check_version()
|
||||
|
||||
# Add target TBB and link all available components
|
||||
if (NOT TARGET TBB)
|
||||
add_library(TBB INTERFACE)
|
||||
foreach(C IN LISTS _TBB_AVAILABLE_COMPONENTS)
|
||||
target_link_libraries(TBB INTERFACE TBB::${C})
|
||||
endforeach()
|
||||
endif()
|
||||
set(TBB_FOUND TRUE)
|
||||
set(TBB_INCLUDE_DIRS "${TBB_INCLUDE_DIR}")
|
||||
return()
|
||||
elseif ((TARGET TBB) OR (NOT _TBB_AVAILABLE_COMPONENTS STREQUAL ""))
|
||||
rk_tbb_error("Ignoring existing TBB targets because required components are missing: ${_TBB_MISSING_COMPONENTS}")
|
||||
endif()
|
||||
endmacro()
|
||||
|
||||
|
||||
#===============================================================================
|
||||
# Find the root directory if a manual override is not specified.
|
||||
# Sets TBB_ROOT in the parent scope, but does not check for failure.
|
||||
#===============================================================================
|
||||
|
||||
function(rk_tbb_find_root)
|
||||
if (NOT TBB_ROOT OR TBB_ROOT STREQUAL "")
|
||||
set(TBB_HINTS "")
|
||||
set(TBB_PATHS "")
|
||||
|
||||
if (WIN32)
|
||||
# workaround for parentheses in variable name / CMP0053
|
||||
set(PROGRAMFILESx86 "PROGRAMFILES(x86)")
|
||||
set(PROGRAMFILES32 "$ENV{${PROGRAMFILESx86}}")
|
||||
if(NOT PROGRAMFILES32)
|
||||
set(PROGRAMFILES32 "$ENV{PROGRAMFILES}")
|
||||
endif()
|
||||
if(NOT PROGRAMFILES32)
|
||||
set(PROGRAMFILES32 "C:/Program Files (x86)")
|
||||
endif()
|
||||
set(TBB_PATHS
|
||||
"${PROJECT_SOURCE_DIR}/../tbb"
|
||||
"${PROGRAMFILES32}/IntelSWTools/compilers_and_libraries/windows/tbb"
|
||||
"${PROGRAMFILES32}/Intel/Composer XE/tbb"
|
||||
"${PROGRAMFILES32}/Intel/compilers_and_libraries/windows/tbb")
|
||||
else()
|
||||
set(TBB_HINTS "/usr/local")
|
||||
set(TBB_PATHS
|
||||
"${PROJECT_SOURCE_DIR}/tbb"
|
||||
"/opt/intel/composerxe/tbb"
|
||||
"/opt/intel/compilers_and_libraries/tbb"
|
||||
"/opt/intel/compilers_and_libraries/linux/tbb"
|
||||
"/opt/intel/tbb")
|
||||
endif()
|
||||
|
||||
set(TBB_ROOT "TBB_ROOT-NOTFOUND")
|
||||
find_path(TBB_ROOT
|
||||
NAMES "${_TBB_INCLUDE_SUBDIR}/${_TBB_HEADER}"
|
||||
HINTS ${TBB_HINTS}
|
||||
PATHS ${TBB_PATHS}
|
||||
NO_PACKAGE_ROOT_PATH)
|
||||
endif()
|
||||
endfunction()
|
||||
|
||||
#===============================================================================
|
||||
# Find the include directory if a manual override is not specified.
|
||||
# Assumes TBB_ROOT to be set.
|
||||
#===============================================================================
|
||||
|
||||
function(rk_tbb_find_include_directory)
|
||||
find_path(TBB_INCLUDE_DIR
|
||||
NAMES "${_TBB_HEADER}"
|
||||
HINTS "${TBB_ROOT}/${_TBB_INCLUDE_SUBDIR}"
|
||||
NO_PACKAGE_ROOT_PATH)
|
||||
endfunction()
|
||||
|
||||
#===============================================================================
|
||||
# Find a specific library and create a target for it.
|
||||
#===============================================================================
|
||||
|
||||
function(rk_tbb_find_library COMPONENT_NAME BUILD_CONFIG)
|
||||
set(LIB_VAR "${COMPONENT_NAME}_LIBRARY_${BUILD_CONFIG}")
|
||||
set(BIN_DIR_VAR "${COMPONENT_NAME}_BIN_DIR_${BUILD_CONFIG}")
|
||||
set(DLL_VAR "${COMPONENT_NAME}_DLL_${BUILD_CONFIG}")
|
||||
if (BUILD_CONFIG STREQUAL "DEBUG")
|
||||
set(LIB_NAME "${COMPONENT_NAME}_debug")
|
||||
else()
|
||||
set(LIB_NAME "${COMPONENT_NAME}")
|
||||
endif()
|
||||
|
||||
unset(LIB_PATHS)
|
||||
|
||||
if (WIN32)
|
||||
if(CMAKE_SIZEOF_VOID_P EQUAL 8)
|
||||
set(TBB_ARCH intel64)
|
||||
else()
|
||||
set(TBB_ARCH ia32)
|
||||
endif()
|
||||
|
||||
if(MSVC10)
|
||||
set(TBB_VCVER vc10)
|
||||
elseif(MSVC11)
|
||||
set(TBB_VCVER vc11)
|
||||
elseif(MSVC12)
|
||||
set(TBB_VCVER vc12)
|
||||
else()
|
||||
set(TBB_VCVER vc14)
|
||||
endif()
|
||||
|
||||
set(LIB_PATHS
|
||||
${TBB_ROOT}/lib/${TBB_ARCH}/${TBB_VCVER}
|
||||
${TBB_ROOT}/lib
|
||||
)
|
||||
|
||||
# On window, also search the DLL so that the client may install it.
|
||||
set(DLL_NAME "${LIB_NAME}.dll")
|
||||
|
||||
# lib name with version suffix to handle oneTBB tbb12.dll
|
||||
set(LIB_NAME_VERSION "")
|
||||
if (${COMPONENT_NAME} STREQUAL "tbb")
|
||||
if (BUILD_CONFIG STREQUAL "DEBUG")
|
||||
set(LIB_NAME_VERSION "tbb12_debug")
|
||||
else()
|
||||
set(LIB_NAME_VERSION "tbb12")
|
||||
endif()
|
||||
endif()
|
||||
set(DLL_NAME_VERSION "${LIB_NAME_VERSION}.dll")
|
||||
|
||||
find_file(BIN_FILE
|
||||
NAMES ${DLL_NAME} ${DLL_NAME_VERSION}
|
||||
PATHS
|
||||
"${TBB_ROOT}/bin/${TBB_ARCH}/${TBB_VCVER}"
|
||||
"${TBB_ROOT}/bin"
|
||||
"${TBB_ROOT}/redist/${TBB_ARCH}/${TBB_VCVER}"
|
||||
"${TBB_ROOT}/../redist/${TBB_ARCH}/tbb/${TBB_VCVER}"
|
||||
"${TBB_ROOT}/../redist/${TBB_ARCH}_win/tbb/${TBB_VCVER}"
|
||||
NO_DEFAULT_PATH)
|
||||
get_filename_component(${BIN_DIR_VAR} ${BIN_FILE} DIRECTORY)
|
||||
set(${DLL_VAR} "${BIN_FILE}" CACHE PATH "${COMPONENT_NAME} ${BUILD_CONFIG} dll path")
|
||||
elseif(APPLE)
|
||||
set(LIB_PATHS ${TBB_ROOT}/lib)
|
||||
else()
|
||||
file(GLOB LIB_PATHS PATHS ${TBB_ROOT}/lib/intel64/gcc*)
|
||||
list(REVERSE LIB_PATHS)
|
||||
list(APPEND LIB_PATHS
|
||||
${TBB_ROOT}/lib
|
||||
${TBB_ROOT}/lib/x86_64-linux-gnu
|
||||
${TBB_ROOT}/lib64
|
||||
${TBB_ROOT}/libx86_64-linux-gnu)
|
||||
endif()
|
||||
|
||||
# We prefer finding the versioned file on Unix so that the library path
|
||||
# variable will not point to a symlink. This makes installing TBB as a
|
||||
# dependency easier.
|
||||
if (UNIX)
|
||||
set(LIB_NAME lib${LIB_NAME}.so.2 ${LIB_NAME})
|
||||
endif()
|
||||
|
||||
find_library(${LIB_VAR}
|
||||
NAMES ${LIB_NAME}
|
||||
PATHS ${LIB_PATHS}
|
||||
NO_DEFAULT_PATH)
|
||||
|
||||
# Hide this variable if we found something, otherwise display it for
|
||||
# easy override.
|
||||
if(${LIB_VAR})
|
||||
mark_as_advanced(${LIB_VAR})
|
||||
endif()
|
||||
if(${BIN_DIR_VAR})
|
||||
mark_as_advanced(${BIN_DIR_VAR})
|
||||
endif()
|
||||
if(${DLL_VAR})
|
||||
mark_as_advanced(${DLL_VAR})
|
||||
endif()
|
||||
endfunction()
|
||||
|
||||
#===============================================================================
|
||||
# Find the given component.
|
||||
# This macro attempts to find both release and debug versions, and falls back
|
||||
# appropriately if only one can be found.
|
||||
# On success, it creates a target ${TARGET}::${COMPONENT_NAME} and links
|
||||
# it to the overall ${TARGET}.
|
||||
#
|
||||
# For more information on the variables set here, see
|
||||
# https://cmake.org/cmake/help/v3.17/manual/cmake-developer.7.html#a-sample-find-module
|
||||
#===============================================================================
|
||||
|
||||
function(rk_tbb_find_and_link_component COMPONENT_NAME)
|
||||
set(COMPONENT_TARGET "TBB::${COMPONENT_NAME}")
|
||||
|
||||
rk_tbb_find_library("${COMPONENT_NAME}" RELEASE)
|
||||
rk_tbb_find_library("${COMPONENT_NAME}" DEBUG)
|
||||
|
||||
if (${COMPONENT_NAME}_LIBRARY_RELEASE OR ${COMPONENT_NAME}_LIBRARY_DEBUG)
|
||||
# Note: We *must* use SHARED here rather than UNKNOWN as our
|
||||
# IMPORTED_NO_SONAME trick a few lines down does not work with
|
||||
# UNKNOWN.
|
||||
add_library(${COMPONENT_TARGET} SHARED IMPORTED)
|
||||
|
||||
if (${COMPONENT_NAME}_LIBRARY_RELEASE)
|
||||
set_property(TARGET ${COMPONENT_TARGET} APPEND PROPERTY
|
||||
IMPORTED_CONFIGURATIONS RELEASE)
|
||||
if(WIN32)
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_LOCATION_RELEASE "${${COMPONENT_NAME}_DLL_RELEASE}")
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_IMPLIB_RELEASE "${${COMPONENT_NAME}_LIBRARY_RELEASE}")
|
||||
else()
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_LOCATION_RELEASE "${${COMPONENT_NAME}_LIBRARY_RELEASE}")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if (${COMPONENT_NAME}_LIBRARY_DEBUG)
|
||||
set_property(TARGET ${COMPONENT_TARGET} APPEND PROPERTY
|
||||
IMPORTED_CONFIGURATIONS DEBUG)
|
||||
if(WIN32)
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_LOCATION_DEBUG "${${COMPONENT_NAME}_DLL_DEBUG}")
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_IMPLIB_DEBUG "${${COMPONENT_NAME}_LIBRARY_DEBUG}")
|
||||
else()
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_LOCATION_DEBUG "${${COMPONENT_NAME}_LIBRARY_DEBUG}")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
INTERFACE_INCLUDE_DIRECTORIES "${TBB_INCLUDE_DIR}"
|
||||
INTERFACE_COMPILE_DEFINITIONS "__TBB_NO_IMPLICIT_LINKAGE=1"
|
||||
)
|
||||
|
||||
|
||||
if(NOT WIN32)
|
||||
# Note: IMPORTED_NO_SONAME must be set or cmake will attempt
|
||||
# to link to the full path of libtbb.so. Instead, we
|
||||
# rely on the linker to find libtbb.so.2.
|
||||
set_target_properties(${COMPONENT_TARGET} PROPERTIES
|
||||
IMPORTED_NO_SONAME TRUE
|
||||
)
|
||||
endif()
|
||||
|
||||
target_link_libraries(TBB INTERFACE ${COMPONENT_TARGET})
|
||||
endif()
|
||||
endfunction()
|
||||
|
||||
#===============================================================================
|
||||
|
||||
# Note: The order of these is important.
|
||||
# Some of these macros create variables that are used in later calls.
|
||||
rk_tbb_list_components()
|
||||
rk_tbb_reuse_existing_target_components()
|
||||
|
||||
rk_tbb_find_root()
|
||||
if (NOT EXISTS "${TBB_ROOT}")
|
||||
rk_tbb_error("Unable to find root directory ${TBB_ROOT}")
|
||||
endif()
|
||||
mark_as_advanced(TBB_ROOT) # Hide, we found something.
|
||||
|
||||
rk_tbb_find_include_directory()
|
||||
if (NOT EXISTS "${TBB_INCLUDE_DIR}")
|
||||
rk_tbb_error("Unable to find include directory ${TBB_INCLUDE_DIR}")
|
||||
endif()
|
||||
mark_as_advanced(TBB_INCLUDE_DIR) # Hide, we found something.
|
||||
|
||||
rk_tbb_check_version()
|
||||
|
||||
add_library(TBB INTERFACE)
|
||||
|
||||
foreach(C IN LISTS _REQUIRED_COMPONENTS _OPTIONAL_COMPONENTS)
|
||||
rk_tbb_find_and_link_component(${C})
|
||||
endforeach()
|
||||
|
||||
rk_tbb_check_components()
|
||||
if (_TBB_MISSING_COMPONENTS)
|
||||
rk_tbb_error("Cannot find required components: "
|
||||
"${_TBB_MISSING_COMPONENTS}")
|
||||
endif()
|
||||
|
||||
set(TBB_FOUND TRUE)
|
||||
set(TBB_INCLUDE_DIRS "${TBB_INCLUDE_DIR}")
|
||||
|
||||
@@ -1,13 +0,0 @@
|
||||
// Copyright 2009-2020 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#if !defined(__ARM_NEON)
|
||||
#error "No ARM Neon support"
|
||||
#endif
|
||||
|
||||
#include <arm_neon.h>
|
||||
|
||||
int main()
|
||||
{
|
||||
return vaddvq_s32(vdupq_n_s32(1));
|
||||
}
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
IF (WIN32 OR APPLE)
|
||||
return()
|
||||
@@ -12,7 +25,7 @@ foreach (line ${output})
|
||||
if ("${line}" MATCHES "O .bss")
|
||||
if (NOT "${line}" MATCHES "std::__ioinit" AND # this is caused by iostream initialization and is likely also ok
|
||||
NOT "${line}" MATCHES "\\(\\)::" AND # this matches a static inside a function which is fine
|
||||
NOT "${line}" MATCHES "function_local_static_" AND # static variable inside a function (explicitly named)
|
||||
NOT "${line}" MATCHES "function_local_static_" AND # static variable inside a function (explicitely named)
|
||||
NOT "${line}" MATCHES "__\\$U") # ICC generated locks for static variable inside a function
|
||||
message(WARNING "\nProblematic global variable in non-SSE code:\n" ${line})
|
||||
endif()
|
||||
|
||||
@@ -17,7 +17,11 @@
|
||||
#if \
|
||||
defined(__AVX512F__) && defined(__AVX512CD__) && \
|
||||
defined(__AVX512DQ__) && defined(__AVX512BW__) && defined(__AVX512VL__)
|
||||
char const *info_isa = "ISA" ":" "AVX512";
|
||||
char const *info_isa = "ISA" ":" "AVX512SKX";
|
||||
#elif \
|
||||
defined(__AVX512F__) && defined(__AVX512CD__) && \
|
||||
defined(__AVX512ER__) && defined(__AVX512PF__)
|
||||
char const *info_isa = "ISA" ":" "AVX512KNL";
|
||||
#elif defined(__AVX2__)
|
||||
char const *info_isa = "ISA" ":" "AVX2";
|
||||
#elif defined(__AVX__)
|
||||
|
||||
@@ -14,29 +14,19 @@
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
SET(CHECK_ISA_DIR "${CMAKE_CURRENT_LIST_DIR}")
|
||||
SET(CHECK_ISA_DIR ${CMAKE_CURRENT_LIST_DIR})
|
||||
FUNCTION(CHECK_ISA_DEFAULT OUTVAR)
|
||||
|
||||
TRY_COMPILE(COMPILER_SUPPORTS_ARM_NEON "${CMAKE_BINARY_DIR}" "${PROJECT_SOURCE_DIR}/common/cmake/check_arm_neon.cpp")
|
||||
IF (COMPILER_SUPPORTS_ARM_NEON)
|
||||
SET(ISA_DEFAULT "NEON")
|
||||
SET(${OUTVAR} ${ISA_DEFAULT} PARENT_SCOPE)
|
||||
RETURN()
|
||||
ENDIF()
|
||||
|
||||
SET(ISA_DEFAULT_BIN "${CMAKE_BINARY_DIR}/${CMAKE_FILES_DIRECTORY}/check_isa_default.bin")
|
||||
SET(SRC "${CHECK_ISA_DIR}/check_isa.cpp")
|
||||
SET(ISA_DEFAULT_BIN ${CMAKE_BINARY_DIR}${CMAKE_FILES_DIRECTORY}/check_isa_default.bin)
|
||||
SET(SRC ${CHECK_ISA_DIR}/check_isa.cpp)
|
||||
TRY_COMPILE(ISA_DEFAULT_COMPILE
|
||||
"${CMAKE_BINARY_DIR}"
|
||||
"${SRC}"
|
||||
COPY_FILE "${ISA_DEFAULT_BIN}"
|
||||
${CMAKE_BINARY_DIR}
|
||||
${SRC}
|
||||
COPY_FILE ${ISA_DEFAULT_BIN}
|
||||
)
|
||||
IF(NOT ISA_DEFAULT_COMPILE)
|
||||
SET(ISA_DEFAULT "SSE2")
|
||||
SET(${OUTVAR} ${ISA_DEFAULT} PARENT_SCOPE)
|
||||
RETURN()
|
||||
ENDIF()
|
||||
|
||||
FILE(STRINGS ${ISA_DEFAULT_BIN} ISA_DEFAULT REGEX "^ISA:")
|
||||
STRING(REPLACE "ISA:" "" ISA_DEFAULT "${ISA_DEFAULT}")
|
||||
SET(${OUTVAR} ${ISA_DEFAULT} PARENT_SCOPE)
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
IF (WIN32 OR APPLE)
|
||||
return()
|
||||
|
||||
+39
-37
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
IF(NOT ${VAR})
|
||||
@@ -7,19 +20,12 @@ MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
SET(FLAGS_SSE2 "-D__SSE__ -D__SSE2__")
|
||||
SET(FLAGS_SSE42 "-D__SSE4_2__ -D__SSE4_1__")
|
||||
SET(FLAGS_AVX "-D__AVX__ -D__SSE4_2__ -D__SSE4_1__ -D__BMI__ -D__BMI2__ -D__LZCNT__")
|
||||
SET(FLAGS_AVX2 "-D__AVX2__ -D__AVX__ -D__SSE4_2__ -D__SSE4_1__ -D__BMI__ -D__BMI2__ -D__LZCNT__")
|
||||
ELSE ()
|
||||
# for `thread` keyword
|
||||
_SET_IF_EMPTY(FLAGS_SSE2 "-msse -msse2 -mno-sse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-msse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-mavx")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512 "-march=skx")
|
||||
ENDIF ()
|
||||
_SET_IF_EMPTY(FLAGS_SSE2 "-msse2")
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-msse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-mavx")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512KNL "-march=knl")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512SKX "-march=skx")
|
||||
|
||||
IF (WIN32)
|
||||
|
||||
@@ -28,8 +34,6 @@ IF (WIN32)
|
||||
# SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /MP") # compile source files in parallel
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GR") # enable runtime type information (on by default)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} -Xclang -fcxx-exceptions") # enable C++ exceptions in Clang
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /w") # disable all warnings
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /Gy") # package individual functions
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GS") # protects against return address overrides
|
||||
ELSE()
|
||||
@@ -44,7 +48,6 @@ IF (WIN32)
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DDEBUG") # enables assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_DEBUG "${CMAKE_EXE_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_DEBUG "${CMAKE_SHARED_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
@@ -52,10 +55,15 @@ IF (WIN32)
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Gy") # package individual functions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -fno-vectorize") # disable auto vectorization
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Gy") # package individual functions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -fno-vectorize") # disable auto vectorization
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
|
||||
@@ -69,6 +77,7 @@ IF (WIN32)
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} ${SECURE_LINKER_FLAGS}")
|
||||
|
||||
INCLUDE(msvc_post)
|
||||
|
||||
ELSE()
|
||||
|
||||
OPTION(EMBREE_IGNORE_CMAKE_CXX_FLAGS "When enabled Embree ignores default CMAKE_CXX_FLAGS." ON)
|
||||
@@ -80,7 +89,6 @@ ELSE()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall") # enables most warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wformat -Wformat-security") # enables string format vulnerability warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsigned-char") # treat char as signed on all processors, including ARM
|
||||
IF (NOT APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIE") # enables support for more secure position independent execution
|
||||
ENDIF()
|
||||
@@ -91,7 +99,6 @@ ELSE()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fvisibility-inlines-hidden") # makes all inline symbols hidden by default
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-aliasing") # disables strict aliasing rules
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-tree-vectorize") # disable auto vectorizer
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fstack-protector") # protects against return address overrides
|
||||
ENDIF()
|
||||
@@ -105,26 +112,23 @@ ELSE()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsanitize=address -fsanitize-address-use-after-scope -fno-omit-frame-pointer -fno-optimize-sibling-calls")
|
||||
ENDIF()
|
||||
|
||||
IF (EMSCRIPTEN)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fexceptions") # enable exceptions
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -pthread") # enable threads
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msimd128") # enable SIMD intrinsics
|
||||
ENDIF()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "")
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O2") # allow inlining of intrinsics
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
IF (APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mmacosx-version-min=10.7") # makes sure code runs on older MacOSX versions
|
||||
@@ -132,13 +136,11 @@ ELSE()
|
||||
ELSE(APPLE)
|
||||
IF (NOT EMBREE_ADDRESS_SANITIZER) # for address sanitizer this causes link errors
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -Wl,--no-undefined") # issues link error for undefined symbols in shared library
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -pie") # enables position independent execution for executable
|
||||
IF (NOT EMSCRIPTEN)
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDIF(APPLE)
|
||||
|
||||
|
||||
@@ -17,7 +17,8 @@ SET(FLAGS_SSE2 "-target-cpu=x86_64")
|
||||
SET(FLAGS_SSE42 "NOT_SUPPORTED")
|
||||
SET(FLAGS_AVX "-target-cpu=sandybridge")
|
||||
SET(FLAGS_AVX2 "-target-cpu=haswell")
|
||||
SET(FLAGS_AVX512 "-target-cpu=x86-skylake")
|
||||
SET(FLAGS_AVX512KNL "-target-cpu=mic-knl")
|
||||
SET(FLAGS_AVX512SKX "-target-cpu=x86-skylake")
|
||||
|
||||
STRING(TOLOWER "${CMAKE_CXX_COMPILER_ID}" _lower_compiler_id)
|
||||
INCLUDE("${CMAKE_CURRENT_LIST_DIR}/${_lower_compiler_id}.cmake" OPTIONAL)
|
||||
|
||||
@@ -1,4 +1,17 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
file(WRITE ${dst} "#include \"${src}\"\n")
|
||||
|
||||
@@ -1,138 +0,0 @@
|
||||
## Copyright 2009-2022 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
IF(NOT ${VAR})
|
||||
SET(${VAR} "${VALUE}")
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
_SET_IF_EMPTY(FLAGS_SSE2 "-msse2")
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-msse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-mavx")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512 "-march=skx")
|
||||
|
||||
IF (WIN32)
|
||||
|
||||
SET(COMMON_CXX_FLAGS "")
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /EHsc") # catch C++ exceptions only and extern "C" functions never throw a C++ exception
|
||||
# SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /MP") # compile source files in parallel
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GR") # enable runtime type information (on by default)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} -Xclang -fcxx-exceptions") # enable C++ exceptions in Clang
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /w") # disable all warnings
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /Gy") # package individual functions
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GS") # protects against return address overrides
|
||||
ELSE()
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GS-") # do not protect against return address overrides
|
||||
ENDIF()
|
||||
MACRO(DISABLE_STACK_PROTECTOR_FOR_FILE file)
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET_SOURCE_FILES_PROPERTIES(${file} PROPERTIES COMPILE_FLAGS "/GS-")
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DDEBUG") # enables assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_DEBUG "${CMAKE_EXE_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_DEBUG "${CMAKE_SHARED_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oi") # inline intrinsic functions
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
|
||||
SET(SECURE_LINKER_FLAGS "")
|
||||
SET(SECURE_LINKER_FLAGS "${SECURE_LINKER_FLAGS} /NXCompat") # compatible with data execution prevention (on by default)
|
||||
SET(SECURE_LINKER_FLAGS "${SECURE_LINKER_FLAGS} /DynamicBase") # random rebase of executable at load time
|
||||
IF (CMAKE_SIZEOF_VOID_P EQUAL 4)
|
||||
SET(SECURE_LINKER_FLAGS "${SECURE_LINKER_FLAGS} /SafeSEH") # invoke known exception handlers (Win32 only, x64 exception handlers are safe by design)
|
||||
ENDIF()
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${SECURE_LINKER_FLAGS}")
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} ${SECURE_LINKER_FLAGS}")
|
||||
|
||||
INCLUDE(msvc_post)
|
||||
|
||||
# workaround for file encoding problems of kernels/embree.rc found here https://gitlab.kitware.com/cmake/cmake/-/issues/18311
|
||||
set(CMAKE_NINJA_CMCLDEPS_RC OFF)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /fp:precise") # dpcpp has fp-model fast as default
|
||||
|
||||
ELSE()
|
||||
|
||||
OPTION(EMBREE_IGNORE_CMAKE_CXX_FLAGS "When enabled Embree ignores default CMAKE_CXX_FLAGS." ON)
|
||||
OPTION(EMBREE_ADDRESS_SANITIZER "Enabled CLANG address sanitizer." OFF)
|
||||
IF (EMBREE_IGNORE_CMAKE_CXX_FLAGS)
|
||||
SET(CMAKE_CXX_FLAGS "")
|
||||
ENDIF()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall") # enables most warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wformat -Wformat-security") # enables string format vulnerability warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsigned-char") # treat char as signed on all processors, including ARM
|
||||
IF (NOT APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIE") # enables support for more secure position independent execution
|
||||
ENDIF()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11") # enables C++11 features
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fvisibility=hidden") # makes all symbols hidden by default
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fvisibility-inlines-hidden") # makes all inline symbols hidden by default
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-aliasing") # disables strict aliasing rules
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-tree-vectorize") # disable auto vectorizer
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fstack-protector") # protects against return address overrides
|
||||
ENDIF()
|
||||
MACRO(DISABLE_STACK_PROTECTOR_FOR_FILE file)
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET_SOURCE_FILES_PROPERTIES(${file} PROPERTIES COMPILE_FLAGS "-fno-stack-protector")
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
IF (EMBREE_ADDRESS_SANITIZER)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsanitize=address -fsanitize-address-use-after-scope -fno-omit-frame-pointer -fno-optimize-sibling-calls")
|
||||
ENDIF()
|
||||
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -static-intel") # links intel runtime statically
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -static-intel") # links intel runtime statically
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -ffp-model=precise") # dpcpp has fp-model fast as default
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "")
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O3") # enable full optimizations
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3") # enable full optimizations
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -O3") # enable full optimizations
|
||||
|
||||
IF (APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mmacosx-version-min=10.7") # makes sure code runs on older MacOSX versions
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++") # link against libc++ which supports C++11 features
|
||||
ELSE(APPLE)
|
||||
IF (NOT EMBREE_ADDRESS_SANITIZER) # for address sanitizer this causes link errors
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -Wl,--no-undefined") # issues link error for undefined symbols in shared library
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -pie") # enables position independent execution for executable
|
||||
ENDIF()
|
||||
ENDIF(APPLE)
|
||||
|
||||
ENDIF()
|
||||
@@ -1,10 +1,23 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
# use default install config
|
||||
INCLUDE("${CMAKE_CURRENT_LIST_DIR}/embree-config-install.cmake")
|
||||
INCLUDE(${CMAKE_CURRENT_LIST_DIR}/embree-config-install.cmake)
|
||||
|
||||
# and override path variables to match for build directory
|
||||
SET(EMBREE_INCLUDE_DIRS @PROJECT_SOURCE_DIR@/include)
|
||||
SET(EMBREE_LIBRARY @PROJECT_BINARY_DIR@/@EMBREE_LIBRARY_FULLNAME@)
|
||||
SET(EMBREE_LIBRARY @PROJECT_BINARY_DIR@/@EMBREE_LIBRARY_NAME@)
|
||||
SET(EMBREE_LIBRARIES ${EMBREE_LIBRARY})
|
||||
|
||||
@@ -1,16 +1,28 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
SET(PACKAGE_VERSION @EMBREE_VERSION@)
|
||||
|
||||
SET(PACKAGE_VERSION_EXACT 0)
|
||||
SET(PACKAGE_VERSION_COMPATIBLE 0)
|
||||
|
||||
IF (PACKAGE_FIND_VERSION VERSION_EQUAL PACKAGE_VERSION)
|
||||
SET(PACKAGE_VERSION_EXACT 1)
|
||||
SET(PACKAGE_VERSION_COMPATIBLE 1)
|
||||
ENDIF()
|
||||
|
||||
IF (PACKAGE_FIND_VERSION_MAJOR EQUAL @EMBREE_VERSION_MAJOR@ AND PACKAGE_FIND_VERSION VERSION_LESS PACKAGE_VERSION)
|
||||
SET(PACKAGE_VERSION_COMPATIBLE 1)
|
||||
IF (${PACKAGE_FIND_VERSION_MAJOR} EQUAL @EMBREE_VERSION_MAJOR@)
|
||||
IF (${PACKAGE_FIND_VERSION} VERSION_LESS @EMBREE_VERSION@)
|
||||
SET(PACKAGE_VERSION_COMPATIBLE 1)
|
||||
ENDIF()
|
||||
IF (${PACKAGE_FIND_VERSION} VERSION_EQUAL @EMBREE_VERSION@)
|
||||
SET(PACKAGE_VERSION_EXACT 1)
|
||||
ENDIF()
|
||||
ELSE()
|
||||
SET(PACKAGE_VERSION_UNSUITABLE 1)
|
||||
ENDIF()
|
||||
|
||||
@@ -1,13 +1,28 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
SET(EMBREE_ROOT_DIR "${CMAKE_CURRENT_LIST_DIR}/@EMBREE_RELATIVE_ROOT_DIR@")
|
||||
SET(EMBREE_ROOT_DIR ${CMAKE_CURRENT_LIST_DIR}/@EMBREE_RELATIV_ROOT_DIR@)
|
||||
GET_FILENAME_COMPONENT(EMBREE_ROOT_DIR "${EMBREE_ROOT_DIR}" ABSOLUTE)
|
||||
|
||||
SET(EMBREE_INCLUDE_DIRS "${EMBREE_ROOT_DIR}/@CMAKE_INSTALL_INCLUDEDIR@")
|
||||
SET(EMBREE_LIBRARY "${EMBREE_ROOT_DIR}/@CMAKE_INSTALL_LIBDIR@/@EMBREE_LIBRARY_FULLNAME@")
|
||||
SET(EMBREE_INCLUDE_DIRS ${EMBREE_ROOT_DIR}/@CMAKE_INSTALL_INCLUDEDIR@)
|
||||
SET(EMBREE_LIBRARY ${EMBREE_ROOT_DIR}/@CMAKE_INSTALL_LIBDIR@/@EMBREE_LIBRARY_NAME@)
|
||||
SET(EMBREE_LIBRARIES ${EMBREE_LIBRARY})
|
||||
|
||||
MARK_AS_ADVANCED(embree_DIR)
|
||||
|
||||
SET(EMBREE_VERSION @EMBREE_VERSION@)
|
||||
SET(EMBREE_VERSION_MAJOR @EMBREE_VERSION_MAJOR@)
|
||||
SET(EMBREE_VERSION_MINOR @EMBREE_VERSION_MINOR@)
|
||||
@@ -19,9 +34,8 @@ SET(EMBREE_ISA_SSE2 @EMBREE_ISA_SSE2@)
|
||||
SET(EMBREE_ISA_SSE42 @EMBREE_ISA_SSE42@)
|
||||
SET(EMBREE_ISA_AVX @EMBREE_ISA_AVX@)
|
||||
SET(EMBREE_ISA_AVX2 @EMBREE_ISA_AVX2@)
|
||||
SET(EMBREE_ISA_AVX512 @EMBREE_ISA_AVX512@)
|
||||
SET(EMBREE_ISA_AVX512SKX @EMBREE_ISA_AVX512@) # just for compatibility
|
||||
SET(EMBREE_ISA_NEON @EMBREE_ISA_NEON@)
|
||||
SET(EMBREE_ISA_AVX512KNL @EMBREE_ISA_AVX512KNL@)
|
||||
SET(EMBREE_ISA_AVX512SKX @EMBREE_ISA_AVX512SKX@)
|
||||
|
||||
SET(EMBREE_BUILD_TYPE @CMAKE_BUILD_TYPE@)
|
||||
SET(EMBREE_ISPC_SUPPORT @EMBREE_ISPC_SUPPORT@)
|
||||
@@ -34,49 +48,10 @@ SET(EMBREE_BACKFACE_CULLING @EMBREE_BACKFACE_CULLING@)
|
||||
SET(EMBREE_FILTER_FUNCTION @EMBREE_FILTER_FUNCTION@)
|
||||
SET(EMBREE_IGNORE_INVALID_RAYS @EMBREE_IGNORE_INVALID_RAYS@)
|
||||
SET(EMBREE_TASKING_SYSTEM @EMBREE_TASKING_SYSTEM@)
|
||||
SET(EMBREE_TBB_COMPONENT @EMBREE_TBB_COMPONENT@)
|
||||
SET(EMBREE_COMPACT_POLYS @EMBREE_COMPACT_POLYS@)
|
||||
|
||||
SET(EMBREE_GEOMETRY_TRIANGLE @EMBREE_GEOMETRY_TRIANGLE@)
|
||||
SET(EMBREE_GEOMETRY_QUAD @EMBREE_GEOMETRY_QUAD@)
|
||||
SET(EMBREE_GEOMETRY_CURVE @EMBREE_GEOMETRY_CURVE@)
|
||||
SET(EMBREE_GEOMETRY_SUBDIVISION @EMBREE_GEOMETRY_SUBDIVISION@)
|
||||
SET(EMBREE_GEOMETRY_USER @EMBREE_GEOMETRY_USER@)
|
||||
SET(EMBREE_GEOMETRY_POINT @EMBREE_GEOMETRY_POINT@)
|
||||
|
||||
SET(EMBREE_RAY_PACKETS @EMBREE_RAY_PACKETS@)
|
||||
SET(EMBREE_MAX_INSTANCE_LEVEL_COUNT @EMBREE_MAX_INSTANCE_LEVEL_COUNT@)
|
||||
SET(EMBREE_CURVE_SELF_INTERSECTION_AVOIDANCE_FACTOR @EMBREE_CURVE_SELF_INTERSECTION_AVOIDANCE_FACTOR@)
|
||||
SET(EMBREE_DISC_POINT_SELF_INTERSECTION_AVOIDANCE @EMBREE_DISC_POINT_SELF_INTERSECTION_AVOIDANCE@)
|
||||
SET(EMBREE_MIN_WIDTH @EMBREE_MIN_WIDTH@)
|
||||
|
||||
IF (EMBREE_TASKING_SYSTEM STREQUAL "TBB")
|
||||
FIND_PACKAGE(TBB REQUIRED ${EMBREE_TBB_COMPONENT})
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/sys-targets.cmake")
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/math-targets.cmake")
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/simd-targets.cmake")
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/lexers-targets.cmake")
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/tasking-targets.cmake")
|
||||
|
||||
IF (EMBREE_ISA_SSE42)
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/embree_sse42-targets.cmake")
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_AVX)
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/embree_avx-targets.cmake")
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_AVX2)
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/embree_avx2-targets.cmake")
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISA_AVX512)
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/embree_avx512-targets.cmake")
|
||||
ENDIF()
|
||||
|
||||
ENDIF()
|
||||
|
||||
INCLUDE("${EMBREE_ROOT_DIR}/@EMBREE_CMAKEEXPORT_DIR@/embree-targets.cmake")
|
||||
|
||||
@@ -1,8 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
|
||||
<plist version="1.0">
|
||||
<dict>
|
||||
<key>com.apple.security.cs.disable-library-validation</key>
|
||||
<true/>
|
||||
</dict>
|
||||
</plist>
|
||||
+42
-61
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
IF(NOT ${VAR})
|
||||
@@ -7,55 +20,31 @@ MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
SET(FLAGS_SSE2 "-D__SSE__ -D__SSE2__")
|
||||
SET(FLAGS_SSE42 "-D__SSE4_2__ -D__SSE4_1__")
|
||||
SET(FLAGS_AVX "-D__AVX__ -D__SSE4_2__ -D__SSE4_1__ -D__BMI__ -D__BMI2__ -D__LZCNT__")
|
||||
SET(FLAGS_AVX2 "-D__AVX2__ -D__AVX__ -D__SSE4_2__ -D__SSE4_1__ -D__BMI__ -D__BMI2__ -D__LZCNT__")
|
||||
ELSE ()
|
||||
_SET_IF_EMPTY(FLAGS_SSE2 "-msse2")
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-msse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-mavx")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512 "-mavx512f -mavx512dq -mavx512cd -mavx512bw -mavx512vl -mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2 -mprefer-vector-width=256")
|
||||
ENDIF ()
|
||||
_SET_IF_EMPTY(FLAGS_SSE2 "-msse2")
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-msse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-mavx")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512KNL "-mavx512f -mavx512pf -mavx512er -mavx512cd -mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512SKX "-mavx512f -mavx512dq -mavx512cd -mavx512bw -mavx512vl -mf16c -mavx2 -mfma -mlzcnt -mbmi -mbmi2")
|
||||
|
||||
OPTION(EMBREE_IGNORE_CMAKE_CXX_FLAGS "When enabled Embree ignores default CMAKE_CXX_FLAGS." ON)
|
||||
IF (EMBREE_IGNORE_CMAKE_CXX_FLAGS)
|
||||
SET(CMAKE_CXX_FLAGS "")
|
||||
ENDIF()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
|
||||
|
||||
IF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsigned-char") # treat 'char' as 'signed char'
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -flax-vector-conversions") # allow lax vector type conversions
|
||||
ENDIF (EMBREE_ARM)
|
||||
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall") # enables most warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wformat -Wformat-security") # enables string format vulnerability warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-class-memaccess") # disables clearing an object of type ‘XXX’ with no trivial copy-assignment; use assignment or value-initialization instead
|
||||
|
||||
# these prevent compile to optimize away security checks
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-overflow") # assume that signed overflow occurs
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-delete-null-pointer-checks") # keep all checks for NULL pointers
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fwrapv") # this option instructs the compiler to assume that signed arithmetic overflow warps around.
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsigned-char") # treat char as signed on all processors, including ARM
|
||||
|
||||
IF (NOT WIN32)
|
||||
IF (NOT APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIE") # enables support for more secure position independent execution
|
||||
ENDIF()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
ENDIF ()
|
||||
|
||||
IF (NOT APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIE") # enables support for more secure position independent execution
|
||||
ENDIF()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fPIC") # generate position independent code suitable for shared libraries
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11") # enables C++11 features
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fvisibility=hidden") # makes all symbols hidden by default
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fvisibility-inlines-hidden") # makes all inline symbols hidden by default
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-aliasing") # disables strict aliasing rules
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-tree-vectorize") # disable auto vectorizer
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fstack-protector") # protects against return address overrides
|
||||
ENDIF()
|
||||
@@ -66,40 +55,32 @@ MACRO(DISABLE_STACK_PROTECTOR_FOR_FILE file)
|
||||
ENDMACRO()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "")
|
||||
IF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fsigned-char") # treat 'char' as 'signed char'
|
||||
ENDIF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O2") # allow inlining of intrinsics
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "")
|
||||
IF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -fsigned-char") # treat 'char' as 'signed char'
|
||||
ENDIF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "")
|
||||
IF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -fsigned-char") # treat 'char' as 'signed char'
|
||||
ENDIF (EMBREE_ARM)
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
IF (APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mmacosx-version-min=10.7") # makes sure code runs on older MacOSX versions
|
||||
# SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++") # link against libc++ which supports C++11 features
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++") # link against libc++ which supports C++11 features
|
||||
ELSE(APPLE)
|
||||
IF (CMAKE_CXX_COMPILER_ID MATCHES "GNU")
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -Wl,--no-undefined") # issues link error for undefined symbols in shared library
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
ENDIF ()
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -pie") # enables position independent execution for executable
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -Wl,--no-undefined") # issues link error for undefined symbols in shared library
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z noexecstack") # we do not need an executable stack
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -z relro -z now") # re-arranges data sections to increase security
|
||||
SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -pie") # enables position independent execution for executable
|
||||
ENDIF(APPLE)
|
||||
|
||||
|
||||
@@ -1,24 +0,0 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
INSTALL(TARGETS TBB EXPORT TBB-targets)
|
||||
INSTALL(EXPORT TBB-targets DESTINATION "${EMBREE_CMAKEEXPORT_DIR}" COMPONENT devel)
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_INSTALL_DEPENDENCIES)
|
||||
IF (TARGET TBB::${EMBREE_TBB_COMPONENT})
|
||||
GET_TARGET_PROPERTY(LIB_PATH TBB::${EMBREE_TBB_COMPONENT} IMPORTED_LOCATION_RELEASE)
|
||||
IF(WIN32)
|
||||
INSTALL(FILES "${LIB_PATH}" DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT examples)
|
||||
GET_TARGET_PROPERTY(IMPLIB_PATH TBB::${EMBREE_TBB_COMPONENT} IMPORTED_IMPLIB_RELEASE)
|
||||
INSTALL(FILES "${IMPLIB_PATH}" DESTINATION "${CMAKE_INSTALL_LIBDIR}" COMPONENT lib)
|
||||
ELSE()
|
||||
GET_FILENAME_COMPONENT(LIB_DIR "${LIB_PATH}" DIRECTORY)
|
||||
FILE(GLOB LIB_FILES ${LIB_DIR}/libtbb.*)
|
||||
INSTALL(FILES ${LIB_FILES} DESTINATION "${CMAKE_INSTALL_LIBDIR}" COMPONENT lib)
|
||||
ENDIF()
|
||||
ELSE()
|
||||
MESSAGE(SEND_ERROR "Target TBB::${EMBREE_TBB_COMPONENT} not found during install.")
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
+45
-33
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
MACRO(_SET_IF_EMPTY VAR VALUE)
|
||||
IF(NOT ${VAR})
|
||||
@@ -13,7 +26,8 @@ IF (WIN32)
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "/QxSSE4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "/arch:AVX")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "/QxCORE-AVX2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512 "/QxCORE-AVX512")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512KNL "/QxMIC-AVX512")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512SKX "/QxCORE-AVX512")
|
||||
|
||||
SET(COMMON_CXX_FLAGS "")
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /EHsc") # catch C++ exceptions only and extern "C" functions never throw a C++ exception
|
||||
@@ -50,7 +64,7 @@ IF (WIN32)
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Qinline-factor=150") # increase default inline factors by 2x
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} ${COMMON_CXX_FLAGS}")
|
||||
# SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Gy") # package individual functions
|
||||
@@ -75,17 +89,12 @@ IF (WIN32)
|
||||
|
||||
# remove libmmd dependency
|
||||
IF (NOT EMBREE_STATIC_RUNTIME)
|
||||
# Use the default math library instead of libmmd[d]
|
||||
STRING(APPEND CMAKE_EXE_LINKER_FLAGS_DEBUG " /nodefaultlib:libmmdd.lib")
|
||||
STRING(APPEND CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO " /nodefaultlib:libmmd.lib")
|
||||
STRING(APPEND CMAKE_EXE_LINKER_FLAGS_RELEASE " /nodefaultlib:libmmd.lib")
|
||||
STRING(APPEND CMAKE_SHARED_LINKER_FLAGS_DEBUG " /nodefaultlib:libmmdd.lib")
|
||||
STRING(APPEND CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO " /nodefaultlib:libmmd.lib")
|
||||
STRING(APPEND CMAKE_SHARED_LINKER_FLAGS_RELEASE " /nodefaultlib:libmmd.lib")
|
||||
|
||||
# Link the static version of SVML
|
||||
string(APPEND CMAKE_EXE_LINKER_FLAGS " /defaultlib:svml_dispmt.lib")
|
||||
string(APPEND CMAKE_SHARED_LINKER_FLAGS " /defaultlib:svml_dispmt.lib")
|
||||
ENDIF()
|
||||
|
||||
ELSE()
|
||||
@@ -98,7 +107,8 @@ ELSE()
|
||||
_SET_IF_EMPTY(FLAGS_SSE42 "-xsse4.2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX "-xAVX")
|
||||
_SET_IF_EMPTY(FLAGS_AVX2 "-xCORE-AVX2")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512 "-xCORE-AVX512")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512KNL "-xMIC-AVX512")
|
||||
_SET_IF_EMPTY(FLAGS_AVX512SKX "-xCORE-AVX512")
|
||||
|
||||
OPTION(EMBREE_IGNORE_CMAKE_CXX_FLAGS "When enabled Embree ignores default CMAKE_CXX_FLAGS." ON)
|
||||
IF (EMBREE_IGNORE_CMAKE_CXX_FLAGS)
|
||||
@@ -108,12 +118,6 @@ ELSE()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall") # enables most warnings
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wformat -Wformat-security") # enables string format vulnerability warnings
|
||||
|
||||
# these prevent compile to optimize away security checks
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-overflow") # assume that signed overflow occurs
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-delete-null-pointer-checks") # keep all checks for NULL pointers
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fwrapv") # this option instructs the compiler to assume that signed arithmetic overflow warps around.
|
||||
|
||||
IF (NOT APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIE") # enables support for more secure position independent execution
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -ftls-model=local-dynamic") # otherwise ICC2019 cannot compile code with -fPIE enabled
|
||||
@@ -126,7 +130,6 @@ ELSE()
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-ansi-alias") # disables strict aliasing rules
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-vec") # disable auto vectorizer
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fasm-blocks") # enable assembly blocks
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fstack-protector") # protects against return address overrides
|
||||
ENDIF()
|
||||
@@ -136,31 +139,40 @@ ELSE()
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -restrict") # enable restrict keyword
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-inline-max-total-size") # no size limit when performing inlining
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-inline-max-per-compile") # no maximal number of inlinings per compilation unit
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -inline-factor=150") # increase default inline factors by 2x
|
||||
|
||||
IF(NOT CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
IF(NOT CMAKE_CXX_COMPILER_WRAPPER STREQUAL "CrayPrgEnv")
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -static-intel") # links intel runtime statically
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-intel-extensions") # disables some intel extensions which cause symbols to be exported
|
||||
ENDIF()
|
||||
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "")
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enables assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DDEBUG") # enable assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DTBB_USE_DEBUG") # configure TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -O2") # allow inlining of intrinsics
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3") # full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -restrict") # enable restrict keyword
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -no-inline-max-total-size") # no size limit when performing inlining
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -no-inline-max-per-compile") # no maximal number of inlinings per compilation unit
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -inline-factor=150") # increase default inline factors limits by 2x
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DNDEBUG") # disable assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -O3") # enable full optimizations
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DDEBUG") # enables assertions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -g") # generate debug information
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -O3") # full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -restrict") # enable restrict keyword
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -no-inline-max-total-size") # no size limit when performing inlining
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -no-inline-max-per-compile") # no maximal number of inlinings per compilation unit
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -inline-factor=150") # increase default inline factors by 2x
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -D_FORTIFY_SOURCE=2") # perform extra security checks for some standard library calls
|
||||
|
||||
#SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -g -debug inline-debug-info")
|
||||
#SET(CMAKE_EXE_LINKER_FLAGS "-g")
|
||||
|
||||
IF (APPLE)
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mmacosx-version-min=10.7") # makes sure code runs on older MacOSX versions
|
||||
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++") # link against libc++ which supports C++11 features
|
||||
|
||||
+29
-25
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
# ##################################################################
|
||||
# add macro INCLUDE_DIRECTORIES_ISPC() that allows to specify search
|
||||
@@ -12,7 +25,7 @@ ENDMACRO ()
|
||||
|
||||
IF (EMBREE_ISPC_SUPPORT)
|
||||
|
||||
# ISPC versions to look for, in descending order (newest first)
|
||||
# ISPC versions to look for, in decending order (newest first)
|
||||
SET(ISPC_VERSION_WORKING "1.9.1" "1.9.0" "1.8.3" "1.8.2")
|
||||
LIST(GET ISPC_VERSION_WORKING -1 ISPC_VERSION_REQUIRED)
|
||||
|
||||
@@ -32,7 +45,7 @@ IF (NOT EMBREE_ISPC_EXECUTABLE)
|
||||
ENDIF()
|
||||
FOREACH(ver ${ISPC_VERSION_WORKING})
|
||||
FOREACH(suffix ${ISPC_DIR_SUFFIX})
|
||||
LIST(APPEND ISPC_DIR_HINT "${PROJECT_SOURCE_DIR}/../ispc-v${ver}-${suffix}")
|
||||
LIST(APPEND ISPC_DIR_HINT ${PROJECT_SOURCE_DIR}/../ispc-v${ver}-${suffix})
|
||||
ENDFOREACH()
|
||||
ENDFOREACH()
|
||||
|
||||
@@ -45,14 +58,7 @@ IF (NOT EMBREE_ISPC_EXECUTABLE)
|
||||
ENDIF()
|
||||
|
||||
# check ISPC version
|
||||
EXECUTE_PROCESS(COMMAND ${EMBREE_ISPC_EXECUTABLE} --version
|
||||
OUTPUT_VARIABLE ISPC_OUTPUT
|
||||
RESULT_VARIABLE ISPC_RESULT)
|
||||
|
||||
IF (NOT ${ISPC_RESULT} STREQUAL "0")
|
||||
MESSAGE(FATAL_ERROR "Error executing ISPC executable '${EMBREE_ISPC_EXECUTABLE}': ${ISPC_RESULT}")
|
||||
ENDIF()
|
||||
|
||||
EXECUTE_PROCESS(COMMAND ${EMBREE_ISPC_EXECUTABLE} --version OUTPUT_VARIABLE ISPC_OUTPUT)
|
||||
STRING(REGEX MATCH "([0-9]+[.][0-9]+[.][0-9]+)" DUMMY "${ISPC_OUTPUT}")
|
||||
SET(ISPC_VERSION ${CMAKE_MATCH_1})
|
||||
|
||||
@@ -62,8 +68,7 @@ ENDIF()
|
||||
|
||||
GET_FILENAME_COMPONENT(ISPC_DIR ${EMBREE_ISPC_EXECUTABLE} PATH)
|
||||
|
||||
SET(EMBREE_ISPC_ADDRESSING 32 CACHE STRING "32vs64 bit addressing in ispc")
|
||||
SET_PROPERTY(CACHE EMBREE_ISPC_ADDRESSING PROPERTY STRINGS 32 64)
|
||||
SET(EMBREE_ISPC_ADDRESSING 32 CACHE INT "32vs64 bit addressing in ispc")
|
||||
MARK_AS_ADVANCED(EMBREE_ISPC_ADDRESSING)
|
||||
|
||||
MACRO (ISPC_COMPILE)
|
||||
@@ -73,16 +78,12 @@ MACRO (ISPC_COMPILE)
|
||||
STRING(REPLACE ";" "," ISPC_TARGET_ARGS "${ISPC_TARGETS}")
|
||||
|
||||
IF (CMAKE_SIZEOF_VOID_P EQUAL 8)
|
||||
IF (${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm64|aarch64")
|
||||
SET(ISPC_ARCHITECTURE "aarch64")
|
||||
ELSE()
|
||||
SET(ISPC_ARCHITECTURE "x86-64")
|
||||
ENDIF()
|
||||
SET(ISPC_ARCHITECTURE "x86-64")
|
||||
ELSE()
|
||||
SET(ISPC_ARCHITECTURE "x86")
|
||||
ENDIF()
|
||||
|
||||
SET(ISPC_TARGET_DIR "${CMAKE_CURRENT_BINARY_DIR}")
|
||||
SET(ISPC_TARGET_DIR ${CMAKE_CURRENT_BINARY_DIR})
|
||||
|
||||
IF(ISPC_INCLUDE_DIR)
|
||||
STRING(REPLACE ";" ";-I;" ISPC_INCLUDE_DIR_PARMS "${ISPC_INCLUDE_DIR}")
|
||||
@@ -108,7 +109,7 @@ MACRO (ISPC_COMPILE)
|
||||
GET_FILENAME_COMPONENT(dir ${src} PATH)
|
||||
|
||||
SET(outdir "${ISPC_TARGET_DIR}/${dir}")
|
||||
SET(input "${CMAKE_CURRENT_SOURCE_DIR}/${src}")
|
||||
SET(input ${CMAKE_CURRENT_SOURCE_DIR}/${src})
|
||||
|
||||
SET(deps "")
|
||||
IF (EXISTS ${outdir}/${fname}.dev.idep)
|
||||
@@ -129,7 +130,10 @@ MACRO (ISPC_COMPILE)
|
||||
LIST(LENGTH ISPC_TARGETS NUM_TARGETS)
|
||||
IF (NUM_TARGETS GREATER 1)
|
||||
FOREACH(target ${ISPC_TARGETS})
|
||||
IF (${target} STREQUAL "avx512skx-i32x16")
|
||||
# in v1.9.0 ISPC changed the ISA suffix of avx512knl-i32x16 to just 'avx512knl'
|
||||
IF (${target} STREQUAL "avx512knl-i32x16" AND NOT ISPC_VERSION VERSION_LESS "1.9.0")
|
||||
SET(target "avx512knl")
|
||||
ELSEIF (${target} STREQUAL "avx512skx-i32x16")
|
||||
SET(target "avx512skx")
|
||||
ENDIF()
|
||||
SET(results ${results} "${outdir}/${fname}.dev_${target}${ISPC_TARGET_EXT}")
|
||||
@@ -137,10 +141,10 @@ MACRO (ISPC_COMPILE)
|
||||
ENDIF()
|
||||
|
||||
ADD_CUSTOM_COMMAND(
|
||||
OUTPUT ${results} "${ISPC_TARGET_DIR}/${fname}_ispc.h"
|
||||
OUTPUT ${results} ${ISPC_TARGET_DIR}/${fname}_ispc.h
|
||||
COMMAND ${CMAKE_COMMAND} -E make_directory ${outdir}
|
||||
COMMAND ${EMBREE_ISPC_EXECUTABLE}
|
||||
-I "${CMAKE_CURRENT_SOURCE_DIR}"
|
||||
-I ${CMAKE_CURRENT_SOURCE_DIR}
|
||||
${ISPC_INCLUDE_DIR_PARMS}
|
||||
${ISPC_DEFINITIONS}
|
||||
--arch=${ISPC_ARCHITECTURE}
|
||||
@@ -150,7 +154,7 @@ MACRO (ISPC_COMPILE)
|
||||
--woff
|
||||
--opt=fast-math
|
||||
${ISPC_ADDITIONAL_ARGS}
|
||||
-h "${ISPC_TARGET_DIR}/${fname}_ispc.h"
|
||||
-h ${ISPC_TARGET_DIR}/${fname}_ispc.h
|
||||
-MMM ${outdir}/${fname}.dev.idep
|
||||
-o ${outdir}/${fname}.dev${ISPC_TARGET_EXT}
|
||||
${input}
|
||||
|
||||
+18
-5
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
SET(FLAGS_SSE2 "/D__SSE__ /D__SSE2__")
|
||||
SET(FLAGS_SSE42 "${FLAGS_SSE2} /D__SSE3__ /D__SSSE3__ /D__SSE4_1__ /D__SSE4_2__")
|
||||
@@ -10,7 +23,6 @@ SET(COMMON_CXX_FLAGS "")
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /EHsc") # catch C++ exceptions only and extern "C" functions never throw a C++ exception
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /MP") # compile source files in parallel
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GR") # enable runtime type information (on by default)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /Gy") # package individual functions
|
||||
IF (EMBREE_STACK_PROTECTOR)
|
||||
SET(COMMON_CXX_FLAGS "${COMMON_CXX_FLAGS} /GS") # protects against return address overrides
|
||||
ELSE()
|
||||
@@ -23,10 +35,8 @@ MACRO(DISABLE_STACK_PROTECTOR_FOR_FILE file)
|
||||
ENDMACRO()
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} ${COMMON_CXX_FLAGS}")
|
||||
STRING(REPLACE "/RTC1" "" CMAKE_CXX_FLAGS_DEBUG ${CMAKE_CXX_FLAGS_DEBUG}) # disable native runtime checks
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DDEBUG") # enables assertions
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_DEBUG "${CMAKE_EXE_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_DEBUG "${CMAKE_SHARED_LINKER_FLAGS_DEBUG} /DEBUG") # generate debug information
|
||||
@@ -34,10 +44,13 @@ SET(CMAKE_SHARED_LINKER_FLAGS_DEBUG "${CMAKE_SHARED_LINKER_FLAGS_DEBUG} /DEBUG")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Gy") # package individual functions
|
||||
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} ${COMMON_CXX_FLAGS}")
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /DTBB_USE_DEBUG") # configures TBB in debug mode
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Ox") # enable full optimizations
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Oi") # inline intrinsic functions
|
||||
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Gy") # package individual functions
|
||||
SET(CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
SET(CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO} /DEBUG") # generate debug information
|
||||
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
# optionally use static runtime library
|
||||
OPTION(EMBREE_STATIC_RUNTIME "Use the static version of the C/C++ runtime library." OFF)
|
||||
@@ -8,6 +21,6 @@ IF (EMBREE_STATIC_RUNTIME)
|
||||
STRING(REPLACE "/MD" "/MT" CMAKE_CXX_FLAGS_RELWITHDEBINFO ${CMAKE_CXX_FLAGS_RELWITHDEBINFO})
|
||||
STRING(REPLACE "/MD" "/MT" CMAKE_CXX_FLAGS_RELEASE ${CMAKE_CXX_FLAGS_RELEASE})
|
||||
ENDIF()
|
||||
|
||||
|
||||
# remove define NDEBUG and instead set define DEBUG for config RelWithDebInfo
|
||||
STRING(REPLACE "NDEBUG" "DEBUG" CMAKE_CXX_FLAGS_RELWITHDEBINFO ${CMAKE_CXX_FLAGS_RELWITHDEBINFO})
|
||||
|
||||
+136
-91
@@ -1,62 +1,62 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
INCLUDE(GNUInstallDirs)
|
||||
|
||||
if (NOT DEFINED EMBREE_ZIP_MODE)
|
||||
set(EMBREE_ZIP_MODE OFF)
|
||||
endif()
|
||||
|
||||
IF (NOT EMBREE_ZIP_MODE AND NOT WIN32 AND NOT APPLE)
|
||||
SET(CMAKE_INSTALL_BINDIR "${CMAKE_INSTALL_BINDIR}/embree${EMBREE_VERSION_MAJOR}")
|
||||
SET(CMAKE_INSTALL_FULL_BINDIR "${CMAKE_INSTALL_FULL_BINDIR}/embree${EMBREE_VERSION_MAJOR}")
|
||||
SET(CMAKE_INSTALL_BINDIR ${CMAKE_INSTALL_BINDIR}/embree${EMBREE_VERSION_MAJOR})
|
||||
SET(CMAKE_INSTALL_FULL_BINDIR ${CMAKE_INSTALL_FULL_BINDIR}/embree${EMBREE_VERSION_MAJOR})
|
||||
ENDIF()
|
||||
|
||||
# use full absolute path as install name
|
||||
IF (NOT EMBREE_ZIP_MODE)
|
||||
SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_FULL_LIBDIR}")
|
||||
SET(CMAKE_INSTALL_NAME_DIR ${CMAKE_INSTALL_FULL_LIBDIR})
|
||||
ELSE()
|
||||
IF(APPLE)
|
||||
SET(CMAKE_INSTALL_RPATH "@loader_path/../${CMAKE_INSTALL_LIBDIR}")
|
||||
SET(CMAKE_INSTALL_RPATH "@loader_path/../lib")
|
||||
ELSE()
|
||||
SET(CMAKE_INSTALL_RPATH "$ORIGIN/../${CMAKE_INSTALL_LIBDIR}")
|
||||
SET(CMAKE_INSTALL_RPATH "$ORIGIN/../lib")
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
##############################################################
|
||||
# Install MSVC runtime
|
||||
##############################################################
|
||||
|
||||
IF (WIN32)
|
||||
SET(CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS_SKIP TRUE)
|
||||
INCLUDE(InstallRequiredSystemLibraries)
|
||||
LIST(FILTER CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS INCLUDE REGEX ".*msvcp[0-9]+\.dll|.*vcruntime[0-9]+\.dll|.*vcruntime[0-9]+_[0-9]+\.dll")
|
||||
INSTALL(FILES ${CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS} DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT lib)
|
||||
ENDIF()
|
||||
|
||||
##############################################################
|
||||
# Install Headers
|
||||
##############################################################
|
||||
|
||||
INSTALL(DIRECTORY include/embree3 DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}" COMPONENT devel)
|
||||
INSTALL(DIRECTORY include/embree3 DESTINATION ${CMAKE_INSTALL_INCLUDEDIR} COMPONENT devel)
|
||||
IF (NOT WIN32)
|
||||
INSTALL(DIRECTORY man/man3 DESTINATION "${CMAKE_INSTALL_MANDIR}" COMPONENT devel)
|
||||
INSTALL(DIRECTORY man/man3 DESTINATION ${CMAKE_INSTALL_MANDIR} COMPONENT devel)
|
||||
ENDIF()
|
||||
|
||||
##############################################################
|
||||
# Install Models
|
||||
##############################################################
|
||||
|
||||
IF (EMBREE_TUTORIALS)
|
||||
INSTALL(DIRECTORY tutorials/models DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT examples)
|
||||
ENDIF()
|
||||
INSTALL(DIRECTORY tutorials/models DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT examples)
|
||||
|
||||
##############################################################
|
||||
# Install Documentation
|
||||
##############################################################
|
||||
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/LICENSE.txt" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/CHANGELOG.md" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/README.md" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/readme.pdf" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/third-party-programs.txt" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_SOURCE_DIR}/third-party-programs-TBB.txt" DESTINATION "${CMAKE_INSTALL_DOCDIR}" COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_SOURCE_DIR}/LICENSE.txt DESTINATION ${CMAKE_INSTALL_DOCDIR} COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_SOURCE_DIR}/CHANGELOG.md DESTINATION ${CMAKE_INSTALL_DOCDIR} COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_SOURCE_DIR}/README.md DESTINATION ${CMAKE_INSTALL_DOCDIR} COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_SOURCE_DIR}/readme.pdf DESTINATION ${CMAKE_INSTALL_DOCDIR} COMPONENT lib)
|
||||
|
||||
##############################################################
|
||||
# Install scripts to set embree paths
|
||||
@@ -65,15 +65,15 @@ INSTALL(FILES "${PROJECT_SOURCE_DIR}/third-party-programs-TBB.txt" DESTINATION "
|
||||
IF (EMBREE_ZIP_MODE)
|
||||
IF (WIN32)
|
||||
ELSEIF(APPLE)
|
||||
CONFIGURE_FILE("${PROJECT_SOURCE_DIR}/scripts/install_macosx/embree-vars.sh" embree-vars.sh @ONLY)
|
||||
CONFIGURE_FILE("${PROJECT_SOURCE_DIR}/scripts/install_macosx/embree-vars.csh" embree-vars.csh @ONLY)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-vars.sh" DESTINATION "." COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-vars.csh" DESTINATION "." COMPONENT lib)
|
||||
CONFIGURE_FILE(${PROJECT_SOURCE_DIR}/scripts/install_macosx/embree-vars.sh embree-vars.sh @ONLY)
|
||||
CONFIGURE_FILE(${PROJECT_SOURCE_DIR}/scripts/install_macosx/embree-vars.csh embree-vars.csh @ONLY)
|
||||
INSTALL(FILES ${PROJECT_BINARY_DIR}/embree-vars.sh DESTINATION "." COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_BINARY_DIR}/embree-vars.csh DESTINATION "." COMPONENT lib)
|
||||
ELSE()
|
||||
CONFIGURE_FILE("${PROJECT_SOURCE_DIR}/scripts/install_linux/embree-vars.sh" embree-vars.sh @ONLY)
|
||||
CONFIGURE_FILE("${PROJECT_SOURCE_DIR}/scripts/install_linux/embree-vars.csh" embree-vars.csh @ONLY)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-vars.sh" DESTINATION "." COMPONENT lib)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-vars.csh" DESTINATION "." COMPONENT lib)
|
||||
CONFIGURE_FILE(${PROJECT_SOURCE_DIR}/scripts/install_linux/embree-vars.sh embree-vars.sh @ONLY)
|
||||
CONFIGURE_FILE(${PROJECT_SOURCE_DIR}/scripts/install_linux/embree-vars.csh embree-vars.csh @ONLY)
|
||||
INSTALL(FILES ${PROJECT_BINARY_DIR}/embree-vars.sh DESTINATION "." COMPONENT lib)
|
||||
INSTALL(FILES ${PROJECT_BINARY_DIR}/embree-vars.csh DESTINATION "." COMPONENT lib)
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
@@ -87,58 +87,52 @@ ELSE()
|
||||
SET(EMBREE_CONFIG_VERSION ${EMBREE_VERSION_MAJOR})
|
||||
ENDIF()
|
||||
|
||||
# why does this have to be so complicated...
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
SET(EMBREE_LIBRARY_FULLNAME ${CMAKE_STATIC_LIBRARY_PREFIX}${EMBREE_LIBRARY_NAME}${CMAKE_STATIC_LIBRARY_SUFFIX})
|
||||
ELSE()
|
||||
IF (WIN32)
|
||||
SET(EMBREE_LIBRARY_FULLNAME ${CMAKE_IMPORT_LIBRARY_PREFIX}${EMBREE_LIBRARY_NAME}${CMAKE_IMPORT_LIBRARY_SUFFIX})
|
||||
ELSEIF (APPLE)
|
||||
SET(EMBREE_LIBRARY_FULLNAME ${CMAKE_SHARED_LIBRARY_PREFIX}${EMBREE_LIBRARY_NAME}.${EMBREE_CONFIG_VERSION}${CMAKE_SHARED_LIBRARY_SUFFIX})
|
||||
ELSE()
|
||||
SET(EMBREE_LIBRARY_FULLNAME ${CMAKE_SHARED_LIBRARY_PREFIX}${EMBREE_LIBRARY_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX}.${EMBREE_CONFIG_VERSION})
|
||||
ENDIF()
|
||||
IF (APPLE AND NOT EMBREE_ZIP_MODE)
|
||||
CONFIGURE_FILE(scripts/install_macosx/uninstall.command uninstall.command @ONLY)
|
||||
INSTALL(PROGRAMS "${PROJECT_BINARY_DIR}/uninstall.command" DESTINATION ${CMAKE_INSTALL_BINDIR}/.. COMPONENT lib)
|
||||
ENDIF()
|
||||
|
||||
#IF (WIN32 OR EMBREE_ZIP_MODE)
|
||||
# why does this have to be so complicated...
|
||||
IF (WIN32)
|
||||
SET(EMBREE_LIBRARY_NAME ${CMAKE_IMPORT_LIBRARY_PREFIX}embree${EMBREE_VERSION_MAJOR}${CMAKE_IMPORT_LIBRARY_SUFFIX})
|
||||
ELSEIF (APPLE)
|
||||
SET(EMBREE_LIBRARY_NAME ${CMAKE_SHARED_LIBRARY_PREFIX}embree${EMBREE_VERSION_MAJOR}.${EMBREE_CONFIG_VERSION}${CMAKE_SHARED_LIBRARY_SUFFIX})
|
||||
ELSE()
|
||||
SET(EMBREE_LIBRARY_NAME ${CMAKE_SHARED_LIBRARY_PREFIX}embree${EMBREE_VERSION_MAJOR}${CMAKE_SHARED_LIBRARY_SUFFIX}.${EMBREE_CONFIG_VERSION})
|
||||
ENDIF()
|
||||
|
||||
IF (WIN32 OR EMBREE_ZIP_MODE)
|
||||
# for local "installs" and on Windows we want the cmake config files placed
|
||||
# in the install root, such that users can point the CMake variable
|
||||
# embree_DIR just to the install folder
|
||||
# SET(EMBREE_CMAKECONFIG_DIR ".")
|
||||
# SET(EMBREE_CMAKEEXPORT_DIR "cmake")
|
||||
# SET(EMBREE_RELATIVE_ROOT_DIR ".")
|
||||
#ELSE()
|
||||
SET(EMBREE_CMAKECONFIG_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/embree-${EMBREE_VERSION}")
|
||||
SET(EMBREE_CMAKEEXPORT_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/embree-${EMBREE_VERSION}")
|
||||
IF (WIN32)
|
||||
SET(EMBREE_RELATIVE_ROOT_DIR "../../../")
|
||||
SET(EMBREE_CMAKECONFIG_DIR ".")
|
||||
SET(EMBREE_RELATIV_ROOT_DIR ".")
|
||||
ELSE()
|
||||
FILE(RELATIVE_PATH EMBREE_RELATIVE_ROOT_DIR "/${EMBREE_CMAKECONFIG_DIR}" "/")
|
||||
SET(EMBREE_CMAKECONFIG_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/embree-${EMBREE_VERSION}")
|
||||
FILE(RELATIVE_PATH EMBREE_RELATIV_ROOT_DIR "/${EMBREE_CMAKECONFIG_DIR}" "/")
|
||||
ENDIF()
|
||||
#ENDIF()
|
||||
|
||||
CONFIGURE_FILE(common/cmake/embree-config.cmake embree-config-install.cmake @ONLY)
|
||||
CONFIGURE_FILE(common/cmake/embree-config-version.cmake embree-config-version.cmake @ONLY)
|
||||
# create a config file for the build directory
|
||||
CONFIGURE_FILE(common/cmake/embree-config-builddir.cmake embree-config.cmake @ONLY)
|
||||
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-config-install.cmake" DESTINATION "${EMBREE_CMAKECONFIG_DIR}" RENAME "embree-config.cmake" COMPONENT devel)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-config-version.cmake" DESTINATION "${EMBREE_CMAKECONFIG_DIR}" COMPONENT devel)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-config-install.cmake" DESTINATION ${EMBREE_CMAKECONFIG_DIR} RENAME "embree-config.cmake" COMPONENT devel)
|
||||
INSTALL(FILES "${PROJECT_BINARY_DIR}/embree-config-version.cmake" DESTINATION ${EMBREE_CMAKECONFIG_DIR} COMPONENT devel)
|
||||
|
||||
##############################################################
|
||||
# CPack specific stuff
|
||||
##############################################################
|
||||
|
||||
SET(CPACK_PACKAGE_NAME "Intel(R) Embree Ray Tracing Kernels")
|
||||
SET(CPACK_PACKAGE_NAME "Embree Ray Tracing Kernels")
|
||||
SET(CPACK_PACKAGE_FILE_NAME "embree-${EMBREE_VERSION}")
|
||||
#SET(CPACK_PACKAGE_ICON "${PROJECT_SOURCE_DIR}/embree-doc/images/icon.png")
|
||||
#SET(CPACK_PACKAGE_ICON ${PROJECT_SOURCE_DIR}/embree-doc/images/icon.png)
|
||||
#SET(CPACK_PACKAGE_RELOCATABLE TRUE)
|
||||
SET(CPACK_STRIP_FILES TRUE)
|
||||
SET(CPACK_STRIP_FILES FALSE) # we strip already as a post_build step
|
||||
|
||||
SET(CPACK_PACKAGE_VERSION_MAJOR ${EMBREE_VERSION_MAJOR})
|
||||
SET(CPACK_PACKAGE_VERSION_MINOR ${EMBREE_VERSION_MINOR})
|
||||
SET(CPACK_PACKAGE_VERSION_PATCH ${EMBREE_VERSION_PATCH})
|
||||
SET(CPACK_PACKAGE_DESCRIPTION_SUMMARY "Intel(R) Embree implements high performance ray tracing kernels including acceleration structure construction and traversal.")
|
||||
SET(CPACK_PACKAGE_DESCRIPTION_SUMMARY "Embree: High Performance Ray Tracing Kernels")
|
||||
SET(CPACK_PACKAGE_VENDOR "Intel Corporation")
|
||||
SET(CPACK_PACKAGE_CONTACT embree_support@intel.com)
|
||||
|
||||
@@ -157,8 +151,8 @@ SET(CPACK_COMPONENT_EXAMPLES_DESCRIPTION "Tutorials demonstrating how to use Emb
|
||||
#SET(CPACK_COMPONENT_LIB_REQUIRED ON) # always install the libs
|
||||
|
||||
# point to readme and license files
|
||||
SET(CPACK_RESOURCE_FILE_README "${PROJECT_SOURCE_DIR}/README.md")
|
||||
SET(CPACK_RESOURCE_FILE_LICENSE "${PROJECT_SOURCE_DIR}/LICENSE.txt")
|
||||
SET(CPACK_RESOURCE_FILE_README ${PROJECT_SOURCE_DIR}/README.md)
|
||||
SET(CPACK_RESOURCE_FILE_LICENSE ${PROJECT_SOURCE_DIR}/LICENSE.txt)
|
||||
|
||||
# Windows specific settings
|
||||
IF(WIN32)
|
||||
@@ -177,39 +171,90 @@ IF(WIN32)
|
||||
SET(VCVER vc14)
|
||||
ENDIF()
|
||||
|
||||
SET(CPACK_GENERATOR ZIP)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.${ARCH}.${VCVER}.windows")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}" COMMAND "${PROJECT_SOURCE_DIR}/scripts/package_win.bat" ${CMAKE_BUILD_TYPE} "${CPACK_PACKAGE_FILE_NAME}.zip")
|
||||
IF (NOT EMBREE_ZIP_MODE)
|
||||
SET(CPACK_GENERATOR WIX)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.${ARCH}.${VCVER}")
|
||||
SET(CPACK_PACKAGE_INSTALL_DIRECTORY "Intel\\\\Embree${EMBREE_VERSION_MAJOR} ${ARCH}")
|
||||
SET(CPACK_WIX_PRODUCT_GUID "331BD5A9-DAD6-486B-A435-18AAB80${EMBREE_VERSION_NUMBER}")
|
||||
SET(CPACK_WIX_UPGRADE_GUID "331BD5A9-DAD6-486B-A435-18AAB80${EMBREE_VERSION_MAJOR}0000") # upgrade as long as major version is the same
|
||||
SET(CPACK_WIX_CMAKE_PACKAGE_REGISTRY TRUE)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_win.bat ${CMAKE_BUILD_TYPE} ${CPACK_PACKAGE_FILE_NAME}.msi ${EMBREE_SIGN_FILE})
|
||||
ENDIF()
|
||||
ELSE()
|
||||
SET(CPACK_GENERATOR ZIP)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.${ARCH}.${VCVER}.windows")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_win.bat ${CMAKE_BUILD_TYPE} ${CPACK_PACKAGE_FILE_NAME}.zip)
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
# MacOSX specific settings
|
||||
ELSEIF(APPLE)
|
||||
|
||||
CONFIGURE_FILE(README.md README.txt)
|
||||
SET(CPACK_RESOURCE_FILE_README "${PROJECT_BINARY_DIR}/README.txt")
|
||||
SET(CPACK_RESOURCE_FILE_README ${PROJECT_BINARY_DIR}/README.txt)
|
||||
|
||||
SET(CPACK_GENERATOR ZIP)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.x86_64.macosx")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}" COMMAND "${PROJECT_SOURCE_DIR}/scripts/package_macosx.sh" ${CMAKE_BUILD_TYPE} "${CPACK_PACKAGE_FILE_NAME}.zip" "${EMBREE_SIGN_FILE}")
|
||||
IF (NOT EMBREE_ZIP_MODE)
|
||||
SET(CPACK_PACKAGING_INSTALL_PREFIX ${CMAKE_INSTALL_PREFIX})
|
||||
SET(CPACK_GENERATOR productbuild)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.x86_64")
|
||||
SET(CPACK_COMPONENTS_ALL lib devel examples)
|
||||
#SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
SET(CPACK_PACKAGE_NAME embree-${EMBREE_VERSION})
|
||||
SET(CPACK_PACKAGE_VENDOR "intel") # creates short name com.intel.embree3.xxx in pkgutil
|
||||
SET(CPACK_OSX_PACKAGE_VERSION 10.7)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_macosx.sh ${CMAKE_BUILD_TYPE} ${CPACK_PACKAGE_FILE_NAME}.pkg ${EMBREE_SIGN_FILE})
|
||||
ENDIF()
|
||||
ELSE()
|
||||
SET(CPACK_GENERATOR TGZ)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.x86_64.macosx")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_macosx.sh ${CMAKE_BUILD_TYPE} ${CPACK_PACKAGE_FILE_NAME}.tgz ${EMBREE_SIGN_FILE})
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
# Linux specific settings
|
||||
ELSE()
|
||||
|
||||
SET(CPACK_GENERATOR TGZ)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.x86_64.linux")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}" COMMAND "${PROJECT_SOURCE_DIR}/scripts/package_linux.sh"
|
||||
${EMBREE_ZIP_MODE} ${EMBREE_LIBRARY_NAME} ${EMBREE_VERSION} ${EMBREE_VERSION_MAJOR} "${EMBREE_SIGN_FILE}")
|
||||
IF (NOT EMBREE_ZIP_MODE)
|
||||
|
||||
SET(CPACK_GENERATOR RPM)
|
||||
SET(CPACK_RPM_PACKAGE_NAME "embree${EMBREE_VERSION_MAJOR}")
|
||||
SET(CPACK_COMPONENTS_ALL devel lib examples)
|
||||
SET(CPACK_RPM_COMPONENT_INSTALL ON)
|
||||
SET(CPACK_RPM_FILE_NAME RPM-DEFAULT)
|
||||
SET(CPACK_RPM_devel_PACKAGE_ARCHITECTURE noarch)
|
||||
SET(CPACK_RPM_PACKAGE_LICENSE "ASL 2.0") # Apache Software License, Version 2.0
|
||||
SET(CPACK_RPM_PACKAGE_GROUP "Development/Libraries")
|
||||
SET(CPACK_RPM_CHANGELOG_FILE ${CMAKE_BINARY_DIR}/rpm_changelog.txt) # ChangeLog of the RPM
|
||||
IF (CMAKE_VERSION VERSION_LESS "3.7.0")
|
||||
EXECUTE_PROCESS(COMMAND date "+%a %b %d %Y" OUTPUT_VARIABLE CHANGELOG_DATE OUTPUT_STRIP_TRAILING_WHITESPACE)
|
||||
ELSE()
|
||||
STRING(TIMESTAMP CHANGELOG_DATE "%a %b %d %Y")
|
||||
ENDIF()
|
||||
SET(RPM_CHANGELOG "* ${CHANGELOG_DATE} Johannes Günther <johannes.guenther@intel.com> - ${EMBREE_VERSION}-${CPACK_RPM_PACKAGE_RELEASE}\n- First package")
|
||||
FILE(WRITE ${CPACK_RPM_CHANGELOG_FILE} ${RPM_CHANGELOG})
|
||||
SET(CPACK_RPM_PACKAGE_URL http://embree.github.io/)
|
||||
SET(CPACK_RPM_DEFAULT_DIR_PERMISSIONS OWNER_READ OWNER_WRITE OWNER_EXECUTE GROUP_READ GROUP_EXECUTE WORLD_READ WORLD_EXECUTE) # otherwise cmake directory has wrong 775 permissions
|
||||
|
||||
# post install and uninstall scripts
|
||||
SET(CPACK_RPM_lib_POST_INSTALL_SCRIPT_FILE ${PROJECT_SOURCE_DIR}/common/cmake/rpm_ldconfig.sh)
|
||||
SET(CPACK_RPM_lib_POST_UNINSTALL_SCRIPT_FILE ${PROJECT_SOURCE_DIR}/common/cmake/rpm_ldconfig.sh)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_linux.sh ${EMBREE_ZIP_MODE} ${EMBREE_VERSION} ${EMBREE_VERSION_MAJOR} ${EMBREE_SIGN_FILE})
|
||||
ENDIF()
|
||||
ELSE()
|
||||
|
||||
SET(CPACK_GENERATOR TGZ)
|
||||
SET(CPACK_PACKAGE_FILE_NAME "${CPACK_PACKAGE_FILE_NAME}.x86_64.linux")
|
||||
SET(CPACK_MONOLITHIC_INSTALL 1)
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
ADD_TEST(NAME "BuildPackage" WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/package_linux.sh ${EMBREE_ZIP_MODE} ${EMBREE_VERSION} ${EMBREE_SIGN_FILE})
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_TESTING_PACKAGE)
|
||||
SET_TESTS_PROPERTIES(BuildPackage PROPERTIES TIMEOUT 1200)
|
||||
ENDIF()
|
||||
|
||||
@@ -1,4 +1 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
/sbin/ldconfig
|
||||
|
||||
+118
-54
@@ -1,14 +1,23 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
INCLUDE(CTest)
|
||||
|
||||
IF (WIN32)
|
||||
IF(${CMAKE_CXX_COMPILER} MATCHES ".*icx")
|
||||
SET(MY_PROJECT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}")
|
||||
ELSE()
|
||||
SET(MY_PROJECT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_BUILD_TYPE}")
|
||||
ENDIF()
|
||||
SET(MY_PROJECT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_BUILD_TYPE}")
|
||||
ELSE()
|
||||
SET(MY_PROJECT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}")
|
||||
ENDIF()
|
||||
@@ -19,8 +28,7 @@ FIND_PATH(EMBREE_TESTING_MODEL_DIR
|
||||
DOC "Path to the folder containing the Embree models for regression testing."
|
||||
NO_DEFAULT_PATHS)
|
||||
|
||||
SET(EMBREE_TESTING_INTENSITY 1 CACHE STRING "Intensity of testing (0 = no testing, 1 = verify and tutorials, 2 = light testing, 3 = intensive testing.")
|
||||
SET_PROPERTY(CACHE EMBREE_TESTING_INTENSITY PROPERTY STRINGS 0 1 2 3)
|
||||
SET(EMBREE_TESTING_INTENSITY 1 CACHE INT "Intensity of testing (0 = no testing, 1 = verify and tutorials, 2 = light testing, 3 = intensive testing.")
|
||||
SET(EMBREE_TESTING_MEMCHECK OFF CACHE BOOL "Turns on memory checking for some tests.")
|
||||
SET(EMBREE_TESTING_BENCHMARK OFF CACHE BOOL "Turns benchmarking on.")
|
||||
SET(EMBREE_TESTING_BENCHMARK_DATABASE "${PROJECT_BINARY_DIR}" CACHE PATH "Path to database for benchmarking.")
|
||||
@@ -31,28 +39,35 @@ SET_PROPERTY(CACHE EMBREE_TESTING_SDE PROPERTY STRINGS OFF pnr nhm wsm snb ivb h
|
||||
|
||||
SET(EMBREE_MODEL_DIR "none")
|
||||
IF (EMBREE_TESTING_MODEL_DIR)
|
||||
SET(EMBREE_MODEL_DIR "${EMBREE_TESTING_MODEL_DIR}")
|
||||
SET(EMBREE_MODEL_DIR ${EMBREE_TESTING_MODEL_DIR})
|
||||
ENDIF()
|
||||
|
||||
MACRO (ADD_EMBREE_NORMAL_CPP_TEST name reference executable args)
|
||||
MACRO (ADD_EMBREE_NORMAL_TEST name reference executable args)
|
||||
IF (BUILD_TESTING)
|
||||
ADD_TEST(NAME ${name}
|
||||
WORKING_DIRECTORY "${MY_PROJECT_BINARY_DIR}"
|
||||
COMMAND ${executable} --compare "${EMBREE_MODEL_DIR}/reference/${reference}.tga" ${args})
|
||||
WORKING_DIRECTORY ${MY_PROJECT_BINARY_DIR}
|
||||
COMMAND python ${PROJECT_SOURCE_DIR}/scripts/invoke_test.py
|
||||
--name ${name}
|
||||
--reference ${reference}
|
||||
--modeldir ${EMBREE_MODEL_DIR}
|
||||
--model default
|
||||
--sde ${EMBREE_TESTING_SDE}
|
||||
--execute ${MY_PROJECT_BINARY_DIR}/${executable} ${args})
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISPC_SUPPORT AND NOT DISABLE_ISPC_TEST)
|
||||
IF (BUILD_TESTING)
|
||||
ADD_TEST(NAME ${name}_ispc
|
||||
WORKING_DIRECTORY ${MY_PROJECT_BINARY_DIR}
|
||||
COMMAND python ${PROJECT_SOURCE_DIR}/scripts/invoke_test.py
|
||||
--name ${name}_ispc
|
||||
--reference ${reference}
|
||||
--modeldir ${EMBREE_MODEL_DIR}
|
||||
--model default
|
||||
--sde ${EMBREE_TESTING_SDE}
|
||||
--execute ${MY_PROJECT_BINARY_DIR}/${executable}_ispc ${args})
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_NORMAL_ISPC_TEST name reference executable args)
|
||||
IF (BUILD_TESTING AND EMBREE_ISPC_SUPPORT AND EMBREE_RAY_PACKETS)
|
||||
ADD_TEST(NAME ${name}_ispc
|
||||
WORKING_DIRECTORY "${MY_PROJECT_BINARY_DIR}"
|
||||
COMMAND ${executable}_ispc --compare "${EMBREE_MODEL_DIR}/reference/${reference}.tga" ${args})
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_NORMAL_TEST name reference executable args)
|
||||
ADD_EMBREE_NORMAL_CPP_TEST(${name} ${reference} ${executable} "${args}")
|
||||
ADD_EMBREE_NORMAL_ISPC_TEST(${name} ${reference} ${executable} "${args}")
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_TEST name)
|
||||
@@ -60,55 +75,104 @@ MACRO (ADD_EMBREE_TEST name)
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_TEST2 name exe args)
|
||||
ADD_EMBREE_NORMAL_TEST(${name} ${exe} ${exe} "${args}")
|
||||
ADD_EMBREE_MODEL_TEST(${name} ${exe} ${exe} "${args}" "default")
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_MODEL_TEST name reference executable args model)
|
||||
IF (BUILD_TESTING)
|
||||
ADD_TEST(NAME ${name}
|
||||
WORKING_DIRECTORY "${MY_PROJECT_BINARY_DIR}"
|
||||
COMMAND ${executable} -c "${EMBREE_MODEL_DIR}/${model}" --compare "${EMBREE_MODEL_DIR}/reference/${reference}.tga" ${args})
|
||||
WORKING_DIRECTORY ${MY_PROJECT_BINARY_DIR}
|
||||
COMMAND python ${PROJECT_SOURCE_DIR}/scripts/invoke_test.py
|
||||
--name ${name}
|
||||
--reference ${reference}
|
||||
--modeldir ${EMBREE_MODEL_DIR}
|
||||
--model ${model}
|
||||
--sde ${EMBREE_TESTING_SDE}
|
||||
--execute ${MY_PROJECT_BINARY_DIR}/${executable} ${args})
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_ISPC_SUPPORT AND EMBREE_RAY_PACKETS)
|
||||
IF (EMBREE_ISPC_SUPPORT AND NOT DISABLE_ISPC_TEST)
|
||||
IF (BUILD_TESTING)
|
||||
ADD_TEST(NAME ${name}_ispc
|
||||
WORKING_DIRECTORY "${MY_PROJECT_BINARY_DIR}"
|
||||
COMMAND COMMAND ${executable}_ispc -c "${EMBREE_MODEL_DIR}/${model}" --compare "${EMBREE_MODEL_DIR}/reference/${reference}.tga" ${args})
|
||||
WORKING_DIRECTORY ${MY_PROJECT_BINARY_DIR}
|
||||
COMMAND python ${PROJECT_SOURCE_DIR}/scripts/invoke_test.py
|
||||
--name ${name}_ispc
|
||||
--reference ${reference}
|
||||
--modeldir ${EMBREE_MODEL_DIR}
|
||||
--model ${model}
|
||||
--sde ${EMBREE_TESTING_SDE}
|
||||
--execute ${MY_PROJECT_BINARY_DIR}/${executable}_ispc ${args})
|
||||
ENDIF()
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_MODELS_TEST model_list_file name reference executable)
|
||||
IF (BUILD_TESTING)
|
||||
IF (EMBREE_TESTING_MODEL_DIR AND EMBREE_TESTING_INTENSITY GREATER 0)
|
||||
|
||||
IF( NOT EXISTS "${EMBREE_TESTING_MODEL_DIR}/test-models-subdiv.txt"
|
||||
OR NOT EXISTS "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity2.txt"
|
||||
OR NOT EXISTS "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity3.txt"
|
||||
OR NOT EXISTS "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity4.txt")
|
||||
MESSAGE(FATAL_ERROR "Invalid Embree testing model repository.")
|
||||
ENDIF()
|
||||
|
||||
FILE(READ "${EMBREE_TESTING_MODEL_DIR}/test-models-subdiv.txt" models_subdiv)
|
||||
STRING(REGEX REPLACE "\n" ";" models_subdiv "${models_subdiv}")
|
||||
|
||||
SET(full_model_list_file "${EMBREE_TESTING_MODEL_DIR}/${model_list_file}")
|
||||
|
||||
IF(NOT EXISTS "${full_model_list_file}")
|
||||
MESSAGE(FATAL_ERROR "File ${EMBREE_TESTING_MODEL_DIR}/${model_list_file} does not exist!")
|
||||
ENDIF()
|
||||
FILE(READ "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity2.txt" models_intensity2)
|
||||
STRING(REGEX REPLACE "\n" ";" models_intensity2 "${models_intensity2}")
|
||||
|
||||
FILE(READ "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity3.txt" models_intensity3)
|
||||
STRING(REGEX REPLACE "\n" ";" models_intensity3 "${models_intensity3}")
|
||||
|
||||
FILE(READ "${EMBREE_TESTING_MODEL_DIR}/test-models-intensity4.txt" models_intensity4)
|
||||
STRING(REGEX REPLACE "\n" ";" models_intensity4 "${models_intensity4}")
|
||||
|
||||
FILE(READ "${full_model_list_file}" models)
|
||||
STRING(REGEX REPLACE "\n" ";" models "${models}")
|
||||
|
||||
IF (EMBREE_TESTING_INTENSITY GREATER 1)
|
||||
LIST(APPEND models ${models_intensity2})
|
||||
ENDIF()
|
||||
|
||||
IF (EMBREE_TESTING_INTENSITY GREATER 2)
|
||||
LIST(APPEND models ${models_intensity3})
|
||||
ENDIF()
|
||||
|
||||
CMAKE_HOST_SYSTEM_INFORMATION(RESULT memory QUERY TOTAL_PHYSICAL_MEMORY)
|
||||
IF (EMBREE_TESTING_INTENSITY GREATER 3 AND ${memory} GREATER 10000)
|
||||
LIST(APPEND models ${models_intensity4})
|
||||
ENDIF()
|
||||
|
||||
MACRO (ADD_EMBREE_MODELS_TEST name reference executable)
|
||||
FOREACH (model ${models})
|
||||
IF (model MATCHES "^#")
|
||||
CONTINUE()
|
||||
ENDIF()
|
||||
STRING(REGEX REPLACE " .*" "" modelname "${model}")
|
||||
STRING(REGEX REPLACE "/" "_" modelname "${modelname}")
|
||||
STRING(REGEX REPLACE "/" "_" modelname "${model}")
|
||||
STRING(REGEX REPLACE ".ecs" "" modelname "${modelname}")
|
||||
ADD_EMBREE_MODEL_TEST(${name}_${modelname} ${reference}_${modelname} ${executable} "${ARGN}" ${model})
|
||||
ENDFOREACH()
|
||||
ENDIF()
|
||||
ENDMACRO()
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_SUBDIV_MODELS_TEST name reference executable)
|
||||
FOREACH (model ${models_subdiv})
|
||||
STRING(REGEX REPLACE "/" "_" modelname "${model}")
|
||||
STRING(REGEX REPLACE ".ecs" "" modelname "${modelname}")
|
||||
ADD_EMBREE_MODEL_TEST(${name}_${modelname} ${reference}_${modelname} ${executable} "${ARGN}" ${model})
|
||||
ENDFOREACH()
|
||||
ENDMACRO()
|
||||
|
||||
ELSE()
|
||||
|
||||
MACRO (ADD_EMBREE_MODEL_TEST name reference executable args model)
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_SUBDIV_MODELS_TEST name reference executable)
|
||||
ENDMACRO()
|
||||
|
||||
MACRO (ADD_EMBREE_MODELS_TEST name)
|
||||
ENDMACRO()
|
||||
|
||||
ENDIF()
|
||||
|
||||
# add klocwork test
|
||||
IF (EMBREE_TESTING_KLOCWORK)
|
||||
ADD_TEST(NAME Klocwork-Build WORKING_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}" COMMAND "${PROJECT_SOURCE_DIR}/scripts/klocwork_build.sh")
|
||||
ADD_TEST(NAME Klocwork-Check WORKING_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}" COMMAND "${PROJECT_SOURCE_DIR}/scripts/klocwork_check.sh")
|
||||
SET_TESTS_PROPERTIES(Klocwork-Build PROPERTIES TIMEOUT 2400)
|
||||
SET_TESTS_PROPERTIES(Klocwork-Check PROPERTIES TIMEOUT 300)
|
||||
ADD_TEST(NAME Klocwork WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR} COMMAND ${PROJECT_SOURCE_DIR}/scripts/klocwork.sh)
|
||||
SET_TESTS_PROPERTIES(Klocwork PROPERTIES TIMEOUT 2400)
|
||||
ENDIF()
|
||||
|
||||
# add valgrind test
|
||||
@@ -121,6 +185,6 @@ IF (EMBREE_TESTING_MEMCHECK)
|
||||
FUNCTION(ADD_MEMCHECK_TEST name binary)
|
||||
set(memcheck_command "${EMBREE_MEMORYCHECK_COMMAND} ${EMBREE_MEMORYCHECK_COMMAND_OPTIONS}")
|
||||
separate_arguments(memcheck_command)
|
||||
add_test(NAME ${name} COMMAND ${memcheck_command} "${MY_PROJECT_BINARY_DIR}/${binary}" ${ARGN})
|
||||
add_test(NAME ${name} COMMAND ${memcheck_command} ${MY_PROJECT_BINARY_DIR}/${binary} ${ARGN})
|
||||
ENDFUNCTION()
|
||||
ENDIF()
|
||||
|
||||
@@ -1,25 +1,38 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
# additional parameters (beyond the name) are treated as additional dependencies
|
||||
# if ADDITIONAL_LIBRARIES is set these will be included during linking
|
||||
|
||||
MACRO (ADD_TUTORIAL TUTORIAL_NAME)
|
||||
ADD_EXECUTABLE(${TUTORIAL_NAME} ../../kernels/embree.rc ${TUTORIAL_NAME}.cpp ${TUTORIAL_NAME}_device.cpp ${ARGN})
|
||||
ADD_EXECUTABLE(${TUTORIAL_NAME} ${TUTORIAL_NAME}.cpp ${TUTORIAL_NAME}_device.cpp ${ARGN})
|
||||
TARGET_LINK_LIBRARIES(${TUTORIAL_NAME} embree image tutorial noise ${ADDITIONAL_LIBRARIES})
|
||||
SET_PROPERTY(TARGET ${TUTORIAL_NAME} PROPERTY FOLDER tutorials/single)
|
||||
SET_PROPERTY(TARGET ${TUTORIAL_NAME} APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
INSTALL(TARGETS ${TUTORIAL_NAME} DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT examples)
|
||||
INSTALL(TARGETS ${TUTORIAL_NAME} DESTINATION ${CMAKE_INSTALL_BINDIR} COMPONENT examples)
|
||||
SIGN_TARGET(${TUTORIAL_NAME})
|
||||
ENDMACRO ()
|
||||
|
||||
MACRO (ADD_TUTORIAL_ISPC TUTORIAL_NAME)
|
||||
IF (EMBREE_ISPC_SUPPORT)
|
||||
ADD_EMBREE_ISPC_EXECUTABLE(${TUTORIAL_NAME}_ispc ../../kernels/embree.rc ${TUTORIAL_NAME}.cpp ${TUTORIAL_NAME}_device.ispc ${ARGN})
|
||||
ADD_EMBREE_ISPC_EXECUTABLE(${TUTORIAL_NAME}_ispc ${TUTORIAL_NAME}.cpp ${TUTORIAL_NAME}_device.ispc)
|
||||
TARGET_LINK_LIBRARIES(${TUTORIAL_NAME}_ispc embree image tutorial_ispc noise noise_ispc)
|
||||
SET_PROPERTY(TARGET ${TUTORIAL_NAME}_ispc PROPERTY FOLDER tutorials/ispc)
|
||||
SET_PROPERTY(TARGET ${TUTORIAL_NAME}_ispc APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
INSTALL(TARGETS ${TUTORIAL_NAME}_ispc DESTINATION "${CMAKE_INSTALL_BINDIR}" COMPONENT examples)
|
||||
INSTALL(TARGETS ${TUTORIAL_NAME}_ispc DESTINATION ${CMAKE_INSTALL_BINDIR} COMPONENT examples)
|
||||
SIGN_TARGET(${TUTORIAL_NAME}_ispc)
|
||||
ENDIF()
|
||||
ENDMACRO ()
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2015 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
IF(NOT EXISTS "@CMAKE_CURRENT_BINARY_DIR@/install_manifest.txt")
|
||||
MESSAGE(FATAL_ERROR "Cannot find install manifest: @CMAKE_CURRENT_BINARY_DIR@/install_manifest.txt")
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
ADD_LIBRARY(lexers STATIC
|
||||
stringstream.cpp
|
||||
@@ -10,8 +23,6 @@ SET_PROPERTY(TARGET lexers PROPERTY FOLDER common)
|
||||
SET_PROPERTY(TARGET lexers APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
INSTALL(TARGETS lexers EXPORT lexers-targets ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}" COMPONENT devel)
|
||||
INSTALL(EXPORT lexers-targets DESTINATION "${EMBREE_CMAKEEXPORT_DIR}" COMPONENT devel)
|
||||
INSTALL(TARGETS lexers ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT devel)
|
||||
SIGN_TARGET(lexers)
|
||||
ENDIF()
|
||||
|
||||
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
+15
-2
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
@@ -1,12 +1,23 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "stringstream.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
static const std::string stringChars = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 _.,+-=:/*\\";
|
||||
|
||||
/* creates map for fast categorization of characters */
|
||||
static void createCharMap(bool map[256], const std::string& chrs) {
|
||||
for (size_t i=0; i<256; i++) map[i] = false;
|
||||
@@ -18,7 +29,6 @@ namespace embree
|
||||
: cin(cin), endl(endl), multiLine(multiLine)
|
||||
{
|
||||
createCharMap(isSepMap,seps);
|
||||
createCharMap(isValidCharMap,stringChars);
|
||||
}
|
||||
|
||||
std::string StringStream::next()
|
||||
@@ -37,11 +47,8 @@ namespace embree
|
||||
|
||||
/* parse everything until the next separator */
|
||||
std::vector<char> str; str.reserve(64);
|
||||
while (cin->peek() != EOF && !isSeparator(cin->peek())) {
|
||||
int c = cin->get();
|
||||
if (!isValidChar(c)) throw std::runtime_error("invalid character "+std::string(1,c)+" in input");
|
||||
str.push_back((char)c);
|
||||
}
|
||||
while (cin->peek() != EOF && !isSeparator(cin->peek()))
|
||||
str.push_back((char)cin->get());
|
||||
str.push_back(0);
|
||||
return std::string(str.data());
|
||||
}
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -17,12 +30,10 @@ namespace embree
|
||||
ParseLocation location() { return cin->loc(); }
|
||||
std::string next();
|
||||
private:
|
||||
__forceinline bool isSeparator(unsigned int c) const { return c<256 && isSepMap[c]; }
|
||||
__forceinline bool isValidChar(unsigned int c) const { return c<256 && isValidCharMap[c]; }
|
||||
__forceinline bool isSeparator(int c) const { return c<256 && isSepMap[c]; }
|
||||
private:
|
||||
Ref<Stream<int> > cin; /*! source character stream */
|
||||
bool isSepMap[256]; /*! map for fast classification of separators */
|
||||
bool isValidCharMap[256]; /*! map for valid characters */
|
||||
std::string endl; /*! the token of the end of line */
|
||||
bool multiLine; /*! whether to parse lines wrapped with \ */
|
||||
};
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "tokenstream.h"
|
||||
#include "../math/math.h"
|
||||
@@ -11,7 +24,6 @@ namespace embree
|
||||
const std::string TokenStream::ALPHA = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
|
||||
const std::string TokenStream::numbers = "0123456789";
|
||||
const std::string TokenStream::separators = "\n\t\r ";
|
||||
const std::string TokenStream::stringChars = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 _.,+-=:/*\\";
|
||||
|
||||
/* creates map for fast categorization of characters */
|
||||
static void createCharMap(bool map[256], const std::string& chrs) {
|
||||
@@ -28,7 +40,6 @@ namespace embree
|
||||
{
|
||||
createCharMap(isAlphaMap,alpha);
|
||||
createCharMap(isSepMap,seps);
|
||||
createCharMap(isStringCharMap,stringChars);
|
||||
}
|
||||
|
||||
bool TokenStream::decDigits(std::string& str_o)
|
||||
@@ -138,11 +149,7 @@ namespace embree
|
||||
std::string str;
|
||||
if (cin->peek() != '\"') return false;
|
||||
cin->drop();
|
||||
while (cin->peek() != '\"') {
|
||||
const int c = cin->get();
|
||||
if (!isStringChar(c)) THROW_RUNTIME_ERROR("invalid string character "+std::string(1,c)+" at "+loc.str());
|
||||
str += (char)c;
|
||||
}
|
||||
while (cin->peek() != '\"') str += (char)cin->get();
|
||||
cin->drop();
|
||||
token = Token(str,Token::TY_STRING,loc);
|
||||
return true;
|
||||
|
||||
+19
-11
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -120,7 +133,6 @@ namespace embree
|
||||
static const std::string ALPHA;
|
||||
static const std::string numbers;
|
||||
static const std::string separators;
|
||||
static const std::string stringChars;
|
||||
|
||||
public:
|
||||
TokenStream(const Ref<Stream<int> >& cin,
|
||||
@@ -145,20 +157,16 @@ namespace embree
|
||||
Ref<Stream<int> > cin;
|
||||
bool isSepMap[256];
|
||||
bool isAlphaMap[256];
|
||||
bool isStringCharMap[256];
|
||||
std::vector<std::string> symbols;
|
||||
|
||||
/*! checks if a character is a separator */
|
||||
__forceinline bool isSeparator(unsigned int c) const { return c<256 && isSepMap[c]; }
|
||||
__forceinline bool isSeparator(int c) const { return c<256 && isSepMap[c]; }
|
||||
|
||||
/*! checks if a character is a number */
|
||||
__forceinline bool isDigit(unsigned int c) const { return c >= '0' && c <= '9'; }
|
||||
|
||||
/*! checks if a character is valid inside a string */
|
||||
__forceinline bool isStringChar(unsigned int c) const { return c<256 && isStringCharMap[c]; }
|
||||
__forceinline bool isDigit(int c) const { return c >= '0' && c <= '9';}
|
||||
|
||||
/*! checks if a character is legal for an identifier */
|
||||
__forceinline bool isAlpha(unsigned int c) const { return c<256 && isAlphaMap[c]; }
|
||||
__forceinline bool isAlphaNum(unsigned int c) const { return isAlpha(c) || isDigit(c); }
|
||||
__forceinline bool isAlpha(int c) const { return c<256 && isAlphaMap[c]; }
|
||||
__forceinline bool isAlphaNum(int c) const { return isAlpha(c) || isDigit(c); }
|
||||
};
|
||||
}
|
||||
|
||||
@@ -1,12 +1,25 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
ADD_LIBRARY(math STATIC constants.cpp)
|
||||
SET_PROPERTY(TARGET math PROPERTY FOLDER common)
|
||||
SET_PROPERTY(TARGET math APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
INSTALL(TARGETS math EXPORT math-targets ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}" COMPONENT devel)
|
||||
INSTALL(EXPORT math-targets DESTINATION "${EMBREE_CMAKEEXPORT_DIR}" COMPONENT devel)
|
||||
INSTALL(TARGETS math ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT devel)
|
||||
SIGN_TARGET(math)
|
||||
ENDIF()
|
||||
|
||||
|
||||
+24
-213
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -7,7 +20,6 @@
|
||||
#include "linearspace3.h"
|
||||
#include "quaternion.h"
|
||||
#include "bbox.h"
|
||||
#include "vec4.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
@@ -69,10 +81,6 @@ namespace embree
|
||||
}
|
||||
|
||||
};
|
||||
|
||||
// template specialization to get correct identity matrix for type AffineSpace3fa
|
||||
template<>
|
||||
__forceinline AffineSpaceT<LinearSpace3ff>::AffineSpaceT( OneTy ) : l(one), p(0.f, 0.f, 0.f, 1.f) {}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
// Unary Operators
|
||||
@@ -132,11 +140,17 @@ namespace embree
|
||||
return AffineSpaceT<L>(select(s,t.l,f.l),select(s,t.p,f.p));
|
||||
}
|
||||
|
||||
/*! blending */
|
||||
template<typename T>
|
||||
__forceinline AffineSpaceT<T> lerp(const AffineSpaceT<T>& l0, const AffineSpaceT<T>& l1, const float t) {
|
||||
return AffineSpaceT<T>(lerp(l0.l,l1.l,t),lerp(l0.p,l1.p,t));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename L> static embree_ostream operator<<(embree_ostream cout, const AffineSpaceT<L>& m) {
|
||||
template<typename L> static std::ostream& operator<<(std::ostream& cout, const AffineSpaceT<L>& m) {
|
||||
return cout << "{ l = " << m.l << ", p = " << m.p << " }";
|
||||
}
|
||||
|
||||
@@ -147,214 +161,11 @@ namespace embree
|
||||
typedef AffineSpaceT<LinearSpace2f> AffineSpace2f;
|
||||
typedef AffineSpaceT<LinearSpace3f> AffineSpace3f;
|
||||
typedef AffineSpaceT<LinearSpace3fa> AffineSpace3fa;
|
||||
typedef AffineSpaceT<LinearSpace3fx> AffineSpace3fx;
|
||||
typedef AffineSpaceT<LinearSpace3ff> AffineSpace3ff;
|
||||
typedef AffineSpaceT<Quaternion3f > OrthonormalSpace3f;
|
||||
|
||||
template<int N> using AffineSpace3vf = AffineSpaceT<LinearSpace3<Vec3<vfloat<N>>>>;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec3<vfloat<4>>>> AffineSpace3vf4;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec3<vfloat<8>>>> AffineSpace3vf8;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec3<vfloat<16>>>> AffineSpace3vf16;
|
||||
|
||||
template<int N> using AffineSpace3vff = AffineSpaceT<LinearSpace3<Vec4<vfloat<N>>>>;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec4<vfloat<4>>>> AffineSpace3vfa4;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec4<vfloat<8>>>> AffineSpace3vfa8;
|
||||
typedef AffineSpaceT<LinearSpace3<Vec4<vfloat<16>>>> AffineSpace3vfa16;
|
||||
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
/// Interpolation
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
template<typename T, typename R>
|
||||
__forceinline AffineSpaceT<T> lerp(const AffineSpaceT<T>& M0,
|
||||
const AffineSpaceT<T>& M1,
|
||||
const R& t)
|
||||
{
|
||||
return AffineSpaceT<T>(lerp(M0.l,M1.l,t),lerp(M0.p,M1.p,t));
|
||||
}
|
||||
|
||||
// slerp interprets the 16 floats of the matrix M = D * R * S as components of
|
||||
// three matrizes (D, R, S) that are interpolated individually.
|
||||
template<typename T> __forceinline AffineSpaceT<LinearSpace3<Vec3<T>>>
|
||||
slerp(const AffineSpaceT<LinearSpace3<Vec4<T>>>& M0,
|
||||
const AffineSpaceT<LinearSpace3<Vec4<T>>>& M1,
|
||||
const T& t)
|
||||
{
|
||||
QuaternionT<T> q0(M0.p.w, M0.l.vx.w, M0.l.vy.w, M0.l.vz.w);
|
||||
QuaternionT<T> q1(M1.p.w, M1.l.vx.w, M1.l.vy.w, M1.l.vz.w);
|
||||
QuaternionT<T> q = slerp(q0, q1, t);
|
||||
|
||||
AffineSpaceT<LinearSpace3<Vec3<T>>> S = lerp(M0, M1, t);
|
||||
AffineSpaceT<LinearSpace3<Vec3<T>>> D(one);
|
||||
D.p.x = S.l.vx.y;
|
||||
D.p.y = S.l.vx.z;
|
||||
D.p.z = S.l.vy.z;
|
||||
S.l.vx.y = 0;
|
||||
S.l.vx.z = 0;
|
||||
S.l.vy.z = 0;
|
||||
|
||||
AffineSpaceT<LinearSpace3<Vec3<T>>> R = LinearSpace3<Vec3<T>>(q);
|
||||
return D * R * S;
|
||||
}
|
||||
|
||||
// this is a specialized version for Vec3fa because that does
|
||||
// not play along nicely with the other templated Vec3/Vec4 types
|
||||
__forceinline AffineSpace3fa slerp(const AffineSpace3ff& M0,
|
||||
const AffineSpace3ff& M1,
|
||||
const float& t)
|
||||
{
|
||||
Quaternion3f q0(M0.p.w, M0.l.vx.w, M0.l.vy.w, M0.l.vz.w);
|
||||
Quaternion3f q1(M1.p.w, M1.l.vx.w, M1.l.vy.w, M1.l.vz.w);
|
||||
Quaternion3f q = slerp(q0, q1, t);
|
||||
|
||||
AffineSpace3fa S = lerp(M0, M1, t);
|
||||
AffineSpace3fa D(one);
|
||||
D.p.x = S.l.vx.y;
|
||||
D.p.y = S.l.vx.z;
|
||||
D.p.z = S.l.vy.z;
|
||||
S.l.vx.y = 0;
|
||||
S.l.vx.z = 0;
|
||||
S.l.vy.z = 0;
|
||||
|
||||
AffineSpace3fa R = LinearSpace3fa(q);
|
||||
return D * R * S;
|
||||
}
|
||||
|
||||
__forceinline AffineSpace3fa quaternionDecompositionToAffineSpace(const AffineSpace3ff& qd)
|
||||
{
|
||||
// compute affine transform from quaternion decomposition
|
||||
Quaternion3f q(qd.p.w, qd.l.vx.w, qd.l.vy.w, qd.l.vz.w);
|
||||
AffineSpace3fa M = qd;
|
||||
AffineSpace3fa D(one);
|
||||
D.p.x = M.l.vx.y;
|
||||
D.p.y = M.l.vx.z;
|
||||
D.p.z = M.l.vy.z;
|
||||
M.l.vx.y = 0;
|
||||
M.l.vx.z = 0;
|
||||
M.l.vy.z = 0;
|
||||
AffineSpace3fa R = LinearSpace3fa(q);
|
||||
return D * R * M;
|
||||
}
|
||||
|
||||
__forceinline void quaternionDecomposition(const AffineSpace3ff& qd, Vec3fa& T, Quaternion3f& q, AffineSpace3fa& S)
|
||||
{
|
||||
q = Quaternion3f(qd.p.w, qd.l.vx.w, qd.l.vy.w, qd.l.vz.w);
|
||||
S = qd;
|
||||
T.x = qd.l.vx.y;
|
||||
T.y = qd.l.vx.z;
|
||||
T.z = qd.l.vy.z;
|
||||
S.l.vx.y = 0;
|
||||
S.l.vx.z = 0;
|
||||
S.l.vy.z = 0;
|
||||
}
|
||||
|
||||
__forceinline AffineSpace3fx quaternionDecomposition(Vec3fa const& T, Quaternion3f const& q, AffineSpace3fa const& S)
|
||||
{
|
||||
AffineSpace3ff M = S;
|
||||
M.l.vx.w = q.i;
|
||||
M.l.vy.w = q.j;
|
||||
M.l.vz.w = q.k;
|
||||
M.p.w = q.r;
|
||||
M.l.vx.y = T.x;
|
||||
M.l.vx.z = T.y;
|
||||
M.l.vy.z = T.z;
|
||||
return M;
|
||||
}
|
||||
|
||||
struct __aligned(16) QuaternionDecomposition
|
||||
{
|
||||
float scale_x = 1.f;
|
||||
float scale_y = 1.f;
|
||||
float scale_z = 1.f;
|
||||
float skew_xy = 0.f;
|
||||
float skew_xz = 0.f;
|
||||
float skew_yz = 0.f;
|
||||
float shift_x = 0.f;
|
||||
float shift_y = 0.f;
|
||||
float shift_z = 0.f;
|
||||
float quaternion_r = 1.f;
|
||||
float quaternion_i = 0.f;
|
||||
float quaternion_j = 0.f;
|
||||
float quaternion_k = 0.f;
|
||||
float translation_x = 0.f;
|
||||
float translation_y = 0.f;
|
||||
float translation_z = 0.f;
|
||||
};
|
||||
|
||||
__forceinline QuaternionDecomposition quaternionDecomposition(AffineSpace3ff const& M)
|
||||
{
|
||||
QuaternionDecomposition qd;
|
||||
qd.scale_x = M.l.vx.x;
|
||||
qd.scale_y = M.l.vy.y;
|
||||
qd.scale_z = M.l.vz.z;
|
||||
qd.shift_x = M.p.x;
|
||||
qd.shift_y = M.p.y;
|
||||
qd.shift_z = M.p.z;
|
||||
qd.translation_x = M.l.vx.y;
|
||||
qd.translation_y = M.l.vx.z;
|
||||
qd.translation_z = M.l.vy.z;
|
||||
qd.skew_xy = M.l.vy.x;
|
||||
qd.skew_xz = M.l.vz.x;
|
||||
qd.skew_yz = M.l.vz.y;
|
||||
qd.quaternion_r = M.p.w;
|
||||
qd.quaternion_i = M.l.vx.w;
|
||||
qd.quaternion_j = M.l.vy.w;
|
||||
qd.quaternion_k = M.l.vz.w;
|
||||
return qd;
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/*
|
||||
* ! Template Specialization for 2D: return matrix for rotation around point
|
||||
* (rotation around arbitrarty vector is not meaningful in 2D)
|
||||
*/
|
||||
template<> __forceinline
|
||||
AffineSpace2f AffineSpace2f::rotate(const Vec2f& p, const float& r) {
|
||||
return translate(+p)*AffineSpace2f(LinearSpace2f::rotate(r))*translate(-p);
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
// Similarity Transform
|
||||
//
|
||||
// checks, if M is a similarity transformation, i.e if there exists a factor D
|
||||
// such that for all x,y: distance(Mx, My) = D * distance(x, y)
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
__forceinline bool similarityTransform(const AffineSpace3fa& M, float* D)
|
||||
{
|
||||
if (D) *D = 0.f;
|
||||
if (abs(dot(M.l.vx, M.l.vy)) > 1e-5f) return false;
|
||||
if (abs(dot(M.l.vx, M.l.vz)) > 1e-5f) return false;
|
||||
if (abs(dot(M.l.vy, M.l.vz)) > 1e-5f) return false;
|
||||
|
||||
const float D_x = dot(M.l.vx, M.l.vx);
|
||||
const float D_y = dot(M.l.vy, M.l.vy);
|
||||
const float D_z = dot(M.l.vz, M.l.vz);
|
||||
|
||||
if (abs(D_x - D_y) > 1e-5f ||
|
||||
abs(D_x - D_z) > 1e-5f ||
|
||||
abs(D_y - D_z) > 1e-5f)
|
||||
return false;
|
||||
|
||||
if (D) *D = sqrtf(D_x);
|
||||
return true;
|
||||
}
|
||||
|
||||
__forceinline void AffineSpace3fa_store_unaligned(const AffineSpace3fa &source, AffineSpace3fa* ptr)
|
||||
{
|
||||
Vec3fa::storeu(&ptr->l.vx, source.l.vx);
|
||||
Vec3fa::storeu(&ptr->l.vy, source.l.vy);
|
||||
Vec3fa::storeu(&ptr->l.vz, source.l.vz);
|
||||
Vec3fa::storeu(&ptr->p, source.p);
|
||||
}
|
||||
|
||||
__forceinline AffineSpace3fa AffineSpace3fa_load_unaligned(AffineSpace3fa* ptr)
|
||||
{
|
||||
AffineSpace3fa space;
|
||||
space.l.vx = Vec3fa::loadu(&ptr->l.vx);
|
||||
space.l.vy = Vec3fa::loadu(&ptr->l.vy);
|
||||
space.l.vz = Vec3fa::loadu(&ptr->l.vz);
|
||||
space.p = Vec3fa::loadu(&ptr->p);
|
||||
return space;
|
||||
}
|
||||
/*! Template Specialization for 2D: return matrix for rotation around point (rotation around arbitrarty vector is not meaningful in 2D) */
|
||||
template<> __forceinline AffineSpace2f AffineSpace2f::rotate(const Vec2f& p, const float& r) { return translate(+p) * AffineSpace2f(LinearSpace2f::rotate(r)) * translate(-p); }
|
||||
|
||||
#undef VectorT
|
||||
#undef ScalarT
|
||||
|
||||
+25
-37
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -8,13 +21,6 @@
|
||||
|
||||
namespace embree
|
||||
{
|
||||
namespace internal {
|
||||
|
||||
template <typename T> __forceinline T divideByTwo(const T& v) { return v / T(2); }
|
||||
template <> __forceinline float divideByTwo<float>(const float& v) { return v * 0.5f; }
|
||||
template <> __forceinline double divideByTwo<double>(const double& v) { return v * 0.5; }
|
||||
|
||||
} // namespace internal
|
||||
template<typename T>
|
||||
struct BBox
|
||||
{
|
||||
@@ -25,8 +31,7 @@ namespace embree
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline BBox ( ) { }
|
||||
template<typename T1>
|
||||
__forceinline BBox ( const BBox<T1>& other ) : lower(other.lower), upper(other.upper) {}
|
||||
__forceinline BBox ( const BBox& other ) { lower = other.lower; upper = other.upper; }
|
||||
__forceinline BBox& operator=( const BBox& other ) { lower = other.lower; upper = other.upper; return *this; }
|
||||
|
||||
__forceinline BBox ( const T& v ) : lower(v), upper(v) {}
|
||||
@@ -46,7 +51,7 @@ namespace embree
|
||||
__forceinline T size() const { return upper - lower; }
|
||||
|
||||
/*! computes the center of the box */
|
||||
__forceinline T center() const { return internal::divideByTwo<T>(lower+upper); }
|
||||
__forceinline T center() const { return 0.5f*(lower+upper); }
|
||||
|
||||
/*! computes twice the center of the box */
|
||||
__forceinline T center2() const { return lower+upper; }
|
||||
@@ -57,7 +62,7 @@ namespace embree
|
||||
}
|
||||
|
||||
/*! enlarge box by some scaling factor */
|
||||
__forceinline BBox enlarge_by(const float a) const {
|
||||
__forceinline BBox enlarge_by(const float a) {
|
||||
return BBox(lower - T(a)*abs(lower), upper + T(a)*abs(upper));
|
||||
}
|
||||
|
||||
@@ -77,13 +82,10 @@ namespace embree
|
||||
return lower > upper;
|
||||
}
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined(__SSE__)
|
||||
template<> __forceinline bool BBox<Vec3fa>::empty() const {
|
||||
return !all(le_mask(lower,upper));
|
||||
}
|
||||
template<> __forceinline bool BBox<Vec3fx>::empty() const {
|
||||
return !all(le_mask(lower,upper));
|
||||
}
|
||||
#endif
|
||||
|
||||
/*! tests if box is finite */
|
||||
@@ -91,11 +93,6 @@ namespace embree
|
||||
return all(gt_mask(v.lower,Vec3fa_t(-FLT_LARGE)) & lt_mask(v.upper,Vec3fa_t(+FLT_LARGE)));
|
||||
}
|
||||
|
||||
/*! tests if box is finite and non-empty*/
|
||||
__forceinline bool isvalid_non_empty( const BBox<Vec3fa>& v ) {
|
||||
return all(gt_mask(v.lower,Vec3fa_t(-FLT_LARGE)) & lt_mask(v.upper,Vec3fa_t(+FLT_LARGE)) & le_mask(v.lower,v.upper));
|
||||
}
|
||||
|
||||
/*! tests if box has finite entries */
|
||||
__forceinline bool is_finite( const BBox<Vec3fa>& b) {
|
||||
return is_finite(b.lower) && is_finite(b.upper);
|
||||
@@ -106,7 +103,7 @@ namespace embree
|
||||
|
||||
/*! computes the center of the box */
|
||||
template<typename T> __forceinline const T center2(const BBox<T>& box) { return box.lower + box.upper; }
|
||||
template<typename T> __forceinline const T center (const BBox<T>& box) { return internal::divideByTwo<T>(center2(box)); }
|
||||
template<typename T> __forceinline const T center (const BBox<T>& box) { return T(0.5f)*center2(box); }
|
||||
|
||||
/*! computes the volume of a bounding box */
|
||||
__forceinline float volume ( const BBox<Vec3fa>& b ) { return reduce_mul(b.size()); }
|
||||
@@ -119,14 +116,11 @@ namespace embree
|
||||
template<typename T> __forceinline const T area( const BBox<Vec2<T> >& b ) { const Vec2<T> d = b.size(); return d.x*d.y; }
|
||||
|
||||
template<typename T> __forceinline const T halfArea( const BBox<Vec3<T> >& b ) { return halfArea(b.size()); }
|
||||
template<typename T> __forceinline const T area( const BBox<Vec3<T> >& b ) { return T(2)*halfArea(b); }
|
||||
template<typename T> __forceinline const T area( const BBox<Vec3<T> >& b ) { return 2.0f*halfArea(b); }
|
||||
|
||||
__forceinline float halfArea( const BBox<Vec3fa>& b ) { return halfArea(b.size()); }
|
||||
__forceinline float area( const BBox<Vec3fa>& b ) { return 2.0f*halfArea(b); }
|
||||
|
||||
__forceinline float halfArea( const BBox<Vec3fx>& b ) { return halfArea(b.size()); }
|
||||
__forceinline float area( const BBox<Vec3fx>& b ) { return 2.0f*halfArea(b); }
|
||||
|
||||
template<typename Vec> __forceinline float safeArea( const BBox<Vec>& b ) { if (b.empty()) return 0.0f; else return area(b); }
|
||||
|
||||
template<typename T> __forceinline float expectedApproxHalfArea(const BBox<T>& box) {
|
||||
@@ -196,11 +190,7 @@ namespace embree
|
||||
}
|
||||
|
||||
template<> __inline bool subset( const BBox<Vec3fa>& a, const BBox<Vec3fa>& b ) {
|
||||
return all(ge_mask(a.lower,b.lower)) && all(le_mask(a.upper,b.upper));
|
||||
}
|
||||
|
||||
template<> __inline bool subset( const BBox<Vec3fx>& a, const BBox<Vec3fx>& b ) {
|
||||
return all(ge_mask(a.lower,b.lower)) && all(le_mask(a.upper,b.upper));
|
||||
return all(ge_mask(a.lower,b.lower)) & all(le_mask(a.upper,b.upper));
|
||||
}
|
||||
|
||||
/*! blending */
|
||||
@@ -210,7 +200,7 @@ namespace embree
|
||||
}
|
||||
|
||||
/*! output operator */
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const BBox<T>& box) {
|
||||
template<typename T> __forceinline std::ostream& operator<<(std::ostream& cout, const BBox<T>& box) {
|
||||
return cout << "[" << box.lower << "; " << box.upper << "]";
|
||||
}
|
||||
|
||||
@@ -220,19 +210,17 @@ namespace embree
|
||||
typedef BBox<Vec2fa> BBox2fa;
|
||||
typedef BBox<Vec3f> BBox3f;
|
||||
typedef BBox<Vec3fa> BBox3fa;
|
||||
typedef BBox<Vec3fx> BBox3fx;
|
||||
typedef BBox<Vec3ff> BBox3ff;
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// SSE / AVX / MIC specializations
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined (__SSE__) || defined(__ARM_NEON)
|
||||
#if defined __SSE__
|
||||
#include "../simd/sse.h"
|
||||
#endif
|
||||
|
||||
#if defined (__AVX__)
|
||||
#if defined __AVX__
|
||||
#include "../simd/avx.h"
|
||||
#endif
|
||||
|
||||
|
||||
+16
-3
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -37,7 +50,7 @@ namespace embree
|
||||
};
|
||||
|
||||
/*! output operator */
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const Col3<T>& a) {
|
||||
template<typename T> inline std::ostream& operator<<(std::ostream& cout, const Col3<T>& a) {
|
||||
return cout << "(" << a.r << ", " << a.g << ", " << a.b << ")";
|
||||
}
|
||||
|
||||
|
||||
+16
-3
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -37,7 +50,7 @@ namespace embree
|
||||
};
|
||||
|
||||
/*! output operator */
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const Col4<T>& a) {
|
||||
template<typename T> inline std::ostream& operator<<(std::ostream& cout, const Col4<T>& a) {
|
||||
return cout << "(" << a.r << ", " << a.g << ", " << a.b << ", " << a.a << ")";
|
||||
}
|
||||
|
||||
|
||||
+17
-21
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -152,38 +165,21 @@ namespace embree
|
||||
}
|
||||
__forceinline const Color rcp ( const Color& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
__m128 reciprocal = _mm_rcp_ps(a.m128);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a.m128, reciprocal), reciprocal);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a.m128, reciprocal), reciprocal);
|
||||
return (const Color)reciprocal;
|
||||
#else
|
||||
#if defined(__AVX512VL__)
|
||||
const Color r = _mm_rcp14_ps(a.m128);
|
||||
#else
|
||||
const Color r = _mm_rcp_ps(a.m128);
|
||||
#endif
|
||||
return _mm_add_ps(r,_mm_mul_ps(r, _mm_sub_ps(_mm_set1_ps(1.0f), _mm_mul_ps(a, r)))); // computes r + r * (1 - a * r)
|
||||
|
||||
#endif //defined(__aarch64__)
|
||||
return _mm_sub_ps(_mm_add_ps(r, r), _mm_mul_ps(_mm_mul_ps(r, r), a));
|
||||
}
|
||||
__forceinline const Color rsqrt( const Color& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
return r;
|
||||
#else
|
||||
|
||||
#if defined(__AVX512VL__)
|
||||
__m128 r = _mm_rsqrt14_ps(a.m128);
|
||||
#else
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
#endif
|
||||
return _mm_add_ps(_mm_mul_ps(_mm_set1_ps(1.5f),r), _mm_mul_ps(_mm_mul_ps(_mm_mul_ps(a, _mm_set1_ps(-0.5f)), r), _mm_mul_ps(r, r)));
|
||||
|
||||
#endif //defined(__aarch64__)
|
||||
}
|
||||
__forceinline const Color sqrt ( const Color& a ) { return _mm_sqrt_ps(a.m128); }
|
||||
|
||||
@@ -252,7 +248,7 @@ namespace embree
|
||||
__forceinline float luminance (const Color& a) { return madd(0.212671f,a.r,madd(0.715160f,a.g,0.072169f*a.b)); }
|
||||
|
||||
/*! output operator */
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Color& a) {
|
||||
inline std::ostream& operator<<(std::ostream& cout, const Color& a) {
|
||||
return cout << "(" << a.r << ", " << a.g << ", " << a.b << ")";
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,8 +1,40 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#include "constants.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
TrueTy True;
|
||||
FalseTy False;
|
||||
ZeroTy zero;
|
||||
OneTy one;
|
||||
NegInfTy neg_inf;
|
||||
PosInfTy inf;
|
||||
PosInfTy pos_inf;
|
||||
NaNTy nan;
|
||||
UlpTy ulp;
|
||||
PiTy pi;
|
||||
OneOverPiTy one_over_pi;
|
||||
TwoPiTy two_pi;
|
||||
OneOverTwoPiTy one_over_two_pi;
|
||||
FourPiTy four_pi;
|
||||
OneOverFourPiTy one_over_four_pi;
|
||||
StepTy step;
|
||||
ReverseStepTy reverse_step;
|
||||
EmptyTy empty;
|
||||
UndefinedTy undefined;
|
||||
}
|
||||
|
||||
+36
-36
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -10,7 +23,6 @@
|
||||
#define _USE_MATH_DEFINES
|
||||
#include <math.h> // using cmath causes issues under Windows
|
||||
#include <cfloat>
|
||||
#include <climits>
|
||||
|
||||
namespace embree
|
||||
{
|
||||
@@ -24,13 +36,13 @@ namespace embree
|
||||
__forceinline operator bool( ) const { return true; }
|
||||
};
|
||||
|
||||
const constexpr TrueTy True = TrueTy();
|
||||
extern MAYBE_UNUSED TrueTy True;
|
||||
|
||||
struct FalseTy {
|
||||
__forceinline operator bool( ) const { return false; }
|
||||
};
|
||||
|
||||
const constexpr FalseTy False = FalseTy();
|
||||
extern MAYBE_UNUSED FalseTy False;
|
||||
|
||||
struct ZeroTy
|
||||
{
|
||||
@@ -48,7 +60,7 @@ namespace embree
|
||||
__forceinline operator unsigned char ( ) const { return 0; }
|
||||
};
|
||||
|
||||
const constexpr ZeroTy zero = ZeroTy();
|
||||
extern MAYBE_UNUSED ZeroTy zero;
|
||||
|
||||
struct OneTy
|
||||
{
|
||||
@@ -66,7 +78,7 @@ namespace embree
|
||||
__forceinline operator unsigned char ( ) const { return 1; }
|
||||
};
|
||||
|
||||
const constexpr OneTy one = OneTy();
|
||||
extern MAYBE_UNUSED OneTy one;
|
||||
|
||||
struct NegInfTy
|
||||
{
|
||||
@@ -85,7 +97,7 @@ namespace embree
|
||||
|
||||
};
|
||||
|
||||
const constexpr NegInfTy neg_inf = NegInfTy();
|
||||
extern MAYBE_UNUSED NegInfTy neg_inf;
|
||||
|
||||
struct PosInfTy
|
||||
{
|
||||
@@ -103,8 +115,8 @@ namespace embree
|
||||
__forceinline operator unsigned char ( ) const { return std::numeric_limits<unsigned char>::max(); }
|
||||
};
|
||||
|
||||
const constexpr PosInfTy inf = PosInfTy();
|
||||
const constexpr PosInfTy pos_inf = PosInfTy();
|
||||
extern MAYBE_UNUSED PosInfTy inf;
|
||||
extern MAYBE_UNUSED PosInfTy pos_inf;
|
||||
|
||||
struct NaNTy
|
||||
{
|
||||
@@ -112,15 +124,15 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return std::numeric_limits<float>::quiet_NaN(); }
|
||||
};
|
||||
|
||||
const constexpr NaNTy nan = NaNTy();
|
||||
extern MAYBE_UNUSED NaNTy nan;
|
||||
|
||||
struct UlpTy
|
||||
{
|
||||
__forceinline operator double( ) const { return std::numeric_limits<double>::epsilon(); }
|
||||
__forceinline operator float ( ) const { return std::numeric_limits<float>::epsilon(); }
|
||||
};
|
||||
|
||||
const constexpr UlpTy ulp = UlpTy();
|
||||
|
||||
extern MAYBE_UNUSED UlpTy ulp;
|
||||
|
||||
struct PiTy
|
||||
{
|
||||
@@ -128,7 +140,7 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(M_PI); }
|
||||
};
|
||||
|
||||
const constexpr PiTy pi = PiTy();
|
||||
extern MAYBE_UNUSED PiTy pi;
|
||||
|
||||
struct OneOverPiTy
|
||||
{
|
||||
@@ -136,7 +148,7 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(M_1_PI); }
|
||||
};
|
||||
|
||||
const constexpr OneOverPiTy one_over_pi = OneOverPiTy();
|
||||
extern MAYBE_UNUSED OneOverPiTy one_over_pi;
|
||||
|
||||
struct TwoPiTy
|
||||
{
|
||||
@@ -144,7 +156,7 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(2.0*M_PI); }
|
||||
};
|
||||
|
||||
const constexpr TwoPiTy two_pi = TwoPiTy();
|
||||
extern MAYBE_UNUSED TwoPiTy two_pi;
|
||||
|
||||
struct OneOverTwoPiTy
|
||||
{
|
||||
@@ -152,7 +164,7 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(0.5*M_1_PI); }
|
||||
};
|
||||
|
||||
const constexpr OneOverTwoPiTy one_over_two_pi = OneOverTwoPiTy();
|
||||
extern MAYBE_UNUSED OneOverTwoPiTy one_over_two_pi;
|
||||
|
||||
struct FourPiTy
|
||||
{
|
||||
@@ -160,7 +172,7 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(4.0*M_PI); }
|
||||
};
|
||||
|
||||
const constexpr FourPiTy four_pi = FourPiTy();
|
||||
extern MAYBE_UNUSED FourPiTy four_pi;
|
||||
|
||||
struct OneOverFourPiTy
|
||||
{
|
||||
@@ -168,42 +180,30 @@ namespace embree
|
||||
__forceinline operator float ( ) const { return float(0.25*M_1_PI); }
|
||||
};
|
||||
|
||||
const constexpr OneOverFourPiTy one_over_four_pi = OneOverFourPiTy();
|
||||
extern MAYBE_UNUSED OneOverFourPiTy one_over_four_pi;
|
||||
|
||||
struct StepTy {
|
||||
__forceinline operator double ( ) const { return 0; }
|
||||
__forceinline operator float ( ) const { return 0; }
|
||||
__forceinline operator long long( ) const { return 0; }
|
||||
__forceinline operator unsigned long long( ) const { return 0; }
|
||||
__forceinline operator long ( ) const { return 0; }
|
||||
__forceinline operator unsigned long ( ) const { return 0; }
|
||||
__forceinline operator int ( ) const { return 0; }
|
||||
__forceinline operator unsigned int ( ) const { return 0; }
|
||||
__forceinline operator short ( ) const { return 0; }
|
||||
__forceinline operator unsigned short ( ) const { return 0; }
|
||||
__forceinline operator char ( ) const { return 0; }
|
||||
__forceinline operator unsigned char ( ) const { return 0; }
|
||||
};
|
||||
|
||||
const constexpr StepTy step = StepTy();
|
||||
extern MAYBE_UNUSED StepTy step;
|
||||
|
||||
struct ReverseStepTy {
|
||||
};
|
||||
|
||||
const constexpr ReverseStepTy reverse_step = ReverseStepTy();
|
||||
extern MAYBE_UNUSED ReverseStepTy reverse_step;
|
||||
|
||||
struct EmptyTy {
|
||||
};
|
||||
|
||||
const constexpr EmptyTy empty = EmptyTy();
|
||||
extern MAYBE_UNUSED EmptyTy empty;
|
||||
|
||||
struct FullTy {
|
||||
};
|
||||
|
||||
const constexpr FullTy full = FullTy();
|
||||
extern MAYBE_UNUSED FullTy full;
|
||||
|
||||
struct UndefinedTy {
|
||||
};
|
||||
|
||||
const constexpr UndefinedTy undefined = UndefinedTy();
|
||||
extern MAYBE_UNUSED UndefinedTy undefined;
|
||||
}
|
||||
|
||||
+16
-34
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -71,7 +84,7 @@ namespace embree
|
||||
__forceinline friend const Interval intersect( const Interval& a, const Interval& b, const Interval& c ) { return intersect(a,intersect(b,c)); }
|
||||
__forceinline friend const Interval intersect( const Interval& a, const Interval& b, const Interval& c, const Interval& d ) { return intersect(intersect(a,b),intersect(c,d)); }
|
||||
|
||||
friend embree_ostream operator<<(embree_ostream cout, const Interval& a) {
|
||||
friend std::ostream& operator<<(std::ostream& cout, const Interval& a) {
|
||||
return cout << "[" << a.lower << ", " << a.upper << "]";
|
||||
}
|
||||
|
||||
@@ -127,35 +140,4 @@ namespace embree
|
||||
typedef Interval<float> Interval1f;
|
||||
typedef Vec2<Interval<float>> Interval2f;
|
||||
typedef Vec3<Interval<float>> Interval3f;
|
||||
|
||||
inline void swap(float& a, float& b) { float tmp = a; a = b; b = tmp; }
|
||||
|
||||
inline Interval1f shift(const Interval1f& v, float shift) { return Interval1f(v.lower + shift, v.upper + shift); }
|
||||
|
||||
#define TWO_PI (2.0*M_PI)
|
||||
inline Interval1f sin(Interval1f interval)
|
||||
{
|
||||
if (interval.upper-interval.lower >= M_PI) { return Interval1f(-1.0, 1.0); }
|
||||
if (interval.upper > TWO_PI) { interval = shift(interval, -TWO_PI*floor(interval.upper/TWO_PI)); }
|
||||
if (interval.lower < 0) { interval = shift(interval, -TWO_PI*floor(interval.lower/TWO_PI)); }
|
||||
float sinLower = sin(interval.lower);
|
||||
float sinUpper = sin(interval.upper);
|
||||
if (sinLower > sinUpper) swap(sinLower, sinUpper);
|
||||
if (interval.lower < M_PI / 2.0 && interval.upper > M_PI / 2.0) sinUpper = 1.0;
|
||||
if (interval.lower < 3.0 * M_PI / 2.0 && interval.upper > 3.0 * M_PI / 2.0) sinLower = -1.0;
|
||||
return Interval1f(sinLower, sinUpper);
|
||||
}
|
||||
|
||||
inline Interval1f cos(Interval1f interval)
|
||||
{
|
||||
if (interval.upper-interval.lower >= M_PI) { return Interval1f(-1.0, 1.0); }
|
||||
if (interval.upper > TWO_PI) { interval = shift(interval, -TWO_PI*floor(interval.upper/TWO_PI)); }
|
||||
if (interval.lower < 0) { interval = shift(interval, -TWO_PI*floor(interval.lower/TWO_PI)); }
|
||||
float cosLower = cos(interval.lower);
|
||||
float cosUpper = cos(interval.upper);
|
||||
if (cosLower > cosUpper) swap(cosLower, cosUpper);
|
||||
if (interval.lower < M_PI && interval.upper > M_PI) cosLower = -1.0;
|
||||
return Interval1f(cosLower, cosUpper);
|
||||
}
|
||||
#undef TWO_PI
|
||||
}
|
||||
|
||||
+22
-18
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -23,10 +36,9 @@ namespace embree
|
||||
public:
|
||||
__forceinline LBBox () {}
|
||||
|
||||
template<typename T1>
|
||||
__forceinline LBBox ( const LBBox<T1>& other )
|
||||
: bounds0(other.bounds0), bounds1(other.bounds1) {}
|
||||
|
||||
__forceinline LBBox ( const LBBox& other ) {
|
||||
bounds0 = other.bounds0; bounds1 = other.bounds1;
|
||||
}
|
||||
__forceinline LBBox& operator= ( const LBBox& other ) {
|
||||
bounds0 = other.bounds0; bounds1 = other.bounds1; return *this;
|
||||
}
|
||||
@@ -222,13 +234,11 @@ namespace embree
|
||||
}
|
||||
|
||||
/*! Comparison Operators */
|
||||
//template<typename TT> friend __forceinline bool operator==( const LBBox<TT>& a, const LBBox<TT>& b ) { return a.bounds0 == b.bounds0 && a.bounds1 == b.bounds1; }
|
||||
//template<typename TT> friend __forceinline bool operator!=( const LBBox<TT>& a, const LBBox<TT>& b ) { return a.bounds0 != b.bounds0 || a.bounds1 != b.bounds1; }
|
||||
friend __forceinline bool operator==( const LBBox& a, const LBBox& b ) { return a.bounds0 == b.bounds0 && a.bounds1 == b.bounds1; }
|
||||
friend __forceinline bool operator!=( const LBBox& a, const LBBox& b ) { return a.bounds0 != b.bounds0 || a.bounds1 != b.bounds1; }
|
||||
|
||||
template<typename TT> friend __forceinline bool operator==( const LBBox<TT>& a, const LBBox<TT>& b ) { return a.bounds0 == b.bounds0 && a.bounds1 == b.bounds1; }
|
||||
template<typename TT> friend __forceinline bool operator!=( const LBBox<TT>& a, const LBBox<TT>& b ) { return a.bounds0 != b.bounds0 || a.bounds1 != b.bounds1; }
|
||||
|
||||
/*! output operator */
|
||||
friend __forceinline embree_ostream operator<<(embree_ostream cout, const LBBox& box) {
|
||||
friend __forceinline std::ostream& operator<<(std::ostream& cout, const LBBox& box) {
|
||||
return cout << "LBBox { " << box.bounds0 << "; " << box.bounds1 << " }";
|
||||
}
|
||||
|
||||
@@ -241,11 +251,6 @@ namespace embree
|
||||
__forceinline bool isvalid( const LBBox<T>& v ) {
|
||||
return isvalid(v.bounds0) && isvalid(v.bounds1);
|
||||
}
|
||||
|
||||
template<typename T>
|
||||
__forceinline bool isvalid_non_empty( const LBBox<T>& v ) {
|
||||
return isvalid_non_empty(v.bounds0) && isvalid_non_empty(v.bounds1);
|
||||
}
|
||||
|
||||
template<typename T>
|
||||
__forceinline T expectedArea(const T& a0, const T& a1, const T& b0, const T& b1)
|
||||
@@ -285,5 +290,4 @@ namespace embree
|
||||
typedef LBBox<Vec2f> LBBox2f;
|
||||
typedef LBBox<Vec3f> LBBox3f;
|
||||
typedef LBBox<Vec3fa> LBBox3fa;
|
||||
typedef LBBox<Vec3fx> LBBox3fx;
|
||||
}
|
||||
|
||||
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -138,7 +151,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> static embree_ostream operator<<(embree_ostream cout, const LinearSpace2<T>& m) {
|
||||
template<typename T> static std::ostream& operator<<(std::ostream& cout, const LinearSpace2<T>& m) {
|
||||
return cout << "{ vx = " << m.vx << ", vy = " << m.vy << "}";
|
||||
}
|
||||
|
||||
|
||||
+16
-22
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -184,30 +197,11 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> static embree_ostream operator<<(embree_ostream cout, const LinearSpace3<T>& m) {
|
||||
template<typename T> static std::ostream& operator<<(std::ostream& cout, const LinearSpace3<T>& m) {
|
||||
return cout << "{ vx = " << m.vx << ", vy = " << m.vy << ", vz = " << m.vz << "}";
|
||||
}
|
||||
|
||||
/*! Shortcuts for common linear spaces. */
|
||||
typedef LinearSpace3<Vec3f> LinearSpace3f;
|
||||
typedef LinearSpace3<Vec3fa> LinearSpace3fa;
|
||||
typedef LinearSpace3<Vec3fx> LinearSpace3fx;
|
||||
typedef LinearSpace3<Vec3ff> LinearSpace3ff;
|
||||
|
||||
template<int N> using LinearSpace3vf = LinearSpace3<Vec3<vfloat<N>>>;
|
||||
typedef LinearSpace3<Vec3<vfloat<4>>> LinearSpace3vf4;
|
||||
typedef LinearSpace3<Vec3<vfloat<8>>> LinearSpace3vf8;
|
||||
typedef LinearSpace3<Vec3<vfloat<16>>> LinearSpace3vf16;
|
||||
|
||||
/*! blending */
|
||||
template<typename T, typename S>
|
||||
__forceinline LinearSpace3<T> lerp(const LinearSpace3<T>& l0,
|
||||
const LinearSpace3<T>& l1,
|
||||
const S& t)
|
||||
{
|
||||
return LinearSpace3<T>(lerp(l0.vx,l1.vx,t),
|
||||
lerp(l0.vy,l1.vy,t),
|
||||
lerp(l0.vz,l1.vz,t));
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
+24
-118
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -8,16 +21,12 @@
|
||||
#include "constants.h"
|
||||
#include <cmath>
|
||||
|
||||
#if defined(__ARM_NEON)
|
||||
#include "../simd/arm/emulation.h"
|
||||
#else
|
||||
#include <emmintrin.h>
|
||||
#include <xmmintrin.h>
|
||||
#include <immintrin.h>
|
||||
#endif
|
||||
|
||||
#if defined(__WIN32__)
|
||||
#if defined(_MSC_VER) && (_MSC_VER <= 1700)
|
||||
#if (__MSV_VER <= 1700)
|
||||
namespace std
|
||||
{
|
||||
__forceinline bool isinf ( const float x ) { return _finite(x) == 0; }
|
||||
@@ -41,9 +50,6 @@ namespace embree
|
||||
union { float f; int i; } v; v.i = i; return v.f;
|
||||
}
|
||||
|
||||
__forceinline int toInt (const float& a) { return int(a); }
|
||||
__forceinline float toFloat(const int& a) { return float(a); }
|
||||
|
||||
#if defined(__WIN32__)
|
||||
__forceinline bool finite ( const float x ) { return _finite(x) != 0; }
|
||||
#endif
|
||||
@@ -53,16 +59,6 @@ namespace embree
|
||||
|
||||
__forceinline float rcp ( const float x )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
// Move scalar to vector register and do rcp.
|
||||
__m128 a;
|
||||
a[0] = x;
|
||||
float32x4_t reciprocal = vrecpeq_f32(a);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a, reciprocal), reciprocal);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a, reciprocal), reciprocal);
|
||||
return reciprocal[0];
|
||||
#else
|
||||
|
||||
const __m128 a = _mm_set_ss(x);
|
||||
|
||||
#if defined(__AVX512VL__)
|
||||
@@ -76,74 +72,31 @@ namespace embree
|
||||
#else
|
||||
return _mm_cvtss_f32(_mm_mul_ss(r,_mm_sub_ss(_mm_set_ss(2.0f), _mm_mul_ss(r, a))));
|
||||
#endif
|
||||
|
||||
#endif //defined(__aarch64__)
|
||||
}
|
||||
|
||||
__forceinline float signmsk ( const float x ) {
|
||||
#if defined(__aarch64__)
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 a;
|
||||
__m128i b;
|
||||
a[0] = x;
|
||||
b[0] = 0x80000000;
|
||||
a = _mm_and_ps(a, vreinterpretq_f32_s32(b));
|
||||
return a[0];
|
||||
#else
|
||||
return _mm_cvtss_f32(_mm_and_ps(_mm_set_ss(x),_mm_castsi128_ps(_mm_set1_epi32(0x80000000))));
|
||||
#endif
|
||||
}
|
||||
__forceinline float xorf( const float x, const float y ) {
|
||||
#if defined(__aarch64__)
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 a;
|
||||
__m128 b;
|
||||
a[0] = x;
|
||||
b[0] = y;
|
||||
a = _mm_xor_ps(a, b);
|
||||
return a[0];
|
||||
#else
|
||||
return _mm_cvtss_f32(_mm_xor_ps(_mm_set_ss(x),_mm_set_ss(y)));
|
||||
#endif
|
||||
}
|
||||
__forceinline float andf( const float x, const unsigned y ) {
|
||||
#if defined(__aarch64__)
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 a;
|
||||
__m128i b;
|
||||
a[0] = x;
|
||||
b[0] = y;
|
||||
a = _mm_and_ps(a, vreinterpretq_f32_s32(b));
|
||||
return a[0];
|
||||
#else
|
||||
return _mm_cvtss_f32(_mm_and_ps(_mm_set_ss(x),_mm_castsi128_ps(_mm_set1_epi32(y))));
|
||||
#endif
|
||||
}
|
||||
__forceinline float rsqrt( const float x )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 a;
|
||||
a[0] = x;
|
||||
__m128 value = _mm_rsqrt_ps(a);
|
||||
value = vmulq_f32(value, vrsqrtsq_f32(vmulq_f32(a, value), value));
|
||||
value = vmulq_f32(value, vrsqrtsq_f32(vmulq_f32(a, value), value));
|
||||
return value[0];
|
||||
#else
|
||||
|
||||
const __m128 a = _mm_set_ss(x);
|
||||
#if defined(__AVX512VL__)
|
||||
__m128 r = _mm_rsqrt14_ss(_mm_set_ss(0.0f),a);
|
||||
const __m128 r = _mm_rsqrt14_ss(_mm_set_ss(0.0f),a);
|
||||
#else
|
||||
__m128 r = _mm_rsqrt_ss(a);
|
||||
const __m128 r = _mm_rsqrt_ss(a);
|
||||
#endif
|
||||
const __m128 c = _mm_add_ss(_mm_mul_ss(_mm_set_ss(1.5f), r),
|
||||
_mm_mul_ss(_mm_mul_ss(_mm_mul_ss(a, _mm_set_ss(-0.5f)), r), _mm_mul_ss(r, r)));
|
||||
return _mm_cvtss_f32(c);
|
||||
#endif
|
||||
}
|
||||
|
||||
#if defined(__WIN32__) && defined(_MSC_VER) && (_MSC_VER <= 1700)
|
||||
#if defined(__WIN32__) && (__MSC_VER <= 1700)
|
||||
__forceinline float nextafter(float x, float y) { if ((x<y) == (x>0)) return x*(1.1f+float(ulp)); else return x*(0.9f-float(ulp)); }
|
||||
__forceinline double nextafter(double x, double y) { return _nextafter(x, y); }
|
||||
__forceinline int roundf(float f) { return (int)(f + 0.5f); }
|
||||
@@ -197,17 +150,7 @@ namespace embree
|
||||
__forceinline double floor( const double x ) { return ::floor (x); }
|
||||
__forceinline double ceil ( const double x ) { return ::ceil (x); }
|
||||
|
||||
#if defined(__aarch64__)
|
||||
__forceinline float mini(float a, float b) {
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 x;
|
||||
__m128 y;
|
||||
x[0] = a;
|
||||
y[0] = b;
|
||||
x = _mm_min_ps(x, y);
|
||||
return x[0];
|
||||
}
|
||||
#elif defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline float mini(float a, float b) {
|
||||
const __m128i ai = _mm_castps_si128(_mm_set_ss(a));
|
||||
const __m128i bi = _mm_castps_si128(_mm_set_ss(b));
|
||||
@@ -216,17 +159,7 @@ namespace embree
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__aarch64__)
|
||||
__forceinline float maxi(float a, float b) {
|
||||
// FP and Neon shares same vector register in arm64
|
||||
__m128 x;
|
||||
__m128 y;
|
||||
x[0] = a;
|
||||
y[0] = b;
|
||||
x = _mm_max_ps(x, y);
|
||||
return x[0];
|
||||
}
|
||||
#elif defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline float maxi(float a, float b) {
|
||||
const __m128i ai = _mm_castps_si128(_mm_set_ss(a));
|
||||
const __m128i bi = _mm_castps_si128(_mm_set_ss(b));
|
||||
@@ -243,12 +176,9 @@ namespace embree
|
||||
__forceinline int64_t min(int64_t a, int64_t b) { return a<b ? a:b; }
|
||||
__forceinline float min(float a, float b) { return a<b ? a:b; }
|
||||
__forceinline double min(double a, double b) { return a<b ? a:b; }
|
||||
#if defined(__64BIT__) || defined(__EMSCRIPTEN__)
|
||||
#if defined(__X86_64__)
|
||||
__forceinline size_t min(size_t a, size_t b) { return a<b ? a:b; }
|
||||
#endif
|
||||
#if defined(__EMSCRIPTEN__)
|
||||
__forceinline long min(long a, long b) { return a<b ? a:b; }
|
||||
#endif
|
||||
|
||||
template<typename T> __forceinline T min(const T& a, const T& b, const T& c) { return min(min(a,b),c); }
|
||||
template<typename T> __forceinline T min(const T& a, const T& b, const T& c, const T& d) { return min(min(a,b),min(c,d)); }
|
||||
@@ -263,12 +193,9 @@ namespace embree
|
||||
__forceinline int64_t max(int64_t a, int64_t b) { return a<b ? b:a; }
|
||||
__forceinline float max(float a, float b) { return a<b ? b:a; }
|
||||
__forceinline double max(double a, double b) { return a<b ? b:a; }
|
||||
#if defined(__64BIT__) || defined(__EMSCRIPTEN__)
|
||||
#if defined(__X86_64__)
|
||||
__forceinline size_t max(size_t a, size_t b) { return a<b ? b:a; }
|
||||
#endif
|
||||
#if defined(__EMSCRIPTEN__)
|
||||
__forceinline long max(long a, long b) { return a<b ? b:a; }
|
||||
#endif
|
||||
|
||||
template<typename T> __forceinline T max(const T& a, const T& b, const T& c) { return max(max(a,b),c); }
|
||||
template<typename T> __forceinline T max(const T& a, const T& b, const T& c, const T& d) { return max(max(a,b),max(c,d)); }
|
||||
@@ -308,15 +235,6 @@ namespace embree
|
||||
__forceinline float msub ( const float a, const float b, const float c) { return _mm_cvtss_f32(_mm_fmsub_ss(_mm_set_ss(a),_mm_set_ss(b),_mm_set_ss(c))); }
|
||||
__forceinline float nmadd ( const float a, const float b, const float c) { return _mm_cvtss_f32(_mm_fnmadd_ss(_mm_set_ss(a),_mm_set_ss(b),_mm_set_ss(c))); }
|
||||
__forceinline float nmsub ( const float a, const float b, const float c) { return _mm_cvtss_f32(_mm_fnmsub_ss(_mm_set_ss(a),_mm_set_ss(b),_mm_set_ss(c))); }
|
||||
|
||||
#elif defined (__aarch64__) && defined(__clang__)
|
||||
#pragma clang fp contract(fast)
|
||||
__forceinline float madd ( const float a, const float b, const float c) { return a*b + c; }
|
||||
__forceinline float msub ( const float a, const float b, const float c) { return a*b - c; }
|
||||
__forceinline float nmadd ( const float a, const float b, const float c) { return c - a*b; }
|
||||
__forceinline float nmsub ( const float a, const float b, const float c) { return -(c + a*b); }
|
||||
#pragma clang fp contract(on)
|
||||
|
||||
#else
|
||||
__forceinline float madd ( const float a, const float b, const float c) { return a*b+c; }
|
||||
__forceinline float msub ( const float a, const float b, const float c) { return a*b-c; }
|
||||
@@ -351,7 +269,6 @@ __forceinline float nmsub ( const float a, const float b, const float c) { retur
|
||||
__forceinline int select(bool s, int t, int f) { return s ? t : f; }
|
||||
__forceinline float select(bool s, float t, float f) { return s ? t : f; }
|
||||
|
||||
__forceinline bool all(bool s) { return s; }
|
||||
|
||||
__forceinline float lerp(const float v0, const float v1, const float t) {
|
||||
return madd(1.0f-t,v0,t*v1);
|
||||
@@ -365,17 +282,6 @@ __forceinline float nmsub ( const float a, const float b, const float c) { retur
|
||||
/*! exchange */
|
||||
template<typename T> __forceinline void xchg ( T& a, T& b ) { const T tmp = a; a = b; b = tmp; }
|
||||
|
||||
/* load/store */
|
||||
template<typename Ty> struct mem;
|
||||
|
||||
template<> struct mem<float> {
|
||||
static __forceinline float load (bool mask, const void* ptr) { return mask ? *(float*)ptr : 0.0f; }
|
||||
static __forceinline float loadu(bool mask, const void* ptr) { return mask ? *(float*)ptr : 0.0f; }
|
||||
|
||||
static __forceinline void store (bool mask, void* ptr, const float v) { if (mask) *(float*)ptr = v; }
|
||||
static __forceinline void storeu(bool mask, void* ptr, const float v) { if (mask) *(float*)ptr = v; }
|
||||
};
|
||||
|
||||
/*! bit reverse operation */
|
||||
template<class T>
|
||||
__forceinline T bitReverse(const T& vin)
|
||||
@@ -412,7 +318,7 @@ __forceinline float nmsub ( const float a, const float b, const float c) { retur
|
||||
return x | (y << 1) | (z << 2);
|
||||
}
|
||||
|
||||
#if defined(__AVX2__) && !defined(__aarch64__)
|
||||
#if defined(__AVX2__)
|
||||
|
||||
template<>
|
||||
__forceinline unsigned int bitInterleave(const unsigned int &xi, const unsigned int& yi, const unsigned int& zi)
|
||||
|
||||
+17
-4
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -25,8 +38,8 @@ namespace embree
|
||||
__forceinline OBBox (const LinearSpace3<T>& space, const BBox<T>& bounds)
|
||||
: space(space), bounds(bounds) {}
|
||||
|
||||
friend embree_ostream operator<<(embree_ostream cout, const OBBox& p) {
|
||||
return cout << "{ space = " << p.space << ", bounds = " << p.bounds << "}";
|
||||
friend std::ostream& operator<<(std::ostream& cout, const OBBox& p) {
|
||||
return std::cout << "{ space = " << p.space << ", bounds = " << p.bounds << "}";
|
||||
}
|
||||
|
||||
public:
|
||||
|
||||
+18
-83
@@ -1,12 +1,22 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "vec3.h"
|
||||
#include "vec4.h"
|
||||
|
||||
#include "transcendental.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
@@ -29,7 +39,6 @@ namespace embree
|
||||
|
||||
__forceinline QuaternionT( const T& r ) : r(r), i(zero), j(zero), k(zero) {}
|
||||
__forceinline explicit QuaternionT( const Vec3<T>& v ) : r(zero), i(v.x), j(v.y), k(v.z) {}
|
||||
__forceinline explicit QuaternionT( const Vec4<T>& v ) : r(v.x), i(v.y), j(v.z), k(v.w) {}
|
||||
__forceinline QuaternionT( const T& r, const T& i, const T& j, const T& k ) : r(r), i(i), j(j), k(k) {}
|
||||
__forceinline QuaternionT( const T& r, const Vec3<T>& v ) : r(r), i(v.x), j(v.y), k(v.z) {}
|
||||
|
||||
@@ -69,25 +78,6 @@ namespace embree
|
||||
template<typename T> __forceinline QuaternionT<T> rcp ( const QuaternionT<T>& a ) { return conj(a)*rcp(a.r*a.r + a.i*a.i + a.j*a.j + a.k*a.k); }
|
||||
template<typename T> __forceinline QuaternionT<T> normalize ( const QuaternionT<T>& a ) { return a*rsqrt(a.r*a.r + a.i*a.i + a.j*a.j + a.k*a.k); }
|
||||
|
||||
// evaluates a*q-r
|
||||
template<typename T> __forceinline QuaternionT<T>
|
||||
msub(const T& a, const QuaternionT<T>& q, const QuaternionT<T>& p)
|
||||
{
|
||||
return QuaternionT<T>(msub(a, q.r, p.r),
|
||||
msub(a, q.i, p.i),
|
||||
msub(a, q.j, p.j),
|
||||
msub(a, q.k, p.k));
|
||||
}
|
||||
// evaluates a*q-r
|
||||
template<typename T> __forceinline QuaternionT<T>
|
||||
madd (const T& a, const QuaternionT<T>& q, const QuaternionT<T>& p)
|
||||
{
|
||||
return QuaternionT<T>(madd(a, q.r, p.r),
|
||||
madd(a, q.i, p.i),
|
||||
madd(a, q.j, p.j),
|
||||
madd(a, q.k, p.k));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////
|
||||
// Binary Operators
|
||||
////////////////////////////////////////////////////////////////
|
||||
@@ -119,22 +109,10 @@ namespace embree
|
||||
template<typename T> __forceinline QuaternionT<T>& operator /=( QuaternionT<T>& a, const T & b ) { return a = a*rcp(b); }
|
||||
template<typename T> __forceinline QuaternionT<T>& operator /=( QuaternionT<T>& a, const QuaternionT<T>& b ) { return a = a*rcp(b); }
|
||||
|
||||
template<typename T, typename M> __forceinline QuaternionT<T>
|
||||
select(const M& m, const QuaternionT<T>& q, const QuaternionT<T>& p)
|
||||
{
|
||||
return QuaternionT<T>(select(m, q.r, p.r),
|
||||
select(m, q.i, p.i),
|
||||
select(m, q.j, p.j),
|
||||
select(m, q.k, p.k));
|
||||
}
|
||||
|
||||
|
||||
template<typename T> __forceinline Vec3<T> xfmPoint ( const QuaternionT<T>& a, const Vec3<T>& b ) { return (a*QuaternionT<T>(b)*conj(a)).v(); }
|
||||
template<typename T> __forceinline Vec3<T> xfmVector( const QuaternionT<T>& a, const Vec3<T>& b ) { return (a*QuaternionT<T>(b)*conj(a)).v(); }
|
||||
template<typename T> __forceinline Vec3<T> xfmNormal( const QuaternionT<T>& a, const Vec3<T>& b ) { return (a*QuaternionT<T>(b)*conj(a)).v(); }
|
||||
|
||||
template<typename T> __forceinline T dot(const QuaternionT<T>& a, const QuaternionT<T>& b) { return a.r*b.r + a.i*b.i + a.j*b.j + a.k*b.k; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Comparison Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -201,58 +179,15 @@ namespace embree
|
||||
k = sro*cya*cpi - cro*sya*spi;
|
||||
}
|
||||
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Output Operators
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> static embree_ostream operator<<(embree_ostream cout, const QuaternionT<T>& q) {
|
||||
template<typename T> static std::ostream& operator<<(std::ostream& cout, const QuaternionT<T>& q) {
|
||||
return cout << "{ r = " << q.r << ", i = " << q.i << ", j = " << q.j << ", k = " << q.k << " }";
|
||||
}
|
||||
|
||||
/*! default template instantiations */
|
||||
typedef QuaternionT<float> Quaternion3f;
|
||||
typedef QuaternionT<double> Quaternion3d;
|
||||
|
||||
template<int N> using Quaternion3vf = QuaternionT<vfloat<N>>;
|
||||
typedef QuaternionT<vfloat<4>> Quaternion3vf4;
|
||||
typedef QuaternionT<vfloat<8>> Quaternion3vf8;
|
||||
typedef QuaternionT<vfloat<16>> Quaternion3vf16;
|
||||
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
/// Interpolation
|
||||
//////////////////////////////////////////////////////////////////////////////
|
||||
template<typename T>
|
||||
__forceinline QuaternionT<T>lerp(const QuaternionT<T>& q0,
|
||||
const QuaternionT<T>& q1,
|
||||
const T& factor)
|
||||
{
|
||||
QuaternionT<T> q;
|
||||
q.r = lerp(q0.r, q1.r, factor);
|
||||
q.i = lerp(q0.i, q1.i, factor);
|
||||
q.j = lerp(q0.j, q1.j, factor);
|
||||
q.k = lerp(q0.k, q1.k, factor);
|
||||
return q;
|
||||
}
|
||||
|
||||
template<typename T>
|
||||
__forceinline QuaternionT<T> slerp(const QuaternionT<T>& q0,
|
||||
const QuaternionT<T>& q1_,
|
||||
const T& t)
|
||||
{
|
||||
T cosTheta = dot(q0, q1_);
|
||||
QuaternionT<T> q1 = select(cosTheta < 0.f, -q1_, q1_);
|
||||
cosTheta = select(cosTheta < 0.f, -cosTheta, cosTheta);
|
||||
|
||||
// spherical linear interpolation
|
||||
const T phi = t * fastapprox::acos(cosTheta);
|
||||
T sinPhi, cosPhi;
|
||||
fastapprox::sincos(phi, sinPhi, cosPhi);
|
||||
QuaternionT<T> qperp = sinPhi * normalize(msub(cosTheta, q0, q1));
|
||||
QuaternionT<T> qslerp = msub(cosPhi, q0, qperp);
|
||||
|
||||
// regular linear interpolation as fallback
|
||||
QuaternionT<T> qlerp = normalize(lerp(q0, q1, t));
|
||||
|
||||
return select(cosTheta > 0.9995f, qlerp, qslerp);
|
||||
}
|
||||
}
|
||||
|
||||
+17
-4
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -74,7 +87,7 @@ namespace embree
|
||||
return r0.size() < r1.size();
|
||||
}
|
||||
|
||||
friend embree_ostream operator<<(embree_ostream cout, const range& r) {
|
||||
friend std::ostream& operator<<(std::ostream& cout, const range& r) {
|
||||
return cout << "range [" << r.begin() << ", " << r.end() << "]";
|
||||
}
|
||||
|
||||
@@ -128,7 +141,7 @@ namespace embree
|
||||
_ext_end += plus;
|
||||
}
|
||||
|
||||
friend embree_ostream operator<<(embree_ostream cout, const extended_range& r) {
|
||||
friend std::ostream& operator<<(std::ostream& cout, const extended_range& r) {
|
||||
return cout << "extended_range [" << r.begin() << ", " << r.end() << " (" << r.ext_end() << ")]";
|
||||
}
|
||||
|
||||
|
||||
@@ -1,525 +0,0 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#pragma once
|
||||
|
||||
// Transcendental functions from "ispc": https://github.com/ispc/ispc/
|
||||
// Most of the transcendental implementations in ispc code come from
|
||||
// Solomon Boulos's "syrah": https://github.com/boulos/syrah/
|
||||
|
||||
#include "../simd/simd.h"
|
||||
|
||||
namespace embree
|
||||
{
|
||||
|
||||
namespace fastapprox
|
||||
{
|
||||
|
||||
template <typename T>
|
||||
__forceinline T sin(const T &v)
|
||||
{
|
||||
static const float piOverTwoVec = 1.57079637050628662109375;
|
||||
static const float twoOverPiVec = 0.636619746685028076171875;
|
||||
auto scaled = v * twoOverPiVec;
|
||||
auto kReal = floor(scaled);
|
||||
auto k = toInt(kReal);
|
||||
|
||||
// Reduced range version of x
|
||||
auto x = v - kReal * piOverTwoVec;
|
||||
auto kMod4 = k & 3;
|
||||
auto sinUseCos = (kMod4 == 1) | (kMod4 == 3);
|
||||
auto flipSign = (kMod4 > 1);
|
||||
|
||||
// These coefficients are from sollya with fpminimax(sin(x)/x, [|0, 2,
|
||||
// 4, 6, 8, 10|], [|single...|], [0;Pi/2]);
|
||||
static const float sinC2 = -0.16666667163372039794921875;
|
||||
static const float sinC4 = +8.333347737789154052734375e-3;
|
||||
static const float sinC6 = -1.9842604524455964565277099609375e-4;
|
||||
static const float sinC8 = +2.760012648650445044040679931640625e-6;
|
||||
static const float sinC10 = -2.50293279435709337121807038784027099609375e-8;
|
||||
|
||||
static const float cosC2 = -0.5;
|
||||
static const float cosC4 = +4.166664183139801025390625e-2;
|
||||
static const float cosC6 = -1.388833043165504932403564453125e-3;
|
||||
static const float cosC8 = +2.47562347794882953166961669921875e-5;
|
||||
static const float cosC10 = -2.59630184018533327616751194000244140625e-7;
|
||||
|
||||
auto outside = select(sinUseCos, 1., x);
|
||||
auto c2 = select(sinUseCos, T(cosC2), T(sinC2));
|
||||
auto c4 = select(sinUseCos, T(cosC4), T(sinC4));
|
||||
auto c6 = select(sinUseCos, T(cosC6), T(sinC6));
|
||||
auto c8 = select(sinUseCos, T(cosC8), T(sinC8));
|
||||
auto c10 = select(sinUseCos, T(cosC10), T(sinC10));
|
||||
|
||||
auto x2 = x * x;
|
||||
auto formula = x2 * c10 + c8;
|
||||
formula = x2 * formula + c6;
|
||||
formula = x2 * formula + c4;
|
||||
formula = x2 * formula + c2;
|
||||
formula = x2 * formula + 1.;
|
||||
formula *= outside;
|
||||
|
||||
formula = select(flipSign, -formula, formula);
|
||||
return formula;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T cos(const T &v)
|
||||
{
|
||||
static const float piOverTwoVec = 1.57079637050628662109375;
|
||||
static const float twoOverPiVec = 0.636619746685028076171875;
|
||||
auto scaled = v * twoOverPiVec;
|
||||
auto kReal = floor(scaled);
|
||||
auto k = toInt(kReal);
|
||||
|
||||
// Reduced range version of x
|
||||
auto x = v - kReal * piOverTwoVec;
|
||||
|
||||
auto kMod4 = k & 3;
|
||||
auto cosUseCos = (kMod4 == 0) | (kMod4 == 2);
|
||||
auto flipSign = (kMod4 == 1) | (kMod4 == 2);
|
||||
|
||||
const float sinC2 = -0.16666667163372039794921875;
|
||||
const float sinC4 = +8.333347737789154052734375e-3;
|
||||
const float sinC6 = -1.9842604524455964565277099609375e-4;
|
||||
const float sinC8 = +2.760012648650445044040679931640625e-6;
|
||||
const float sinC10 = -2.50293279435709337121807038784027099609375e-8;
|
||||
|
||||
const float cosC2 = -0.5;
|
||||
const float cosC4 = +4.166664183139801025390625e-2;
|
||||
const float cosC6 = -1.388833043165504932403564453125e-3;
|
||||
const float cosC8 = +2.47562347794882953166961669921875e-5;
|
||||
const float cosC10 = -2.59630184018533327616751194000244140625e-7;
|
||||
|
||||
auto outside = select(cosUseCos, 1., x);
|
||||
auto c2 = select(cosUseCos, T(cosC2), T(sinC2));
|
||||
auto c4 = select(cosUseCos, T(cosC4), T(sinC4));
|
||||
auto c6 = select(cosUseCos, T(cosC6), T(sinC6));
|
||||
auto c8 = select(cosUseCos, T(cosC8), T(sinC8));
|
||||
auto c10 = select(cosUseCos, T(cosC10), T(sinC10));
|
||||
|
||||
auto x2 = x * x;
|
||||
auto formula = x2 * c10 + c8;
|
||||
formula = x2 * formula + c6;
|
||||
formula = x2 * formula + c4;
|
||||
formula = x2 * formula + c2;
|
||||
formula = x2 * formula + 1.;
|
||||
formula *= outside;
|
||||
|
||||
formula = select(flipSign, -formula, formula);
|
||||
return formula;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline void sincos(const T &v, T &sinResult, T &cosResult)
|
||||
{
|
||||
const float piOverTwoVec = 1.57079637050628662109375;
|
||||
const float twoOverPiVec = 0.636619746685028076171875;
|
||||
auto scaled = v * twoOverPiVec;
|
||||
auto kReal = floor(scaled);
|
||||
auto k = toInt(kReal);
|
||||
|
||||
// Reduced range version of x
|
||||
auto x = v - kReal * piOverTwoVec;
|
||||
auto kMod4 = k & 3;
|
||||
auto cosUseCos = ((kMod4 == 0) | (kMod4 == 2));
|
||||
auto sinUseCos = ((kMod4 == 1) | (kMod4 == 3));
|
||||
auto sinFlipSign = (kMod4 > 1);
|
||||
auto cosFlipSign = ((kMod4 == 1) | (kMod4 == 2));
|
||||
|
||||
const float oneVec = +1.;
|
||||
const float sinC2 = -0.16666667163372039794921875;
|
||||
const float sinC4 = +8.333347737789154052734375e-3;
|
||||
const float sinC6 = -1.9842604524455964565277099609375e-4;
|
||||
const float sinC8 = +2.760012648650445044040679931640625e-6;
|
||||
const float sinC10 = -2.50293279435709337121807038784027099609375e-8;
|
||||
|
||||
const float cosC2 = -0.5;
|
||||
const float cosC4 = +4.166664183139801025390625e-2;
|
||||
const float cosC6 = -1.388833043165504932403564453125e-3;
|
||||
const float cosC8 = +2.47562347794882953166961669921875e-5;
|
||||
const float cosC10 = -2.59630184018533327616751194000244140625e-7;
|
||||
|
||||
auto x2 = x * x;
|
||||
|
||||
auto sinFormula = x2 * sinC10 + sinC8;
|
||||
auto cosFormula = x2 * cosC10 + cosC8;
|
||||
sinFormula = x2 * sinFormula + sinC6;
|
||||
cosFormula = x2 * cosFormula + cosC6;
|
||||
|
||||
sinFormula = x2 * sinFormula + sinC4;
|
||||
cosFormula = x2 * cosFormula + cosC4;
|
||||
|
||||
sinFormula = x2 * sinFormula + sinC2;
|
||||
cosFormula = x2 * cosFormula + cosC2;
|
||||
|
||||
sinFormula = x2 * sinFormula + oneVec;
|
||||
cosFormula = x2 * cosFormula + oneVec;
|
||||
|
||||
sinFormula *= x;
|
||||
|
||||
sinResult = select(sinUseCos, cosFormula, sinFormula);
|
||||
cosResult = select(cosUseCos, cosFormula, sinFormula);
|
||||
|
||||
sinResult = select(sinFlipSign, -sinResult, sinResult);
|
||||
cosResult = select(cosFlipSign, -cosResult, cosResult);
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T tan(const T &v)
|
||||
{
|
||||
const float piOverFourVec = 0.785398185253143310546875;
|
||||
const float fourOverPiVec = 1.27323949337005615234375;
|
||||
|
||||
auto xLt0 = v < 0.;
|
||||
auto y = select(xLt0, -v, v);
|
||||
auto scaled = y * fourOverPiVec;
|
||||
|
||||
auto kReal = floor(scaled);
|
||||
auto k = toInt(kReal);
|
||||
|
||||
auto x = y - kReal * piOverFourVec;
|
||||
|
||||
// If k & 1, x -= Pi/4
|
||||
auto needOffset = (k & 1) != 0;
|
||||
x = select(needOffset, x - piOverFourVec, x);
|
||||
|
||||
// If k & 3 == (0 or 3) let z = tan_In...(y) otherwise z = -cot_In0To...
|
||||
auto kMod4 = k & 3;
|
||||
auto useCotan = (kMod4 == 1) | (kMod4 == 2);
|
||||
|
||||
const float oneVec = 1.0;
|
||||
|
||||
const float tanC2 = +0.33333075046539306640625;
|
||||
const float tanC4 = +0.13339905440807342529296875;
|
||||
const float tanC6 = +5.3348250687122344970703125e-2;
|
||||
const float tanC8 = +2.46033705770969390869140625e-2;
|
||||
const float tanC10 = +2.892402000725269317626953125e-3;
|
||||
const float tanC12 = +9.500005282461643218994140625e-3;
|
||||
|
||||
const float cotC2 = -0.3333333432674407958984375;
|
||||
const float cotC4 = -2.222204394638538360595703125e-2;
|
||||
const float cotC6 = -2.11752182804048061370849609375e-3;
|
||||
const float cotC8 = -2.0846328698098659515380859375e-4;
|
||||
const float cotC10 = -2.548247357481159269809722900390625e-5;
|
||||
const float cotC12 = -3.5257363606433500535786151885986328125e-7;
|
||||
|
||||
auto x2 = x * x;
|
||||
T z;
|
||||
if (any(useCotan))
|
||||
{
|
||||
auto cotVal = x2 * cotC12 + cotC10;
|
||||
cotVal = x2 * cotVal + cotC8;
|
||||
cotVal = x2 * cotVal + cotC6;
|
||||
cotVal = x2 * cotVal + cotC4;
|
||||
cotVal = x2 * cotVal + cotC2;
|
||||
cotVal = x2 * cotVal + oneVec;
|
||||
// The equation is for x * cot(x) but we need -x * cot(x) for the tan part.
|
||||
cotVal /= -x;
|
||||
z = cotVal;
|
||||
}
|
||||
auto useTan = !useCotan;
|
||||
if (any(useTan))
|
||||
{
|
||||
auto tanVal = x2 * tanC12 + tanC10;
|
||||
tanVal = x2 * tanVal + tanC8;
|
||||
tanVal = x2 * tanVal + tanC6;
|
||||
tanVal = x2 * tanVal + tanC4;
|
||||
tanVal = x2 * tanVal + tanC2;
|
||||
tanVal = x2 * tanVal + oneVec;
|
||||
// Equation was for tan(x)/x
|
||||
tanVal *= x;
|
||||
z = select(useTan, tanVal, z);
|
||||
}
|
||||
return select(xLt0, -z, z);
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T asin(const T &x0)
|
||||
{
|
||||
auto isneg = (x0 < 0.f);
|
||||
auto x = abs(x0);
|
||||
auto isnan = (x > 1.f);
|
||||
|
||||
// sollya
|
||||
// fpminimax(((asin(x)-pi/2)/-sqrt(1-x)), [|0,1,2,3,4,5|],[|single...|],
|
||||
// [1e-20;.9999999999999999]);
|
||||
// avg error: 1.1105439e-06, max error 1.3187528e-06
|
||||
auto v = 1.57079517841339111328125f +
|
||||
x * (-0.21450997889041900634765625f +
|
||||
x * (8.78556668758392333984375e-2f +
|
||||
x * (-4.489909112453460693359375e-2f +
|
||||
x * (1.928029954433441162109375e-2f +
|
||||
x * (-4.3095736764371395111083984375e-3f)))));
|
||||
|
||||
v *= -sqrt(1.f - x);
|
||||
v = v + 1.57079637050628662109375f;
|
||||
|
||||
v = select(v < 0.f, T(0.f), v);
|
||||
v = select(isneg, -v, v);
|
||||
v = select(isnan, T(cast_i2f(0x7fc00000)), v);
|
||||
|
||||
return v;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T acos(const T &v)
|
||||
{
|
||||
return 1.57079637050628662109375f - asin(v);
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T atan(const T &v)
|
||||
{
|
||||
const float piOverTwoVec = 1.57079637050628662109375;
|
||||
// atan(-x) = -atan(x) (so flip from negative to positive first)
|
||||
// If x > 1 -> atan(x) = Pi/2 - atan(1/x)
|
||||
auto xNeg = v < 0.f;
|
||||
auto xFlipped = select(xNeg, -v, v);
|
||||
|
||||
auto xGt1 = xFlipped > 1.;
|
||||
auto x = select(xGt1, rcpSafe(xFlipped), xFlipped);
|
||||
|
||||
// These coefficients approximate atan(x)/x
|
||||
const float atanC0 = +0.99999988079071044921875;
|
||||
const float atanC2 = -0.3333191573619842529296875;
|
||||
const float atanC4 = +0.199689209461212158203125;
|
||||
const float atanC6 = -0.14015688002109527587890625;
|
||||
const float atanC8 = +9.905083477497100830078125e-2;
|
||||
const float atanC10 = -5.93664981424808502197265625e-2;
|
||||
const float atanC12 = +2.417283318936824798583984375e-2;
|
||||
const float atanC14 = -4.6721356920897960662841796875e-3;
|
||||
|
||||
auto x2 = x * x;
|
||||
auto result = x2 * atanC14 + atanC12;
|
||||
result = x2 * result + atanC10;
|
||||
result = x2 * result + atanC8;
|
||||
result = x2 * result + atanC6;
|
||||
result = x2 * result + atanC4;
|
||||
result = x2 * result + atanC2;
|
||||
result = x2 * result + atanC0;
|
||||
result *= x;
|
||||
|
||||
result = select(xGt1, piOverTwoVec - result, result);
|
||||
result = select(xNeg, -result, result);
|
||||
return result;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T atan2(const T &y, const T &x)
|
||||
{
|
||||
const float piVec = 3.1415926536;
|
||||
// atan2(y, x) =
|
||||
//
|
||||
// atan2(y > 0, x = +-0) -> Pi/2
|
||||
// atan2(y < 0, x = +-0) -> -Pi/2
|
||||
// atan2(y = +-0, x < +0) -> +-Pi
|
||||
// atan2(y = +-0, x >= +0) -> +-0
|
||||
//
|
||||
// atan2(y >= 0, x < 0) -> Pi + atan(y/x)
|
||||
// atan2(y < 0, x < 0) -> -Pi + atan(y/x)
|
||||
// atan2(y, x > 0) -> atan(y/x)
|
||||
//
|
||||
// and then a bunch of code for dealing with infinities.
|
||||
auto yOverX = y * rcpSafe(x);
|
||||
auto atanArg = atan(yOverX);
|
||||
auto xLt0 = x < 0.f;
|
||||
auto yLt0 = y < 0.f;
|
||||
auto offset = select(xLt0,
|
||||
select(yLt0, T(-piVec), T(piVec)), 0.f);
|
||||
return offset + atanArg;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T exp(const T &v)
|
||||
{
|
||||
const float ln2Part1 = 0.6931457519;
|
||||
const float ln2Part2 = 1.4286067653e-6;
|
||||
const float oneOverLn2 = 1.44269502162933349609375;
|
||||
|
||||
auto scaled = v * oneOverLn2;
|
||||
auto kReal = floor(scaled);
|
||||
auto k = toInt(kReal);
|
||||
|
||||
// Reduced range version of x
|
||||
auto x = v - kReal * ln2Part1;
|
||||
x -= kReal * ln2Part2;
|
||||
|
||||
// These coefficients are for e^x in [0, ln(2)]
|
||||
const float one = 1.;
|
||||
const float c2 = 0.4999999105930328369140625;
|
||||
const float c3 = 0.166668415069580078125;
|
||||
const float c4 = 4.16539050638675689697265625e-2;
|
||||
const float c5 = 8.378830738365650177001953125e-3;
|
||||
const float c6 = 1.304379315115511417388916015625e-3;
|
||||
const float c7 = 2.7555381529964506626129150390625e-4;
|
||||
|
||||
auto result = x * c7 + c6;
|
||||
result = x * result + c5;
|
||||
result = x * result + c4;
|
||||
result = x * result + c3;
|
||||
result = x * result + c2;
|
||||
result = x * result + one;
|
||||
result = x * result + one;
|
||||
|
||||
// Compute 2^k (should differ for float and double, but I'll avoid
|
||||
// it for now and just do floats)
|
||||
const int fpbias = 127;
|
||||
auto biasedN = k + fpbias;
|
||||
auto overflow = kReal > fpbias;
|
||||
// Minimum exponent is -126, so if k is <= -127 (k + 127 <= 0)
|
||||
// we've got underflow. -127 * ln(2) -> -88.02. So the most
|
||||
// negative float input that doesn't result in zero is like -88.
|
||||
auto underflow = kReal <= -fpbias;
|
||||
const int infBits = 0x7f800000;
|
||||
biasedN <<= 23;
|
||||
// Reinterpret this thing as float
|
||||
auto twoToTheN = asFloat(biasedN);
|
||||
// Handle both doubles and floats (hopefully eliding the copy for float)
|
||||
auto elemtype2n = twoToTheN;
|
||||
result *= elemtype2n;
|
||||
result = select(overflow, cast_i2f(infBits), result);
|
||||
result = select(underflow, 0., result);
|
||||
return result;
|
||||
}
|
||||
|
||||
// Range reduction for logarithms takes log(x) -> log(2^n * y) -> n
|
||||
// * log(2) + log(y) where y is the reduced range (usually in [1/2, 1)).
|
||||
template <typename T, typename R>
|
||||
__forceinline void __rangeReduceLog(const T &input,
|
||||
T &reduced,
|
||||
R &exponent)
|
||||
{
|
||||
auto intVersion = asInt(input);
|
||||
// single precision = SEEE EEEE EMMM MMMM MMMM MMMM MMMM MMMM
|
||||
// exponent mask = 0111 1111 1000 0000 0000 0000 0000 0000
|
||||
// 0x7 0xF 0x8 0x0 0x0 0x0 0x0 0x0
|
||||
// non-exponent = 1000 0000 0111 1111 1111 1111 1111 1111
|
||||
// = 0x8 0x0 0x7 0xF 0xF 0xF 0xF 0xF
|
||||
|
||||
//const int exponentMask(0x7F800000)
|
||||
static const int nonexponentMask = 0x807FFFFF;
|
||||
|
||||
// We want the reduced version to have an exponent of -1 which is
|
||||
// -1 + 127 after biasing or 126
|
||||
static const int exponentNeg1 = (126l << 23);
|
||||
// NOTE(boulos): We don't need to mask anything out since we know
|
||||
// the sign bit has to be 0. If it's 1, we need to return infinity/nan
|
||||
// anyway (log(x), x = +-0 -> infinity, x < 0 -> NaN).
|
||||
auto biasedExponent = intVersion >> 23; // This number is [0, 255] but it means [-127, 128]
|
||||
|
||||
auto offsetExponent = biasedExponent + 1; // Treat the number as if it were 2^{e+1} * (1.m)/2
|
||||
exponent = offsetExponent - 127; // get the real value
|
||||
|
||||
// Blend the offset_exponent with the original input (do this in
|
||||
// int for now, until I decide if float can have & and ¬)
|
||||
auto blended = (intVersion & nonexponentMask) | (exponentNeg1);
|
||||
reduced = asFloat(blended);
|
||||
}
|
||||
|
||||
template <typename T> struct ExponentType { };
|
||||
template <int N> struct ExponentType<vfloat_impl<N>> { typedef vint<N> Ty; };
|
||||
template <> struct ExponentType<float> { typedef int Ty; };
|
||||
|
||||
template <typename T>
|
||||
__forceinline T log(const T &v)
|
||||
{
|
||||
T reduced;
|
||||
typename ExponentType<T>::Ty exponent;
|
||||
|
||||
const int nanBits = 0x7fc00000;
|
||||
const int negInfBits = 0xFF800000;
|
||||
const float nan = cast_i2f(nanBits);
|
||||
const float negInf = cast_i2f(negInfBits);
|
||||
auto useNan = v < 0.;
|
||||
auto useInf = v == 0.;
|
||||
auto exceptional = useNan | useInf;
|
||||
const float one = 1.0;
|
||||
|
||||
auto patched = select(exceptional, one, v);
|
||||
__rangeReduceLog(patched, reduced, exponent);
|
||||
|
||||
const float ln2 = 0.693147182464599609375;
|
||||
|
||||
auto x1 = one - reduced;
|
||||
const float c1 = +0.50000095367431640625;
|
||||
const float c2 = +0.33326041698455810546875;
|
||||
const float c3 = +0.2519190013408660888671875;
|
||||
const float c4 = +0.17541764676570892333984375;
|
||||
const float c5 = +0.3424419462680816650390625;
|
||||
const float c6 = -0.599632322788238525390625;
|
||||
const float c7 = +1.98442304134368896484375;
|
||||
const float c8 = -2.4899270534515380859375;
|
||||
const float c9 = +1.7491014003753662109375;
|
||||
|
||||
auto result = x1 * c9 + c8;
|
||||
result = x1 * result + c7;
|
||||
result = x1 * result + c6;
|
||||
result = x1 * result + c5;
|
||||
result = x1 * result + c4;
|
||||
result = x1 * result + c3;
|
||||
result = x1 * result + c2;
|
||||
result = x1 * result + c1;
|
||||
result = x1 * result + one;
|
||||
|
||||
// Equation was for -(ln(red)/(1-red))
|
||||
result *= -x1;
|
||||
result += toFloat(exponent) * ln2;
|
||||
|
||||
return select(exceptional,
|
||||
select(useNan, T(nan), T(negInf)),
|
||||
result);
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T pow(const T &x, const T &y)
|
||||
{
|
||||
auto x1 = abs(x);
|
||||
auto z = exp(y * log(x1));
|
||||
|
||||
// Handle special cases
|
||||
const float twoOver23 = 8388608.0f;
|
||||
auto yInt = y == round(y);
|
||||
auto yOddInt = select(yInt, asInt(abs(y) + twoOver23) << 31, 0); // set sign bit
|
||||
|
||||
// x == 0
|
||||
z = select(x == 0.0f,
|
||||
select(y < 0.0f, T(inf) | signmsk(x),
|
||||
select(y == 0.0f, T(1.0f), asFloat(yOddInt) & x)), z);
|
||||
|
||||
// x < 0
|
||||
auto xNegative = x < 0.0f;
|
||||
if (any(xNegative))
|
||||
{
|
||||
auto z1 = z | asFloat(yOddInt);
|
||||
z1 = select(yInt, z1, std::numeric_limits<float>::quiet_NaN());
|
||||
z = select(xNegative, z1, z);
|
||||
}
|
||||
|
||||
auto xFinite = isfinite(x);
|
||||
auto yFinite = isfinite(y);
|
||||
if (all(xFinite & yFinite))
|
||||
return z;
|
||||
|
||||
// x finite and y infinite
|
||||
z = select(andn(xFinite, yFinite),
|
||||
select(x1 == 1.0f, 1.0f,
|
||||
select((x1 > 1.0f) ^ (y < 0.0f), inf, T(0.0f))), z);
|
||||
|
||||
// x infinite
|
||||
z = select(xFinite, z,
|
||||
select(y == 0.0f, 1.0f,
|
||||
select(y < 0.0f, T(0.0f), inf) | (asFloat(yOddInt) & x)));
|
||||
|
||||
return z;
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
__forceinline T pow(const T &x, float y)
|
||||
{
|
||||
return pow(x, T(y));
|
||||
}
|
||||
|
||||
} // namespace fastapprox
|
||||
|
||||
} // namespace embree
|
||||
+24
-23
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -15,15 +28,10 @@ namespace embree
|
||||
|
||||
template<typename T> struct Vec2
|
||||
{
|
||||
enum { N = 2 };
|
||||
union {
|
||||
struct { T x, y; };
|
||||
#if !(defined(__WIN32__) && _MSC_VER == 1800) // workaround for older VS 2013 compiler
|
||||
T components[N];
|
||||
#endif
|
||||
};
|
||||
T x, y;
|
||||
|
||||
typedef T Scalar;
|
||||
enum { N = 2 };
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Construction
|
||||
@@ -39,8 +47,6 @@ namespace embree
|
||||
template<typename T1> __forceinline Vec2( const Vec2<T1>& a ) : x(T(a.x)), y(T(a.y)) {}
|
||||
template<typename T1> __forceinline Vec2& operator =( const Vec2<T1>& other ) { x = other.x; y = other.y; return *this; }
|
||||
|
||||
__forceinline Vec2& operator =( const Vec2& other ) { x = other.x; y = other.y; return *this; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Constants
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -50,13 +56,8 @@ namespace embree
|
||||
__forceinline Vec2( PosInfTy ) : x(pos_inf), y(pos_inf) {}
|
||||
__forceinline Vec2( NegInfTy ) : x(neg_inf), y(neg_inf) {}
|
||||
|
||||
#if defined(__WIN32__) && _MSC_VER == 1800 // workaround for older VS 2013 compiler
|
||||
__forceinline const T& operator [](const size_t axis) const { assert(axis < 2); return (&x)[axis]; }
|
||||
__forceinline T& operator [](const size_t axis) { assert(axis < 2); return (&x)[axis]; }
|
||||
#else
|
||||
__forceinline const T& operator [](const size_t axis) const { assert(axis < 2); return components[axis]; }
|
||||
__forceinline T& operator [](const size_t axis ) { assert(axis < 2); return components[axis]; }
|
||||
#endif
|
||||
__forceinline const T& operator []( const size_t axis ) const { assert(axis < 2); return (&x)[axis]; }
|
||||
__forceinline T& operator []( const size_t axis ) { assert(axis < 2); return (&x)[axis]; }
|
||||
};
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -144,7 +145,7 @@ namespace embree
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
/// Euclidian Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline T dot ( const Vec2<T>& a, const Vec2<T>& b ) { return madd(a.x,b.x,a.y*b.y); }
|
||||
@@ -190,7 +191,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const Vec2<T>& a) {
|
||||
template<typename T> inline std::ostream& operator<<(std::ostream& cout, const Vec2<T>& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ")";
|
||||
}
|
||||
|
||||
@@ -205,11 +206,11 @@ namespace embree
|
||||
|
||||
#include "vec2fa.h"
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined __SSE__
|
||||
#include "../simd/sse.h"
|
||||
#endif
|
||||
|
||||
#if defined(__AVX__)
|
||||
#if defined __AVX__
|
||||
#include "../simd/avx.h"
|
||||
#endif
|
||||
|
||||
@@ -221,7 +222,7 @@ namespace embree
|
||||
{
|
||||
template<> __forceinline Vec2<float>::Vec2(const Vec2fa& a) : x(a.x), y(a.y) {}
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined(__SSE__)
|
||||
template<> __forceinline Vec2<vfloat4>::Vec2(const Vec2fa& a) : x(a.x), y(a.y) {}
|
||||
#endif
|
||||
|
||||
|
||||
+23
-31
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -97,12 +110,6 @@ namespace embree
|
||||
|
||||
__forceinline Vec2fa rcp ( const Vec2fa& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
__m128 reciprocal = _mm_rcp_ps(a.m128);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a.m128, reciprocal), reciprocal);
|
||||
reciprocal = vmulq_f32(vrecpsq_f32(a.m128, reciprocal), reciprocal);
|
||||
return (const Vec2fa)reciprocal;
|
||||
#else
|
||||
#if defined(__AVX512VL__)
|
||||
const Vec2fa r = _mm_rcp14_ps(a.m128);
|
||||
#else
|
||||
@@ -110,15 +117,13 @@ namespace embree
|
||||
#endif
|
||||
|
||||
#if defined(__AVX2__)
|
||||
const Vec2fa h_n = _mm_fnmadd_ps(a, r, vfloat4(1.0)); // First, compute 1 - a * r (which will be very close to 0)
|
||||
const Vec2fa res = _mm_fmadd_ps(r, h_n, r); // Then compute r + r * h_n
|
||||
const Vec2fa res = _mm_mul_ps(r,_mm_fnmadd_ps(r, a, vfloat4(2.0f)));
|
||||
#else
|
||||
const Vec2fa h_n = _mm_sub_ps(vfloat4(1.0f), _mm_mul_ps(a, r)); // First, compute 1 - a * r (which will be very close to 0)
|
||||
const Vec2fa res = _mm_add_ps(r,_mm_mul_ps(r, h_n)); // Then compute r + r * h_n
|
||||
const Vec2fa res = _mm_mul_ps(r,_mm_sub_ps(vfloat4(2.0f), _mm_mul_ps(r, a)));
|
||||
//return _mm_sub_ps(_mm_add_ps(r, r), _mm_mul_ps(_mm_mul_ps(r, r), a));
|
||||
#endif
|
||||
|
||||
return res;
|
||||
#endif //defined(__aarch64__)
|
||||
}
|
||||
|
||||
__forceinline Vec2fa sqrt ( const Vec2fa& a ) { return _mm_sqrt_ps(a.m128); }
|
||||
@@ -126,21 +131,12 @@ namespace embree
|
||||
|
||||
__forceinline Vec2fa rsqrt( const Vec2fa& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
return r;
|
||||
#else
|
||||
|
||||
#if defined(__AVX512VL__)
|
||||
__m128 r = _mm_rsqrt14_ps(a.m128);
|
||||
#else
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
#endif
|
||||
return _mm_add_ps(_mm_mul_ps(_mm_set1_ps(1.5f),r), _mm_mul_ps(_mm_mul_ps(_mm_mul_ps(a, _mm_set1_ps(-0.5f)), r), _mm_mul_ps(r, r)));
|
||||
|
||||
#endif
|
||||
}
|
||||
|
||||
__forceinline Vec2fa zero_fix(const Vec2fa& a) {
|
||||
@@ -173,7 +169,7 @@ namespace embree
|
||||
__forceinline Vec2fa min( const Vec2fa& a, const Vec2fa& b ) { return _mm_min_ps(a.m128,b.m128); }
|
||||
__forceinline Vec2fa max( const Vec2fa& a, const Vec2fa& b ) { return _mm_max_ps(a.m128,b.m128); }
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec2fa mini(const Vec2fa& a, const Vec2fa& b) {
|
||||
const vint4 ai = _mm_castps_si128(a);
|
||||
const vint4 bi = _mm_castps_si128(b);
|
||||
@@ -182,7 +178,7 @@ namespace embree
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec2fa maxi(const Vec2fa& a, const Vec2fa& b) {
|
||||
const vint4 ai = _mm_castps_si128(a);
|
||||
const vint4 bi = _mm_castps_si128(b);
|
||||
@@ -244,7 +240,7 @@ namespace embree
|
||||
__forceinline bool operator !=( const Vec2fa& a, const Vec2fa& b ) { return (_mm_movemask_ps(_mm_cmpneq_ps(a.m128, b.m128)) & 3) != 0; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
/// Euclidian Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__SSE4_1__)
|
||||
@@ -292,11 +288,7 @@ namespace embree
|
||||
/// Rounding Functions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__aarch64__)
|
||||
//__forceinline Vec2fa trunc(const Vec2fa& a) { return vrndq_f32(a); }
|
||||
__forceinline Vec2fa floor(const Vec2fa& a) { return vrndmq_f32(a); }
|
||||
__forceinline Vec2fa ceil (const Vec2fa& a) { return vrndpq_f32(a); }
|
||||
#elif defined (__SSE4_1__)
|
||||
#if defined (__SSE4_1__)
|
||||
//__forceinline Vec2fa trunc( const Vec2fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_NEAREST_INT); }
|
||||
__forceinline Vec2fa floor( const Vec2fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_NEG_INF ); }
|
||||
__forceinline Vec2fa ceil ( const Vec2fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_POS_INF ); }
|
||||
@@ -310,7 +302,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Vec2fa& a) {
|
||||
inline std::ostream& operator<<(std::ostream& cout, const Vec2fa& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ")";
|
||||
}
|
||||
|
||||
|
||||
+41
-30
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -15,18 +28,10 @@ namespace embree
|
||||
|
||||
template<typename T> struct Vec3
|
||||
{
|
||||
enum { N = 3 };
|
||||
|
||||
union {
|
||||
struct {
|
||||
T x, y, z;
|
||||
};
|
||||
#if !(defined(__WIN32__) && _MSC_VER == 1800) // workaround for older VS 2013 compiler
|
||||
T components[N];
|
||||
#endif
|
||||
};
|
||||
T x, y, z;
|
||||
|
||||
typedef T Scalar;
|
||||
enum { N = 3 };
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Construction
|
||||
@@ -42,8 +47,6 @@ namespace embree
|
||||
template<typename T1> __forceinline Vec3( const Vec3<T1>& a ) : x(T(a.x)), y(T(a.y)), z(T(a.z)) {}
|
||||
template<typename T1> __forceinline Vec3& operator =(const Vec3<T1>& other) { x = other.x; y = other.y; z = other.z; return *this; }
|
||||
|
||||
__forceinline Vec3& operator =(const Vec3& other) { x = other.x; y = other.y; z = other.z; return *this; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Constants
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -53,13 +56,8 @@ namespace embree
|
||||
__forceinline Vec3( PosInfTy ) : x(pos_inf), y(pos_inf), z(pos_inf) {}
|
||||
__forceinline Vec3( NegInfTy ) : x(neg_inf), y(neg_inf), z(neg_inf) {}
|
||||
|
||||
#if defined(__WIN32__) && (_MSC_VER == 1800) // workaround for older VS 2013 compiler
|
||||
__forceinline const T& operator []( const size_t axis ) const { assert(axis < 3); return (&x)[axis]; }
|
||||
__forceinline T& operator []( const size_t axis ) { assert(axis < 3); return (&x)[axis]; }
|
||||
#else
|
||||
__forceinline const T& operator [](const size_t axis) const { assert(axis < 3); return components[axis]; }
|
||||
__forceinline T& operator [](const size_t axis) { assert(axis < 3); return components[axis]; }
|
||||
#endif
|
||||
};
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -197,16 +195,16 @@ namespace embree
|
||||
template<typename T> __forceinline Vec3<bool> ge_mask( const Vec3<T>& a, const Vec3<T>& b ) { return Vec3<bool>(a.x>=b.x,a.y>=b.y,a.z>=b.z); }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
/// Euclidian Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline T sqr ( const Vec3<T>& a ) { return dot(a,a); }
|
||||
template<typename T> __forceinline T dot ( const Vec3<T>& a, const Vec3<T>& b ) { return madd(a.x,b.x,madd(a.y,b.y,a.z*b.z)); }
|
||||
template<typename T> __forceinline T length ( const Vec3<T>& a ) { return sqrt(sqr(a)); }
|
||||
template<typename T> __forceinline T rcp_length( const Vec3<T>& a ) { return rsqrt(sqr(a)); }
|
||||
template<typename T> __forceinline Vec3<T> normalize( const Vec3<T>& a ) { return a*rsqrt(sqr(a)); }
|
||||
template<typename T> __forceinline T length ( const Vec3<T>& a ) { return sqrt(dot(a,a)); }
|
||||
template<typename T> __forceinline T rcp_length( const Vec3<T>& a ) { return rsqrt(dot(a,a)); }
|
||||
template<typename T> __forceinline Vec3<T> normalize( const Vec3<T>& a ) { return a*rsqrt(dot(a,a)); }
|
||||
template<typename T> __forceinline T distance ( const Vec3<T>& a, const Vec3<T>& b ) { return length(a-b); }
|
||||
template<typename T> __forceinline Vec3<T> cross ( const Vec3<T>& a, const Vec3<T>& b ) { return Vec3<T>(msub(a.y,b.z,a.z*b.y), msub(a.z,b.x,a.x*b.z), msub(a.x,b.y,a.y*b.x)); }
|
||||
|
||||
template<typename T> __forceinline Vec3<T> stable_triangle_normal( const Vec3<T>& a, const Vec3<T>& b, const Vec3<T>& c )
|
||||
{
|
||||
const T ab_x = a.z*b.y, ab_y = a.x*b.z, ab_z = a.y*b.x;
|
||||
@@ -248,7 +246,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const Vec3<T>& a) {
|
||||
template<typename T> inline std::ostream& operator<<(std::ostream& cout, const Vec3<T>& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ", " << a.z << ")";
|
||||
}
|
||||
|
||||
@@ -265,11 +263,11 @@ namespace embree
|
||||
/// SSE / AVX / MIC specializations
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined __SSE__
|
||||
#include "../simd/sse.h"
|
||||
#endif
|
||||
|
||||
#if defined(__AVX__)
|
||||
#if defined __AVX__
|
||||
#include "../simd/avx.h"
|
||||
#endif
|
||||
|
||||
@@ -290,14 +288,18 @@ namespace embree
|
||||
template<> __forceinline Vec3<vfloat4>::Vec3(const Vec3fa& a) {
|
||||
x = a.x; y = a.y; z = a.z;
|
||||
}
|
||||
#elif defined(__SSE__) || defined(__ARM_NEON)
|
||||
#elif defined(__SSE__)
|
||||
template<>
|
||||
__forceinline Vec3<vfloat4>::Vec3(const Vec3fa& a) {
|
||||
const vfloat4 v = vfloat4(a.m128); x = shuffle<0,0,0,0>(v); y = shuffle<1,1,1,1>(v); z = shuffle<2,2,2,2>(v);
|
||||
const vfloat4 v = vfloat4(a); x = shuffle<0,0,0,0>(v); y = shuffle<1,1,1,1>(v); z = shuffle<2,2,2,2>(v);
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined(__SSE__)
|
||||
__forceinline Vec3<vfloat4> broadcast4f(const Vec3<vfloat4>& a, const size_t k) {
|
||||
return Vec3<vfloat4>(vfloat4::broadcast(&a.x[k]), vfloat4::broadcast(&a.y[k]), vfloat4::broadcast(&a.z[k]));
|
||||
}
|
||||
|
||||
template<>
|
||||
__forceinline Vec3<vfloat4> broadcast<vfloat4,vfloat4>(const Vec3<vfloat4>& a, const size_t k) {
|
||||
return Vec3<vfloat4>(vfloat4::broadcast(&a.x[k]), vfloat4::broadcast(&a.y[k]), vfloat4::broadcast(&a.z[k]));
|
||||
@@ -314,6 +316,15 @@ namespace embree
|
||||
__forceinline Vec3<vfloat8>::Vec3(const Vec3fa& a) {
|
||||
x = a.x; y = a.y; z = a.z;
|
||||
}
|
||||
__forceinline Vec3<vfloat4> broadcast4f(const Vec3<vfloat8>& a, const size_t k) {
|
||||
return Vec3<vfloat4>(vfloat4::broadcast(&a.x[k]), vfloat4::broadcast(&a.y[k]), vfloat4::broadcast(&a.z[k]));
|
||||
}
|
||||
__forceinline Vec3<vfloat8> broadcast8f(const Vec3<vfloat4>& a, const size_t k) {
|
||||
return Vec3<vfloat8>(vfloat8::broadcast(&a.x[k]), vfloat8::broadcast(&a.y[k]), vfloat8::broadcast(&a.z[k]));
|
||||
}
|
||||
__forceinline Vec3<vfloat8> broadcast8f(const Vec3<vfloat8>& a, const size_t k) {
|
||||
return Vec3<vfloat8>(vfloat8::broadcast(&a.x[k]), vfloat8::broadcast(&a.y[k]), vfloat8::broadcast(&a.z[k]));
|
||||
}
|
||||
|
||||
template<>
|
||||
__forceinline Vec3<vfloat8> broadcast<vfloat8,vfloat4>(const Vec3<vfloat4>& a, const size_t k) {
|
||||
|
||||
+17
-6
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -19,7 +32,7 @@ namespace embree
|
||||
|
||||
union {
|
||||
__m128 m128;
|
||||
struct { int x,y,z; };
|
||||
struct { int x,y,z; int a; };
|
||||
};
|
||||
|
||||
typedef int Scalar;
|
||||
@@ -108,13 +121,11 @@ namespace embree
|
||||
__forceinline bool any ( const Vec3ba& b ) { return (_mm_movemask_ps(b) & 0x7) != 0x0; }
|
||||
__forceinline bool none ( const Vec3ba& b ) { return (_mm_movemask_ps(b) & 0x7) == 0x0; }
|
||||
|
||||
__forceinline size_t movemask(const Vec3ba& a) { return _mm_movemask_ps(a) & 0x7; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Vec3ba& a) {
|
||||
inline std::ostream& operator<<(std::ostream& cout, const Vec3ba& a) {
|
||||
return cout << "(" << (a.x ? "1" : "0") << ", " << (a.y ? "1" : "0") << ", " << (a.z ? "1" : "0") << ")";
|
||||
}
|
||||
}
|
||||
|
||||
+79
-486
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -21,7 +34,7 @@ namespace embree
|
||||
enum { N = 3 };
|
||||
union {
|
||||
__m128 m128;
|
||||
struct { float x,y,z; };
|
||||
struct { float x,y,z; union { int a; unsigned u; float w; }; };
|
||||
};
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -31,37 +44,47 @@ namespace embree
|
||||
__forceinline Vec3fa( ) {}
|
||||
__forceinline Vec3fa( const __m128 a ) : m128(a) {}
|
||||
|
||||
__forceinline Vec3fa ( const Vec3<float>& other ) { m128 = _mm_set_ps(0, other.z, other.y, other.x); }
|
||||
//__forceinline Vec3fa& operator =( const Vec3<float>& other ) { m128 = _mm_set_ps(0, other.z, other.y, other.x); return *this; }
|
||||
__forceinline Vec3fa ( const Vec3<float>& other ) { x = other.x; y = other.y; z = other.z; }
|
||||
__forceinline Vec3fa& operator =( const Vec3<float>& other ) { x = other.x; y = other.y; z = other.z; return *this; }
|
||||
|
||||
__forceinline Vec3fa ( const Vec3fa& other ) { m128 = other.m128; }
|
||||
|
||||
__forceinline Vec3fa& operator =( const Vec3fa& other ) { m128 = other.m128; return *this; }
|
||||
|
||||
__forceinline explicit Vec3fa( const float a ) : m128(_mm_set1_ps(a)) {}
|
||||
__forceinline Vec3fa( const float x, const float y, const float z) : m128(_mm_set_ps(0, z, y, x)) {}
|
||||
__forceinline Vec3fa( const float x, const float y, const float z) : m128(_mm_set_ps(z, z, y, x)) {}
|
||||
|
||||
__forceinline Vec3fa( const Vec3fa& other, const int a1) { m128 = other.m128; a = a1; }
|
||||
__forceinline Vec3fa( const Vec3fa& other, const unsigned a1) { m128 = other.m128; u = a1; }
|
||||
__forceinline Vec3fa( const Vec3fa& other, const float w1) {
|
||||
#if defined (__SSE4_1__)
|
||||
m128 = _mm_insert_ps(other.m128, _mm_set_ss(w1),3 << 4);
|
||||
#else
|
||||
const vint4 mask(-1,-1,-1,0);
|
||||
m128 = select(vboolf4(_mm_castsi128_ps(mask)),vfloat4(other.m128),vfloat4(w1));
|
||||
#endif
|
||||
}
|
||||
//__forceinline Vec3fa( const float x, const float y, const float z, const int a) : x(x), y(y), z(z), a(a) {} // not working properly!
|
||||
//__forceinline Vec3fa( const float x, const float y, const float z, const unsigned a) : x(x), y(y), z(z), u(a) {} // not working properly!
|
||||
__forceinline Vec3fa( const float x, const float y, const float z, const float w) : m128(_mm_set_ps(w, z, y, x)) {}
|
||||
|
||||
__forceinline explicit Vec3fa( const __m128i a ) : m128(_mm_cvtepi32_ps(a)) {}
|
||||
|
||||
__forceinline explicit operator const vfloat4() const { return vfloat4(m128); }
|
||||
__forceinline explicit operator const vint4() const { return vint4(_mm_cvtps_epi32(m128)); }
|
||||
__forceinline explicit operator const Vec2fa() const { return Vec2fa(m128); }
|
||||
__forceinline explicit operator const Vec3ia() const { return Vec3ia(_mm_cvtps_epi32(m128)); }
|
||||
|
||||
//__forceinline operator const __m128&() const { return m128; }
|
||||
//__forceinline operator __m128&() { return m128; }
|
||||
__forceinline operator const __m128&() const { return m128; }
|
||||
__forceinline operator __m128&() { return m128; }
|
||||
|
||||
#if defined (__SSE4_1__)
|
||||
friend __forceinline Vec3fa copy_a( const Vec3fa& a, const Vec3fa& b ) { return _mm_insert_ps(a, b, (3 << 4) | (3 << 6)); }
|
||||
#else
|
||||
friend __forceinline Vec3fa copy_a( const Vec3fa& a, const Vec3fa& b ) { Vec3fa c = a; c.a = b.a; return c; }
|
||||
#endif
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Loads and Stores
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
static __forceinline Vec3fa load( const void* const a ) {
|
||||
#if defined(__aarch64__)
|
||||
__m128 t = _mm_load_ps((float*)a);
|
||||
t[3] = 0.0f;
|
||||
return Vec3fa(t);
|
||||
#else
|
||||
return Vec3fa(_mm_and_ps(_mm_load_ps((float*)a),_mm_castsi128_ps(_mm_set_epi32(0, -1, -1, -1))));
|
||||
#endif
|
||||
}
|
||||
|
||||
static __forceinline Vec3fa loadu( const void* const a ) {
|
||||
@@ -69,7 +92,7 @@ namespace embree
|
||||
}
|
||||
|
||||
static __forceinline void storeu ( void* ptr, const Vec3fa& v ) {
|
||||
_mm_storeu_ps((float*)ptr,v.m128);
|
||||
_mm_storeu_ps((float*)ptr,v);
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -95,31 +118,19 @@ namespace embree
|
||||
|
||||
__forceinline Vec3fa operator +( const Vec3fa& a ) { return a; }
|
||||
__forceinline Vec3fa operator -( const Vec3fa& a ) {
|
||||
#if defined(__aarch64__)
|
||||
return vnegq_f32(a.m128);
|
||||
#else
|
||||
const __m128 mask = _mm_castsi128_ps(_mm_set1_epi32(0x80000000));
|
||||
return _mm_xor_ps(a.m128, mask);
|
||||
#endif
|
||||
}
|
||||
__forceinline Vec3fa abs ( const Vec3fa& a ) {
|
||||
#if defined(__aarch64__)
|
||||
return _mm_abs_ps(a.m128);
|
||||
#else
|
||||
const __m128 mask = _mm_castsi128_ps(_mm_set1_epi32(0x7fffffff));
|
||||
return _mm_and_ps(a.m128, mask);
|
||||
#endif
|
||||
}
|
||||
__forceinline Vec3fa sign ( const Vec3fa& a ) {
|
||||
return blendv_ps(Vec3fa(one).m128, (-Vec3fa(one)).m128, _mm_cmplt_ps (a.m128,Vec3fa(zero).m128));
|
||||
return blendv_ps(Vec3fa(one), -Vec3fa(one), _mm_cmplt_ps (a,Vec3fa(zero)));
|
||||
}
|
||||
|
||||
__forceinline Vec3fa rcp ( const Vec3fa& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
return vdivq_f32(vdupq_n_f32(1.0f),a.m128);
|
||||
#else
|
||||
|
||||
#if defined(__AVX512VL__)
|
||||
const Vec3fa r = _mm_rcp14_ps(a.m128);
|
||||
#else
|
||||
@@ -127,40 +138,30 @@ namespace embree
|
||||
#endif
|
||||
|
||||
#if defined(__AVX2__)
|
||||
const Vec3fa h_n = _mm_fnmadd_ps(a.m128, r.m128, vfloat4(1.0)); // First, compute 1 - a * r (which will be very close to 0)
|
||||
const Vec3fa res = _mm_fmadd_ps(r.m128, h_n.m128, r.m128); // Then compute r + r * h_n
|
||||
const Vec3fa res = _mm_mul_ps(r,_mm_fnmadd_ps(r, a, vfloat4(2.0f)));
|
||||
#else
|
||||
const Vec3fa h_n = _mm_sub_ps(vfloat4(1.0f), _mm_mul_ps(a.m128, r.m128)); // First, compute 1 - a * r (which will be very close to 0)
|
||||
const Vec3fa res = _mm_add_ps(r.m128,_mm_mul_ps(r.m128, h_n.m128)); // Then compute r + r * h_n
|
||||
const Vec3fa res = _mm_mul_ps(r,_mm_sub_ps(vfloat4(2.0f), _mm_mul_ps(r, a)));
|
||||
//return _mm_sub_ps(_mm_add_ps(r, r), _mm_mul_ps(_mm_mul_ps(r, r), a));
|
||||
#endif
|
||||
|
||||
return res;
|
||||
#endif //defined(__aarch64__)
|
||||
}
|
||||
|
||||
__forceinline Vec3fa sqrt ( const Vec3fa& a ) { return _mm_sqrt_ps(a.m128); }
|
||||
__forceinline Vec3fa sqr ( const Vec3fa& a ) { return _mm_mul_ps(a.m128,a.m128); }
|
||||
__forceinline Vec3fa sqr ( const Vec3fa& a ) { return _mm_mul_ps(a,a); }
|
||||
|
||||
__forceinline Vec3fa rsqrt( const Vec3fa& a )
|
||||
{
|
||||
#if defined(__aarch64__)
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
r = vmulq_f32(r, vrsqrtsq_f32(vmulq_f32(a.m128, r), r));
|
||||
return r;
|
||||
#else
|
||||
|
||||
#if defined(__AVX512VL__)
|
||||
__m128 r = _mm_rsqrt14_ps(a.m128);
|
||||
#else
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
#endif
|
||||
return _mm_add_ps(_mm_mul_ps(_mm_set1_ps(1.5f),r), _mm_mul_ps(_mm_mul_ps(_mm_mul_ps(a.m128, _mm_set1_ps(-0.5f)), r), _mm_mul_ps(r, r)));
|
||||
#endif
|
||||
return _mm_add_ps(_mm_mul_ps(_mm_set1_ps(1.5f),r), _mm_mul_ps(_mm_mul_ps(_mm_mul_ps(a, _mm_set1_ps(-0.5f)), r), _mm_mul_ps(r, r)));
|
||||
}
|
||||
|
||||
__forceinline Vec3fa zero_fix(const Vec3fa& a) {
|
||||
return blendv_ps(a.m128, _mm_set1_ps(min_rcp_input), _mm_cmplt_ps (abs(a).m128, _mm_set1_ps(min_rcp_input)));
|
||||
return blendv_ps(a, _mm_set1_ps(min_rcp_input), _mm_cmplt_ps (abs(a).m128, _mm_set1_ps(min_rcp_input)));
|
||||
}
|
||||
__forceinline Vec3fa rcp_safe(const Vec3fa& a) {
|
||||
return rcp(zero_fix(a));
|
||||
@@ -189,19 +190,19 @@ namespace embree
|
||||
__forceinline Vec3fa min( const Vec3fa& a, const Vec3fa& b ) { return _mm_min_ps(a.m128,b.m128); }
|
||||
__forceinline Vec3fa max( const Vec3fa& a, const Vec3fa& b ) { return _mm_max_ps(a.m128,b.m128); }
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec3fa mini(const Vec3fa& a, const Vec3fa& b) {
|
||||
const vint4 ai = _mm_castps_si128(a.m128);
|
||||
const vint4 bi = _mm_castps_si128(b.m128);
|
||||
const vint4 ai = _mm_castps_si128(a);
|
||||
const vint4 bi = _mm_castps_si128(b);
|
||||
const vint4 ci = _mm_min_epi32(ai,bi);
|
||||
return _mm_castsi128_ps(ci);
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec3fa maxi(const Vec3fa& a, const Vec3fa& b) {
|
||||
const vint4 ai = _mm_castps_si128(a.m128);
|
||||
const vint4 bi = _mm_castps_si128(b.m128);
|
||||
const vint4 ai = _mm_castps_si128(a);
|
||||
const vint4 bi = _mm_castps_si128(b);
|
||||
const vint4 ci = _mm_max_epi32(ai,bi);
|
||||
return _mm_castsi128_ps(ci);
|
||||
}
|
||||
@@ -215,16 +216,16 @@ namespace embree
|
||||
/// Ternary Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__AVX2__) || defined(__ARM_NEON)
|
||||
__forceinline Vec3fa madd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fmadd_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fa msub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fmsub_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fa nmadd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fnmadd_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fa nmsub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fnmsub_ps(a.m128,b.m128,c.m128); }
|
||||
#if defined(__AVX2__)
|
||||
__forceinline Vec3fa madd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fmadd_ps(a,b,c); }
|
||||
__forceinline Vec3fa msub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fmsub_ps(a,b,c); }
|
||||
__forceinline Vec3fa nmadd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fnmadd_ps(a,b,c); }
|
||||
__forceinline Vec3fa nmsub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return _mm_fnmsub_ps(a,b,c); }
|
||||
#else
|
||||
__forceinline Vec3fa madd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return a*b+c; }
|
||||
__forceinline Vec3fa msub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return a*b-c; }
|
||||
__forceinline Vec3fa nmadd ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return -a*b+c;}
|
||||
__forceinline Vec3fa nmsub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return -a*b-c; }
|
||||
__forceinline Vec3fa msub ( const Vec3fa& a, const Vec3fa& b, const Vec3fa& c) { return a*b-c; }
|
||||
#endif
|
||||
|
||||
__forceinline Vec3fa madd ( const float a, const Vec3fa& b, const Vec3fa& c) { return madd(Vec3fa(a),b,c); }
|
||||
@@ -246,27 +247,9 @@ namespace embree
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Reductions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
#if defined(__aarch64__)
|
||||
__forceinline float reduce_add(const Vec3fa& v) {
|
||||
float32x4_t t = v.m128;
|
||||
t[3] = 0.0f;
|
||||
return vaddvq_f32(t);
|
||||
}
|
||||
|
||||
__forceinline float reduce_mul(const Vec3fa& v) { return v.x*v.y*v.z; }
|
||||
__forceinline float reduce_min(const Vec3fa& v) {
|
||||
float32x4_t t = v.m128;
|
||||
t[3] = t[2];
|
||||
return vminvq_f32(t);
|
||||
}
|
||||
__forceinline float reduce_max(const Vec3fa& v) {
|
||||
float32x4_t t = v.m128;
|
||||
t[3] = t[2];
|
||||
return vmaxvq_f32(t);
|
||||
}
|
||||
#else
|
||||
__forceinline float reduce_add(const Vec3fa& v) {
|
||||
const vfloat4 a(v.m128);
|
||||
__forceinline float reduce_add(const Vec3fa& v) {
|
||||
const vfloat4 a(v);
|
||||
const vfloat4 b = shuffle<1>(a);
|
||||
const vfloat4 c = shuffle<2>(a);
|
||||
return _mm_cvtss_f32(a+b+c);
|
||||
@@ -275,7 +258,6 @@ namespace embree
|
||||
__forceinline float reduce_mul(const Vec3fa& v) { return v.x*v.y*v.z; }
|
||||
__forceinline float reduce_min(const Vec3fa& v) { return min(v.x,v.y,v.z); }
|
||||
__forceinline float reduce_max(const Vec3fa& v) { return max(v.x,v.y,v.z); }
|
||||
#endif
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Comparison Operators
|
||||
@@ -288,13 +270,8 @@ namespace embree
|
||||
__forceinline Vec3ba neq_mask(const Vec3fa& a, const Vec3fa& b ) { return _mm_cmpneq_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3ba lt_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmplt_ps (a.m128, b.m128); }
|
||||
__forceinline Vec3ba le_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmple_ps (a.m128, b.m128); }
|
||||
#if defined(__aarch64__)
|
||||
__forceinline Vec3ba gt_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmpgt_ps (a.m128, b.m128); }
|
||||
__forceinline Vec3ba ge_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmpge_ps (a.m128, b.m128); }
|
||||
#else
|
||||
__forceinline Vec3ba gt_mask(const Vec3fa& a, const Vec3fa& b) { return _mm_cmpnle_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3ba ge_mask(const Vec3fa& a, const Vec3fa& b) { return _mm_cmpnlt_ps(a.m128, b.m128); }
|
||||
#endif
|
||||
__forceinline Vec3ba gt_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmpnle_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3ba ge_mask( const Vec3fa& a, const Vec3fa& b ) { return _mm_cmpnlt_ps(a.m128, b.m128); }
|
||||
|
||||
__forceinline bool isvalid ( const Vec3fa& v ) {
|
||||
return all(gt_mask(v,Vec3fa(-FLT_LARGE)) & lt_mask(v,Vec3fa(+FLT_LARGE)));
|
||||
@@ -304,21 +281,13 @@ namespace embree
|
||||
return all(ge_mask(a,Vec3fa(-FLT_MAX)) & le_mask(a,Vec3fa(+FLT_MAX)));
|
||||
}
|
||||
|
||||
__forceinline bool isvalid4 ( const Vec3fa& v ) {
|
||||
return all((vfloat4(v.m128) > vfloat4(-FLT_LARGE)) & (vfloat4(v.m128) < vfloat4(+FLT_LARGE)));
|
||||
}
|
||||
|
||||
__forceinline bool is_finite4 ( const Vec3fa& a ) {
|
||||
return all((vfloat4(a.m128) >= vfloat4(-FLT_MAX)) & (vfloat4(a.m128) <= vfloat4(+FLT_MAX)));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
/// Euclidian Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline float dot ( const Vec3fa& a, const Vec3fa& b ) {
|
||||
return _mm_cvtss_f32(_mm_dp_ps(a.m128,b.m128,0x7F));
|
||||
return _mm_cvtss_f32(_mm_dp_ps(a,b,0x7F));
|
||||
}
|
||||
#else
|
||||
__forceinline float dot ( const Vec3fa& a, const Vec3fa& b ) {
|
||||
@@ -328,10 +297,10 @@ namespace embree
|
||||
|
||||
__forceinline Vec3fa cross ( const Vec3fa& a, const Vec3fa& b )
|
||||
{
|
||||
vfloat4 a0 = vfloat4(a.m128);
|
||||
vfloat4 b0 = shuffle<1,2,0,3>(vfloat4(b.m128));
|
||||
vfloat4 a1 = shuffle<1,2,0,3>(vfloat4(a.m128));
|
||||
vfloat4 b1 = vfloat4(b.m128);
|
||||
vfloat4 a0 = vfloat4(a);
|
||||
vfloat4 b0 = shuffle<1,2,0,3>(vfloat4(b));
|
||||
vfloat4 a1 = shuffle<1,2,0,3>(vfloat4(a));
|
||||
vfloat4 b1 = vfloat4(b);
|
||||
return Vec3fa(shuffle<1,2,0,3>(msub(a0,b0,a1*b1)));
|
||||
}
|
||||
|
||||
@@ -362,11 +331,11 @@ namespace embree
|
||||
|
||||
__forceinline Vec3fa select( bool s, const Vec3fa& t, const Vec3fa& f ) {
|
||||
__m128 mask = s ? _mm_castsi128_ps(_mm_cmpeq_epi32(_mm_setzero_si128(), _mm_setzero_si128())) : _mm_setzero_ps();
|
||||
return blendv_ps(f.m128, t.m128, mask);
|
||||
return blendv_ps(f, t, mask);
|
||||
}
|
||||
|
||||
__forceinline Vec3fa select( const Vec3ba& s, const Vec3fa& t, const Vec3fa& f ) {
|
||||
return blendv_ps(f.m128, t.m128, s);
|
||||
return blendv_ps(f, t, s);
|
||||
}
|
||||
|
||||
__forceinline Vec3fa lerp(const Vec3fa& v0, const Vec3fa& v1, const float t) {
|
||||
@@ -387,14 +356,10 @@ namespace embree
|
||||
/// Rounding Functions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__aarch64__)
|
||||
__forceinline Vec3fa floor(const Vec3fa& a) { return vrndmq_f32(a.m128); }
|
||||
__forceinline Vec3fa ceil (const Vec3fa& a) { return vrndpq_f32(a.m128); }
|
||||
__forceinline Vec3fa trunc(const Vec3fa& a) { return vrndq_f32(a.m128); }
|
||||
#elif defined (__SSE4_1__)
|
||||
__forceinline Vec3fa trunc( const Vec3fa& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_NEAREST_INT); }
|
||||
__forceinline Vec3fa floor( const Vec3fa& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_NEG_INF ); }
|
||||
__forceinline Vec3fa ceil ( const Vec3fa& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_POS_INF ); }
|
||||
#if defined (__SSE4_1__)
|
||||
__forceinline Vec3fa trunc( const Vec3fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_NEAREST_INT); }
|
||||
__forceinline Vec3fa floor( const Vec3fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_NEG_INF ); }
|
||||
__forceinline Vec3fa ceil ( const Vec3fa& a ) { return _mm_round_ps(a, _MM_FROUND_TO_POS_INF ); }
|
||||
#else
|
||||
__forceinline Vec3fa trunc( const Vec3fa& a ) { return Vec3fa(truncf(a.x),truncf(a.y),truncf(a.z)); }
|
||||
__forceinline Vec3fa floor( const Vec3fa& a ) { return Vec3fa(floorf(a.x),floorf(a.y),floorf(a.z)); }
|
||||
@@ -405,381 +370,9 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Vec3fa& a) {
|
||||
inline std::ostream& operator<<(std::ostream& cout, const Vec3fa& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ", " << a.z << ")";
|
||||
}
|
||||
|
||||
typedef Vec3fa Vec3fa_t;
|
||||
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// SSE Vec3fx Type
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
struct __aligned(16) Vec3fx
|
||||
{
|
||||
ALIGNED_STRUCT_(16);
|
||||
|
||||
typedef float Scalar;
|
||||
enum { N = 3 };
|
||||
union {
|
||||
__m128 m128;
|
||||
struct { float x,y,z; union { int a; unsigned u; float w; }; };
|
||||
};
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Constructors, Assignment & Cast Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx( ) {}
|
||||
__forceinline Vec3fx( const __m128 a ) : m128(a) {}
|
||||
|
||||
__forceinline explicit Vec3fx(const Vec3fa& v) : m128(v.m128) {}
|
||||
__forceinline operator Vec3fa () const { return Vec3fa(m128); }
|
||||
|
||||
__forceinline explicit Vec3fx ( const Vec3<float>& other ) { m128 = _mm_set_ps(0, other.z, other.y, other.x); }
|
||||
//__forceinline Vec3fx& operator =( const Vec3<float>& other ) { m128 = _mm_set_ps(0, other.z, other.y, other.x); return *this; }
|
||||
|
||||
__forceinline Vec3fx ( const Vec3fx& other ) { m128 = other.m128; }
|
||||
|
||||
__forceinline Vec3fx& operator =( const Vec3fx& other ) { m128 = other.m128; return *this; }
|
||||
|
||||
__forceinline explicit Vec3fx( const float a ) : m128(_mm_set1_ps(a)) {}
|
||||
__forceinline Vec3fx( const float x, const float y, const float z) : m128(_mm_set_ps(0, z, y, x)) {}
|
||||
|
||||
__forceinline Vec3fx( const Vec3fa& other, const int a1) { m128 = other.m128; a = a1; }
|
||||
__forceinline Vec3fx( const Vec3fa& other, const unsigned a1) { m128 = other.m128; u = a1; }
|
||||
__forceinline Vec3fx( const Vec3fa& other, const float w1) {
|
||||
#if defined (__aarch64__)
|
||||
m128 = other.m128; m128[3] = w1;
|
||||
#elif defined (__SSE4_1__)
|
||||
m128 = _mm_insert_ps(other.m128, _mm_set_ss(w1),3 << 4);
|
||||
#else
|
||||
const vint4 mask(-1,-1,-1,0);
|
||||
m128 = select(vboolf4(_mm_castsi128_ps(mask)),vfloat4(other.m128),vfloat4(w1));
|
||||
#endif
|
||||
}
|
||||
//__forceinline Vec3fx( const float x, const float y, const float z, const int a) : x(x), y(y), z(z), a(a) {} // not working properly!
|
||||
//__forceinline Vec3fx( const float x, const float y, const float z, const unsigned a) : x(x), y(y), z(z), u(a) {} // not working properly!
|
||||
__forceinline Vec3fx( const float x, const float y, const float z, const float w) : m128(_mm_set_ps(w, z, y, x)) {}
|
||||
|
||||
//__forceinline explicit Vec3fx( const __m128i a ) : m128(_mm_cvtepi32_ps(a)) {}
|
||||
|
||||
__forceinline explicit operator const vfloat4() const { return vfloat4(m128); }
|
||||
__forceinline explicit operator const vint4() const { return vint4(_mm_cvtps_epi32(m128)); }
|
||||
__forceinline explicit operator const Vec2fa() const { return Vec2fa(m128); }
|
||||
__forceinline explicit operator const Vec3ia() const { return Vec3ia(_mm_cvtps_epi32(m128)); }
|
||||
|
||||
//__forceinline operator const __m128&() const { return m128; }
|
||||
//__forceinline operator __m128&() { return m128; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Loads and Stores
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
static __forceinline Vec3fx load( const void* const a ) {
|
||||
return Vec3fx(_mm_and_ps(_mm_load_ps((float*)a),_mm_castsi128_ps(_mm_set_epi32(0, -1, -1, -1))));
|
||||
}
|
||||
|
||||
static __forceinline Vec3fx loadu( const void* const a ) {
|
||||
return Vec3fx(_mm_loadu_ps((float*)a));
|
||||
}
|
||||
|
||||
static __forceinline void storeu ( void* ptr, const Vec3fx& v ) {
|
||||
_mm_storeu_ps((float*)ptr,v.m128);
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Constants
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx( ZeroTy ) : m128(_mm_setzero_ps()) {}
|
||||
__forceinline Vec3fx( OneTy ) : m128(_mm_set1_ps(1.0f)) {}
|
||||
__forceinline Vec3fx( PosInfTy ) : m128(_mm_set1_ps(pos_inf)) {}
|
||||
__forceinline Vec3fx( NegInfTy ) : m128(_mm_set1_ps(neg_inf)) {}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Array Access
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline const float& operator []( const size_t index ) const { assert(index < 3); return (&x)[index]; }
|
||||
__forceinline float& operator []( const size_t index ) { assert(index < 3); return (&x)[index]; }
|
||||
};
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Unary Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx operator +( const Vec3fx& a ) { return a; }
|
||||
__forceinline Vec3fx operator -( const Vec3fx& a ) {
|
||||
const __m128 mask = _mm_castsi128_ps(_mm_set1_epi32(0x80000000));
|
||||
return _mm_xor_ps(a.m128, mask);
|
||||
}
|
||||
__forceinline Vec3fx abs ( const Vec3fx& a ) {
|
||||
const __m128 mask = _mm_castsi128_ps(_mm_set1_epi32(0x7fffffff));
|
||||
return _mm_and_ps(a.m128, mask);
|
||||
}
|
||||
__forceinline Vec3fx sign ( const Vec3fx& a ) {
|
||||
return blendv_ps(Vec3fx(one).m128, (-Vec3fx(one)).m128, _mm_cmplt_ps (a.m128,Vec3fx(zero).m128));
|
||||
}
|
||||
|
||||
__forceinline Vec3fx rcp ( const Vec3fx& a )
|
||||
{
|
||||
#if defined(__AVX512VL__)
|
||||
const Vec3fx r = _mm_rcp14_ps(a.m128);
|
||||
#else
|
||||
const Vec3fx r = _mm_rcp_ps(a.m128);
|
||||
#endif
|
||||
|
||||
#if defined(__AVX2__)
|
||||
const Vec3fx res = _mm_mul_ps(r.m128,_mm_fnmadd_ps(r.m128, a.m128, vfloat4(2.0f)));
|
||||
#else
|
||||
const Vec3fx res = _mm_mul_ps(r.m128,_mm_sub_ps(vfloat4(2.0f), _mm_mul_ps(r.m128, a.m128)));
|
||||
//return _mm_sub_ps(_mm_add_ps(r, r), _mm_mul_ps(_mm_mul_ps(r, r), a));
|
||||
#endif
|
||||
|
||||
return res;
|
||||
}
|
||||
|
||||
__forceinline Vec3fx sqrt ( const Vec3fx& a ) { return _mm_sqrt_ps(a.m128); }
|
||||
__forceinline Vec3fx sqr ( const Vec3fx& a ) { return _mm_mul_ps(a.m128,a.m128); }
|
||||
|
||||
__forceinline Vec3fx rsqrt( const Vec3fx& a )
|
||||
{
|
||||
#if defined(__AVX512VL__)
|
||||
__m128 r = _mm_rsqrt14_ps(a.m128);
|
||||
#else
|
||||
__m128 r = _mm_rsqrt_ps(a.m128);
|
||||
#endif
|
||||
return _mm_add_ps(_mm_mul_ps(_mm_set1_ps(1.5f),r), _mm_mul_ps(_mm_mul_ps(_mm_mul_ps(a.m128, _mm_set1_ps(-0.5f)), r), _mm_mul_ps(r, r)));
|
||||
}
|
||||
|
||||
__forceinline Vec3fx zero_fix(const Vec3fx& a) {
|
||||
return blendv_ps(a.m128, _mm_set1_ps(min_rcp_input), _mm_cmplt_ps (abs(a).m128, _mm_set1_ps(min_rcp_input)));
|
||||
}
|
||||
__forceinline Vec3fx rcp_safe(const Vec3fx& a) {
|
||||
return rcp(zero_fix(a));
|
||||
}
|
||||
__forceinline Vec3fx log ( const Vec3fx& a ) {
|
||||
return Vec3fx(logf(a.x),logf(a.y),logf(a.z));
|
||||
}
|
||||
|
||||
__forceinline Vec3fx exp ( const Vec3fx& a ) {
|
||||
return Vec3fx(expf(a.x),expf(a.y),expf(a.z));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Binary Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx operator +( const Vec3fx& a, const Vec3fx& b ) { return _mm_add_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3fx operator -( const Vec3fx& a, const Vec3fx& b ) { return _mm_sub_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3fx operator *( const Vec3fx& a, const Vec3fx& b ) { return _mm_mul_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3fx operator *( const Vec3fx& a, const float b ) { return a * Vec3fx(b); }
|
||||
__forceinline Vec3fx operator *( const float a, const Vec3fx& b ) { return Vec3fx(a) * b; }
|
||||
__forceinline Vec3fx operator /( const Vec3fx& a, const Vec3fx& b ) { return _mm_div_ps(a.m128,b.m128); }
|
||||
__forceinline Vec3fx operator /( const Vec3fx& a, const float b ) { return _mm_div_ps(a.m128,_mm_set1_ps(b)); }
|
||||
__forceinline Vec3fx operator /( const float a, const Vec3fx& b ) { return _mm_div_ps(_mm_set1_ps(a),b.m128); }
|
||||
|
||||
__forceinline Vec3fx min( const Vec3fx& a, const Vec3fx& b ) { return _mm_min_ps(a.m128,b.m128); }
|
||||
__forceinline Vec3fx max( const Vec3fx& a, const Vec3fx& b ) { return _mm_max_ps(a.m128,b.m128); }
|
||||
|
||||
#if defined(__SSE4_1__) || defined(__aarch64__)
|
||||
__forceinline Vec3fx mini(const Vec3fx& a, const Vec3fx& b) {
|
||||
const vint4 ai = _mm_castps_si128(a.m128);
|
||||
const vint4 bi = _mm_castps_si128(b.m128);
|
||||
const vint4 ci = _mm_min_epi32(ai,bi);
|
||||
return _mm_castsi128_ps(ci);
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__SSE4_1__) || defined(__aarch64__)
|
||||
__forceinline Vec3fx maxi(const Vec3fx& a, const Vec3fx& b) {
|
||||
const vint4 ai = _mm_castps_si128(a.m128);
|
||||
const vint4 bi = _mm_castps_si128(b.m128);
|
||||
const vint4 ci = _mm_max_epi32(ai,bi);
|
||||
return _mm_castsi128_ps(ci);
|
||||
}
|
||||
#endif
|
||||
|
||||
__forceinline Vec3fx pow ( const Vec3fx& a, const float& b ) {
|
||||
return Vec3fx(powf(a.x,b),powf(a.y,b),powf(a.z,b));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Ternary Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__AVX2__)
|
||||
__forceinline Vec3fx madd ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return _mm_fmadd_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fx msub ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return _mm_fmsub_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fx nmadd ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return _mm_fnmadd_ps(a.m128,b.m128,c.m128); }
|
||||
__forceinline Vec3fx nmsub ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return _mm_fnmsub_ps(a.m128,b.m128,c.m128); }
|
||||
#else
|
||||
__forceinline Vec3fx madd ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return a*b+c; }
|
||||
__forceinline Vec3fx msub ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return a*b-c; }
|
||||
__forceinline Vec3fx nmadd ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return -a*b+c;}
|
||||
__forceinline Vec3fx nmsub ( const Vec3fx& a, const Vec3fx& b, const Vec3fx& c) { return -a*b-c; }
|
||||
#endif
|
||||
|
||||
__forceinline Vec3fx madd ( const float a, const Vec3fx& b, const Vec3fx& c) { return madd(Vec3fx(a),b,c); }
|
||||
__forceinline Vec3fx msub ( const float a, const Vec3fx& b, const Vec3fx& c) { return msub(Vec3fx(a),b,c); }
|
||||
__forceinline Vec3fx nmadd ( const float a, const Vec3fx& b, const Vec3fx& c) { return nmadd(Vec3fx(a),b,c); }
|
||||
__forceinline Vec3fx nmsub ( const float a, const Vec3fx& b, const Vec3fx& c) { return nmsub(Vec3fx(a),b,c); }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Assignment Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx& operator +=( Vec3fx& a, const Vec3fx& b ) { return a = a + b; }
|
||||
__forceinline Vec3fx& operator -=( Vec3fx& a, const Vec3fx& b ) { return a = a - b; }
|
||||
__forceinline Vec3fx& operator *=( Vec3fx& a, const Vec3fx& b ) { return a = a * b; }
|
||||
__forceinline Vec3fx& operator *=( Vec3fx& a, const float b ) { return a = a * b; }
|
||||
__forceinline Vec3fx& operator /=( Vec3fx& a, const Vec3fx& b ) { return a = a / b; }
|
||||
__forceinline Vec3fx& operator /=( Vec3fx& a, const float b ) { return a = a / b; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Reductions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline float reduce_add(const Vec3fx& v) {
|
||||
const vfloat4 a(v.m128);
|
||||
const vfloat4 b = shuffle<1>(a);
|
||||
const vfloat4 c = shuffle<2>(a);
|
||||
return _mm_cvtss_f32(a+b+c);
|
||||
}
|
||||
|
||||
__forceinline float reduce_mul(const Vec3fx& v) { return v.x*v.y*v.z; }
|
||||
__forceinline float reduce_min(const Vec3fx& v) { return min(v.x,v.y,v.z); }
|
||||
__forceinline float reduce_max(const Vec3fx& v) { return max(v.x,v.y,v.z); }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Comparison Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline bool operator ==( const Vec3fx& a, const Vec3fx& b ) { return (_mm_movemask_ps(_mm_cmpeq_ps (a.m128, b.m128)) & 7) == 7; }
|
||||
__forceinline bool operator !=( const Vec3fx& a, const Vec3fx& b ) { return (_mm_movemask_ps(_mm_cmpneq_ps(a.m128, b.m128)) & 7) != 0; }
|
||||
|
||||
__forceinline Vec3ba eq_mask( const Vec3fx& a, const Vec3fx& b ) { return _mm_cmpeq_ps (a.m128, b.m128); }
|
||||
__forceinline Vec3ba neq_mask(const Vec3fx& a, const Vec3fx& b ) { return _mm_cmpneq_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3ba lt_mask( const Vec3fx& a, const Vec3fx& b ) { return _mm_cmplt_ps (a.m128, b.m128); }
|
||||
__forceinline Vec3ba le_mask( const Vec3fx& a, const Vec3fx& b ) { return _mm_cmple_ps (a.m128, b.m128); }
|
||||
__forceinline Vec3ba gt_mask( const Vec3fx& a, const Vec3fx& b ) { return _mm_cmpnle_ps(a.m128, b.m128); }
|
||||
__forceinline Vec3ba ge_mask( const Vec3fx& a, const Vec3fx& b ) { return _mm_cmpnlt_ps(a.m128, b.m128); }
|
||||
|
||||
__forceinline bool isvalid ( const Vec3fx& v ) {
|
||||
return all(gt_mask(v,Vec3fx(-FLT_LARGE)) & lt_mask(v,Vec3fx(+FLT_LARGE)));
|
||||
}
|
||||
|
||||
__forceinline bool is_finite ( const Vec3fx& a ) {
|
||||
return all(ge_mask(a,Vec3fx(-FLT_MAX)) & le_mask(a,Vec3fx(+FLT_MAX)));
|
||||
}
|
||||
|
||||
__forceinline bool isvalid4 ( const Vec3fx& v ) {
|
||||
return all((vfloat4(v.m128) > vfloat4(-FLT_LARGE)) & (vfloat4(v.m128) < vfloat4(+FLT_LARGE)));
|
||||
}
|
||||
|
||||
__forceinline bool is_finite4 ( const Vec3fx& a ) {
|
||||
return all((vfloat4(a.m128) >= vfloat4(-FLT_MAX)) & (vfloat4(a.m128) <= vfloat4(+FLT_MAX)));
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline float dot ( const Vec3fx& a, const Vec3fx& b ) {
|
||||
return _mm_cvtss_f32(_mm_dp_ps(a.m128,b.m128,0x7F));
|
||||
}
|
||||
#else
|
||||
__forceinline float dot ( const Vec3fx& a, const Vec3fx& b ) {
|
||||
return reduce_add(a*b);
|
||||
}
|
||||
#endif
|
||||
|
||||
__forceinline Vec3fx cross ( const Vec3fx& a, const Vec3fx& b )
|
||||
{
|
||||
vfloat4 a0 = vfloat4(a.m128);
|
||||
vfloat4 b0 = shuffle<1,2,0,3>(vfloat4(b.m128));
|
||||
vfloat4 a1 = shuffle<1,2,0,3>(vfloat4(a.m128));
|
||||
vfloat4 b1 = vfloat4(b.m128);
|
||||
return Vec3fx(shuffle<1,2,0,3>(msub(a0,b0,a1*b1)));
|
||||
}
|
||||
|
||||
__forceinline float sqr_length ( const Vec3fx& a ) { return dot(a,a); }
|
||||
__forceinline float rcp_length ( const Vec3fx& a ) { return rsqrt(dot(a,a)); }
|
||||
__forceinline float rcp_length2( const Vec3fx& a ) { return rcp(dot(a,a)); }
|
||||
__forceinline float length ( const Vec3fx& a ) { return sqrt(dot(a,a)); }
|
||||
__forceinline Vec3fx normalize( const Vec3fx& a ) { return a*rsqrt(dot(a,a)); }
|
||||
__forceinline float distance ( const Vec3fx& a, const Vec3fx& b ) { return length(a-b); }
|
||||
__forceinline float halfArea ( const Vec3fx& d ) { return madd(d.x,(d.y+d.z),d.y*d.z); }
|
||||
__forceinline float area ( const Vec3fx& d ) { return 2.0f*halfArea(d); }
|
||||
|
||||
__forceinline Vec3fx normalize_safe( const Vec3fx& a ) {
|
||||
const float d = dot(a,a); if (unlikely(d == 0.0f)) return a; else return a*rsqrt(d);
|
||||
}
|
||||
|
||||
/*! differentiated normalization */
|
||||
__forceinline Vec3fx dnormalize(const Vec3fx& p, const Vec3fx& dp)
|
||||
{
|
||||
const float pp = dot(p,p);
|
||||
const float pdp = dot(p,dp);
|
||||
return (pp*dp-pdp*p)*rcp(pp)*rsqrt(pp);
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Select
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3fx select( bool s, const Vec3fx& t, const Vec3fx& f ) {
|
||||
__m128 mask = s ? _mm_castsi128_ps(_mm_cmpeq_epi32(_mm_setzero_si128(), _mm_setzero_si128())) : _mm_setzero_ps();
|
||||
return blendv_ps(f.m128, t.m128, mask);
|
||||
}
|
||||
|
||||
__forceinline Vec3fx select( const Vec3ba& s, const Vec3fx& t, const Vec3fx& f ) {
|
||||
return blendv_ps(f.m128, t.m128, s);
|
||||
}
|
||||
|
||||
__forceinline Vec3fx lerp(const Vec3fx& v0, const Vec3fx& v1, const float t) {
|
||||
return madd(1.0f-t,v0,t*v1);
|
||||
}
|
||||
|
||||
__forceinline int maxDim ( const Vec3fx& a )
|
||||
{
|
||||
const Vec3fx b = abs(a);
|
||||
if (b.x > b.y) {
|
||||
if (b.x > b.z) return 0; else return 2;
|
||||
} else {
|
||||
if (b.y > b.z) return 1; else return 2;
|
||||
}
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Rounding Functions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__aarch64__)
|
||||
__forceinline Vec3fx trunc(const Vec3fx& a) { return vrndq_f32(a.m128); }
|
||||
__forceinline Vec3fx floor(const Vec3fx& a) { return vrndmq_f32(a.m128); }
|
||||
__forceinline Vec3fx ceil (const Vec3fx& a) { return vrndpq_f32(a.m128); }
|
||||
#elif defined (__SSE4_1__)
|
||||
__forceinline Vec3fx trunc( const Vec3fx& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_NEAREST_INT); }
|
||||
__forceinline Vec3fx floor( const Vec3fx& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_NEG_INF ); }
|
||||
__forceinline Vec3fx ceil ( const Vec3fx& a ) { return _mm_round_ps(a.m128, _MM_FROUND_TO_POS_INF ); }
|
||||
#else
|
||||
__forceinline Vec3fx trunc( const Vec3fx& a ) { return Vec3fx(truncf(a.x),truncf(a.y),truncf(a.z)); }
|
||||
__forceinline Vec3fx floor( const Vec3fx& a ) { return Vec3fx(floorf(a.x),floorf(a.y),floorf(a.z)); }
|
||||
__forceinline Vec3fx ceil ( const Vec3fx& a ) { return Vec3fx(ceilf (a.x),ceilf (a.y),ceilf (a.z)); }
|
||||
#endif
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Vec3fx& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ", " << a.z << ")";
|
||||
}
|
||||
|
||||
|
||||
typedef Vec3fx Vec3ff;
|
||||
}
|
||||
|
||||
+35
-32
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -19,7 +32,7 @@ namespace embree
|
||||
|
||||
union {
|
||||
__m128i m128;
|
||||
struct { int x,y,z; };
|
||||
struct { int x,y,z; int a; };
|
||||
};
|
||||
|
||||
typedef int Scalar;
|
||||
@@ -65,9 +78,7 @@ namespace embree
|
||||
|
||||
__forceinline Vec3ia operator +( const Vec3ia& a ) { return a; }
|
||||
__forceinline Vec3ia operator -( const Vec3ia& a ) { return _mm_sub_epi32(_mm_setzero_si128(), a.m128); }
|
||||
#if (defined(__aarch64__))
|
||||
__forceinline Vec3ia abs ( const Vec3ia& a ) { return vabsq_s32(a.m128); }
|
||||
#elif defined(__SSSE3__)
|
||||
#if defined(__SSSE3__)
|
||||
__forceinline Vec3ia abs ( const Vec3ia& a ) { return _mm_abs_epi32(a.m128); }
|
||||
#endif
|
||||
|
||||
@@ -83,7 +94,7 @@ namespace embree
|
||||
__forceinline Vec3ia operator -( const Vec3ia& a, const int b ) { return a-Vec3ia(b); }
|
||||
__forceinline Vec3ia operator -( const int a, const Vec3ia& b ) { return Vec3ia(a)-b; }
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec3ia operator *( const Vec3ia& a, const Vec3ia& b ) { return _mm_mullo_epi32(a.m128, b.m128); }
|
||||
__forceinline Vec3ia operator *( const Vec3ia& a, const int b ) { return a * Vec3ia(b); }
|
||||
__forceinline Vec3ia operator *( const int a, const Vec3ia& b ) { return Vec3ia(a) * b; }
|
||||
@@ -118,7 +129,7 @@ namespace embree
|
||||
__forceinline Vec3ia& operator -=( Vec3ia& a, const Vec3ia& b ) { return a = a - b; }
|
||||
__forceinline Vec3ia& operator -=( Vec3ia& a, const int& b ) { return a = a - b; }
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec3ia& operator *=( Vec3ia& a, const Vec3ia& b ) { return a = a * b; }
|
||||
__forceinline Vec3ia& operator *=( Vec3ia& a, const int& b ) { return a = a * b; }
|
||||
#endif
|
||||
@@ -129,38 +140,18 @@ namespace embree
|
||||
__forceinline Vec3ia& operator |=( Vec3ia& a, const Vec3ia& b ) { return a = a | b; }
|
||||
__forceinline Vec3ia& operator |=( Vec3ia& a, const int& b ) { return a = a | b; }
|
||||
|
||||
#if !defined(__ARM_NEON)
|
||||
__forceinline Vec3ia& operator <<=( Vec3ia& a, const int& b ) { return a = a << b; }
|
||||
__forceinline Vec3ia& operator >>=( Vec3ia& a, const int& b ) { return a = a >> b; }
|
||||
#endif
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Select
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3ia select( const Vec3ba& m, const Vec3ia& t, const Vec3ia& f ) {
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
return _mm_castps_si128(_mm_blendv_ps(_mm_castsi128_ps(f), _mm_castsi128_ps(t), m));
|
||||
#else
|
||||
return _mm_or_si128(_mm_and_si128(_mm_castps_si128(m), t), _mm_andnot_si128(_mm_castps_si128(m), f));
|
||||
#endif
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Reductions
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
#if defined(__aarch64__)
|
||||
__forceinline int reduce_add(const Vec3ia& v) { return vaddvq_s32(select(Vec3ba(1,1,1),v,Vec3ia(0))); }
|
||||
__forceinline int reduce_mul(const Vec3ia& v) { return v.x*v.y*v.z; }
|
||||
__forceinline int reduce_min(const Vec3ia& v) { return vminvq_s32(select(Vec3ba(1,1,1),v,Vec3ia(0x7FFFFFFF))); }
|
||||
__forceinline int reduce_max(const Vec3ia& v) { return vmaxvq_s32(select(Vec3ba(1,1,1),v,Vec3ia(0x80000000))); }
|
||||
#else
|
||||
|
||||
__forceinline int reduce_add(const Vec3ia& v) { return v.x+v.y+v.z; }
|
||||
__forceinline int reduce_mul(const Vec3ia& v) { return v.x*v.y*v.z; }
|
||||
__forceinline int reduce_min(const Vec3ia& v) { return min(v.x,v.y,v.z); }
|
||||
__forceinline int reduce_max(const Vec3ia& v) { return max(v.x,v.y,v.z); }
|
||||
#endif
|
||||
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Comparison Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -178,7 +169,19 @@ namespace embree
|
||||
__forceinline Vec3ba lt_mask( const Vec3ia& a, const Vec3ia& b ) { return _mm_castsi128_ps(_mm_cmplt_epi32 (a.m128, b.m128)); }
|
||||
__forceinline Vec3ba gt_mask( const Vec3ia& a, const Vec3ia& b ) { return _mm_castsi128_ps(_mm_cmpgt_epi32 (a.m128, b.m128)); }
|
||||
|
||||
#if defined(__aarch64__) || defined(__SSE4_1__)
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Select
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline Vec3ia select( const Vec3ba& m, const Vec3ia& t, const Vec3ia& f ) {
|
||||
#if defined(__SSE4_1__)
|
||||
return _mm_castps_si128(_mm_blendv_ps(_mm_castsi128_ps(f), _mm_castsi128_ps(t), m));
|
||||
#else
|
||||
return _mm_or_si128(_mm_and_si128(_mm_castps_si128(m), t), _mm_andnot_si128(_mm_castps_si128(m), f));
|
||||
#endif
|
||||
}
|
||||
|
||||
#if defined(__SSE4_1__)
|
||||
__forceinline Vec3ia min( const Vec3ia& a, const Vec3ia& b ) { return _mm_min_epi32(a.m128,b.m128); }
|
||||
__forceinline Vec3ia max( const Vec3ia& a, const Vec3ia& b ) { return _mm_max_epi32(a.m128,b.m128); }
|
||||
#else
|
||||
@@ -190,7 +193,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
__forceinline embree_ostream operator<<(embree_ostream cout, const Vec3ia& a) {
|
||||
inline std::ostream& operator<<(std::ostream& cout, const Vec3ia& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ", " << a.z << ")";
|
||||
}
|
||||
}
|
||||
|
||||
+45
-29
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
@@ -14,15 +27,10 @@ namespace embree
|
||||
|
||||
template<typename T> struct Vec4
|
||||
{
|
||||
enum { N = 4 };
|
||||
union {
|
||||
struct { T x, y, z, w; };
|
||||
#if !(defined(__WIN32__) && _MSC_VER == 1800) // workaround for older VS 2013 compiler
|
||||
T components[N];
|
||||
#endif
|
||||
};
|
||||
T x, y, z, w;
|
||||
|
||||
typedef T Scalar;
|
||||
enum { N = 4 };
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Construction
|
||||
@@ -34,13 +42,11 @@ namespace embree
|
||||
__forceinline Vec4( const Vec3<T>& xyz, const T& w ) : x(xyz.x), y(xyz.y), z(xyz.z), w(w) {}
|
||||
|
||||
__forceinline Vec4( const Vec4& other ) { x = other.x; y = other.y; z = other.z; w = other.w; }
|
||||
__forceinline Vec4( const Vec3fx& other );
|
||||
__forceinline Vec4( const Vec3fa& other );
|
||||
|
||||
template<typename T1> __forceinline Vec4( const Vec4<T1>& a ) : x(T(a.x)), y(T(a.y)), z(T(a.z)), w(T(a.w)) {}
|
||||
template<typename T1> __forceinline Vec4& operator =(const Vec4<T1>& other) { x = other.x; y = other.y; z = other.z; w = other.w; return *this; }
|
||||
|
||||
__forceinline Vec4& operator =(const Vec4& other) { x = other.x; y = other.y; z = other.z; w = other.w; return *this; }
|
||||
|
||||
__forceinline operator Vec3<T> () const { return Vec3<T>(x,y,z); }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
@@ -52,13 +58,8 @@ namespace embree
|
||||
__forceinline Vec4( PosInfTy ) : x(pos_inf), y(pos_inf), z(pos_inf), w(pos_inf) {}
|
||||
__forceinline Vec4( NegInfTy ) : x(neg_inf), y(neg_inf), z(neg_inf), w(neg_inf) {}
|
||||
|
||||
#if defined(__WIN32__) && (_MSC_VER == 1800) // workaround for older VS 2013 compiler
|
||||
__forceinline const T& operator [](const size_t axis) const { assert(axis < 4); return (&x)[axis]; }
|
||||
__forceinline T& operator [](const size_t axis) { assert(axis < 4); return (&x)[axis]; }
|
||||
#else
|
||||
__forceinline const T& operator [](const size_t axis ) const { assert(axis < 4); return components[axis]; }
|
||||
__forceinline T& operator [](const size_t axis) { assert(axis < 4); return components[axis]; }
|
||||
#endif
|
||||
__forceinline const T& operator []( const size_t axis ) const { assert(axis < 4); return (&x)[axis]; }
|
||||
__forceinline T& operator []( const size_t axis ) { assert(axis < 4); return (&x)[axis]; }
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Swizzles
|
||||
@@ -149,7 +150,7 @@ namespace embree
|
||||
}
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
/// Euclidean Space Operators
|
||||
/// Euclidian Space Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline T dot ( const Vec4<T>& a, const Vec4<T>& b ) { return madd(a.x,b.x,madd(a.y,b.y,madd(a.z,b.z,a.w*b.w))); }
|
||||
@@ -183,7 +184,7 @@ namespace embree
|
||||
/// Output Operators
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
template<typename T> __forceinline embree_ostream operator<<(embree_ostream cout, const Vec4<T>& a) {
|
||||
template<typename T> inline std::ostream& operator<<(std::ostream& cout, const Vec4<T>& a) {
|
||||
return cout << "(" << a.x << ", " << a.y << ", " << a.z << ", " << a.w << ")";
|
||||
}
|
||||
|
||||
@@ -205,7 +206,7 @@ namespace embree
|
||||
/// SSE / AVX / MIC specializations
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
#if defined(__SSE__) || defined(__ARM_NEON)
|
||||
#if defined __SSE__
|
||||
#include "../simd/sse.h"
|
||||
#endif
|
||||
|
||||
@@ -219,25 +220,40 @@ namespace embree
|
||||
|
||||
namespace embree
|
||||
{
|
||||
template<> __forceinline Vec4<float>::Vec4( const Vec3fx& a ) { x = a.x; y = a.y; z = a.z; w = a.w; }
|
||||
template<> __forceinline Vec4<float>::Vec4( const Vec3fa& a ) { x = a.x; y = a.y; z = a.z; w = a.w; }
|
||||
|
||||
#if defined(__AVX__)
|
||||
template<> __forceinline Vec4<vfloat4>::Vec4( const Vec3fx& a ) {
|
||||
template<> __forceinline Vec4<vfloat4>::Vec4( const Vec3fa& a ) {
|
||||
x = a.x; y = a.y; z = a.z; w = a.w;
|
||||
}
|
||||
#elif defined(__SSE__) || defined(__ARM_NEON)
|
||||
template<> __forceinline Vec4<vfloat4>::Vec4( const Vec3fx& a ) {
|
||||
const vfloat4 v = vfloat4(a.m128); x = shuffle<0,0,0,0>(v); y = shuffle<1,1,1,1>(v); z = shuffle<2,2,2,2>(v); w = shuffle<3,3,3,3>(v);
|
||||
#elif defined(__SSE__)
|
||||
template<> __forceinline Vec4<vfloat4>::Vec4( const Vec3fa& a ) {
|
||||
const vfloat4 v = vfloat4(a); x = shuffle<0,0,0,0>(v); y = shuffle<1,1,1,1>(v); z = shuffle<2,2,2,2>(v); w = shuffle<3,3,3,3>(v);
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__SSE__)
|
||||
__forceinline Vec4<vfloat4> broadcast4f( const Vec4<vfloat4>& a, const size_t k ) {
|
||||
return Vec4<vfloat4>(vfloat4::broadcast(&a.x[k]), vfloat4::broadcast(&a.y[k]), vfloat4::broadcast(&a.z[k]), vfloat4::broadcast(&a.w[k]));
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__AVX__)
|
||||
template<> __forceinline Vec4<vfloat8>::Vec4( const Vec3fx& a ) {
|
||||
template<> __forceinline Vec4<vfloat8>::Vec4( const Vec3fa& a ) {
|
||||
x = a.x; y = a.y; z = a.z; w = a.w;
|
||||
}
|
||||
__forceinline Vec4<vfloat4> broadcast4f( const Vec4<vfloat8>& a, const size_t k ) {
|
||||
return Vec4<vfloat4>(vfloat4::broadcast(&a.x[k]), vfloat4::broadcast(&a.y[k]), vfloat4::broadcast(&a.z[k]), vfloat4::broadcast(&a.w[k]));
|
||||
}
|
||||
__forceinline Vec4<vfloat8> broadcast8f( const Vec4<vfloat4>& a, const size_t k ) {
|
||||
return Vec4<vfloat8>(vfloat8::broadcast(&a.x[k]), vfloat8::broadcast(&a.y[k]), vfloat8::broadcast(&a.z[k]), vfloat8::broadcast(&a.w[k]));
|
||||
}
|
||||
__forceinline Vec4<vfloat8> broadcast8f( const Vec4<vfloat8>& a, const size_t k ) {
|
||||
return Vec4<vfloat8>(vfloat8::broadcast(&a.x[k]), vfloat8::broadcast(&a.y[k]), vfloat8::broadcast(&a.z[k]), vfloat8::broadcast(&a.w[k]));
|
||||
}
|
||||
#endif
|
||||
|
||||
#if defined(__AVX512F__)
|
||||
template<> __forceinline Vec4<vfloat16>::Vec4( const Vec3fx& a ) : x(a.x), y(a.y), z(a.z), w(a.w) {}
|
||||
template<> __forceinline Vec4<vfloat16>::Vec4( const Vec3fa& a ) : x(a.x), y(a.y), z(a.z), w(a.w) {}
|
||||
#endif
|
||||
}
|
||||
|
||||
@@ -1,11 +1,24 @@
|
||||
## Copyright 2009-2021 Intel Corporation
|
||||
## SPDX-License-Identifier: Apache-2.0
|
||||
## ======================================================================== ##
|
||||
## Copyright 2009-2018 Intel Corporation ##
|
||||
## ##
|
||||
## Licensed under the Apache License, Version 2.0 (the "License"); ##
|
||||
## you may not use this file except in compliance with the License. ##
|
||||
## You may obtain a copy of the License at ##
|
||||
## ##
|
||||
## http://www.apache.org/licenses/LICENSE-2.0 ##
|
||||
## ##
|
||||
## Unless required by applicable law or agreed to in writing, software ##
|
||||
## distributed under the License is distributed on an "AS IS" BASIS, ##
|
||||
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ##
|
||||
## See the License for the specific language governing permissions and ##
|
||||
## limitations under the License. ##
|
||||
## ======================================================================== ##
|
||||
|
||||
ADD_LIBRARY(simd STATIC sse.cpp)
|
||||
SET_PROPERTY(TARGET simd PROPERTY FOLDER common)
|
||||
SET_PROPERTY(TARGET simd APPEND PROPERTY COMPILE_FLAGS " ${FLAGS_LOWEST}")
|
||||
|
||||
IF (EMBREE_STATIC_LIB)
|
||||
INSTALL(TARGETS simd EXPORT simd-targets ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}" COMPONENT devel)
|
||||
INSTALL(EXPORT simd-targets DESTINATION "${EMBREE_CMAKEEXPORT_DIR}" COMPONENT devel)
|
||||
INSTALL(TARGETS simd ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT devel)
|
||||
SIGN_TARGET(simd)
|
||||
ENDIF()
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@@ -1,85 +0,0 @@
|
||||
// Copyright 2009-2020 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#pragma once
|
||||
|
||||
/* Make precision match SSE, at the cost of some performance */
|
||||
#if !defined(__aarch64__)
|
||||
# define SSE2NEON_PRECISE_DIV 1
|
||||
# define SSE2NEON_PRECISE_SQRT 1
|
||||
#endif
|
||||
|
||||
#include "sse2neon.h"
|
||||
|
||||
__forceinline __m128 _mm_abs_ps(__m128 a) { return vabsq_f32(a); }
|
||||
|
||||
__forceinline __m128 _mm_fmadd_ps (__m128 a, __m128 b, __m128 c) { return vfmaq_f32(c, a, b); }
|
||||
__forceinline __m128 _mm_fnmadd_ps(__m128 a, __m128 b, __m128 c) { return vfmsq_f32(c, a, b); }
|
||||
__forceinline __m128 _mm_fnmsub_ps(__m128 a, __m128 b, __m128 c) { return vnegq_f32(vfmaq_f32(c, a, b)); }
|
||||
__forceinline __m128 _mm_fmsub_ps (__m128 a, __m128 b, __m128 c) { return vnegq_f32(vfmsq_f32(c, a, b)); }
|
||||
|
||||
__forceinline __m128 _mm_broadcast_ss (float const * mem_addr)
|
||||
{
|
||||
return vdupq_n_f32(*mem_addr);
|
||||
}
|
||||
|
||||
// AVX2 emulation leverages Intel FMA defs above. Include after them.
|
||||
#include "avx2neon.h"
|
||||
|
||||
/* Dummy defines for floating point control */
|
||||
#define _MM_MASK_MASK 0x1f80
|
||||
#define _MM_MASK_DIV_ZERO 0x200
|
||||
// #define _MM_FLUSH_ZERO_ON 0x8000
|
||||
#define _MM_MASK_DENORM 0x100
|
||||
#define _MM_SET_EXCEPTION_MASK(x)
|
||||
// #define _MM_SET_FLUSH_ZERO_MODE(x)
|
||||
|
||||
__forceinline int _mm_getcsr()
|
||||
{
|
||||
return 0;
|
||||
}
|
||||
|
||||
__forceinline void _mm_mfence()
|
||||
{
|
||||
__sync_synchronize();
|
||||
}
|
||||
|
||||
__forceinline __m128i _mm_load4epu8_epi32(__m128i *ptr)
|
||||
{
|
||||
uint8x8_t t0 = vld1_u8((uint8_t*)ptr);
|
||||
uint16x8_t t1 = vmovl_u8(t0);
|
||||
uint32x4_t t2 = vmovl_u16(vget_low_u16(t1));
|
||||
return vreinterpretq_s32_u32(t2);
|
||||
}
|
||||
|
||||
__forceinline __m128i _mm_load4epu16_epi32(__m128i *ptr)
|
||||
{
|
||||
uint16x8_t t0 = vld1q_u16((uint16_t*)ptr);
|
||||
uint32x4_t t1 = vmovl_u16(vget_low_u16(t0));
|
||||
return vreinterpretq_s32_u32(t1);
|
||||
}
|
||||
|
||||
__forceinline __m128i _mm_load4epi8_f32(__m128i *ptr)
|
||||
{
|
||||
int8x8_t t0 = vld1_s8((int8_t*)ptr);
|
||||
int16x8_t t1 = vmovl_s8(t0);
|
||||
int32x4_t t2 = vmovl_s16(vget_low_s16(t1));
|
||||
float32x4_t t3 = vcvtq_f32_s32(t2);
|
||||
return vreinterpretq_s32_f32(t3);
|
||||
}
|
||||
|
||||
__forceinline __m128i _mm_load4epu8_f32(__m128i *ptr)
|
||||
{
|
||||
uint8x8_t t0 = vld1_u8((uint8_t*)ptr);
|
||||
uint16x8_t t1 = vmovl_u8(t0);
|
||||
uint32x4_t t2 = vmovl_u16(vget_low_u16(t1));
|
||||
return vreinterpretq_s32_u32(t2);
|
||||
}
|
||||
|
||||
__forceinline __m128i _mm_load4epi16_f32(__m128i *ptr)
|
||||
{
|
||||
int16x8_t t0 = vld1q_s16((int16_t*)ptr);
|
||||
int32x4_t t1 = vmovl_s16(vget_low_s16(t0));
|
||||
float32x4_t t2 = vcvtq_f32_s32(t1);
|
||||
return vreinterpretq_s32_f32(t2);
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
+15
-2
@@ -1,5 +1,18 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
|
||||
+15
-3
@@ -1,12 +1,24 @@
|
||||
// Copyright 2009-2021 Intel Corporation
|
||||
// SPDX-License-Identifier: Apache-2.0
|
||||
// ======================================================================== //
|
||||
// Copyright 2009-2018 Intel Corporation //
|
||||
// //
|
||||
// Licensed under the Apache License, Version 2.0 (the "License"); //
|
||||
// you may not use this file except in compliance with the License. //
|
||||
// You may obtain a copy of the License at //
|
||||
// //
|
||||
// http://www.apache.org/licenses/LICENSE-2.0 //
|
||||
// //
|
||||
// Unless required by applicable law or agreed to in writing, software //
|
||||
// distributed under the License is distributed on an "AS IS" BASIS, //
|
||||
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
|
||||
// See the License for the specific language governing permissions and //
|
||||
// limitations under the License. //
|
||||
// ======================================================================== //
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "../sys/platform.h"
|
||||
#include "../sys/intrinsics.h"
|
||||
#include "../math/constants.h"
|
||||
#include "../sys/alloc.h"
|
||||
#include "varying.h"
|
||||
|
||||
#include "vboolf16_avx512.h"
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user