Adsk Contrib - Add support for ARM Neon intrinsics and Universal builds#1775
Closed
cedrik-fuoco-adsk wants to merge 84 commits into
Closed
Conversation
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…s to fix issues with NaN handling. Fixing issues in the unit tests for some tests. For some tests, the neon implementation is matching the C++ implementation instead of the SSE2 implementation. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Now check for OCIO_USE_SSE and HAVE_NEON before integrating sse2neon to the build. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…d on Apple and might interfere with optimization if not careful. Updated CheckSupportSSE2.cmake and adding some checks in SSE.h to support universal build. Added comments and new define (USING_INTEL_SSE, USING_ARM_NEON and USING_CPP) in LogOpCPU_tests which should help with understand what is going on. Checking for SSE2 and ARM Neon only when OCIO_USE_SSE is ON. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…SE_SSE (they are sync up). Will replace OCIO_USE_SSE eventually. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Reverted defined changes in LogOpCPU_tests Updated ocio.bat with OCIO_USE_SIMD Minors comments changes Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
documentation tweak Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <105517825+cedrik-fuoco-adsk@users.noreply.github.com>
Removing Findsse2neon as it is not needed Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
doug-walker
changed the title
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…s to fix issues with NaN handling. Fixing issues in the unit tests for some tests. For some tests, the neon implementation is matching the C++ implementation instead of the SSE2 implementation. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Now check for OCIO_USE_SSE and HAVE_NEON before integrating sse2neon to the build. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…d on Apple and might interfere with optimization if not careful. Updated CheckSupportSSE2.cmake and adding some checks in SSE.h to support universal build. Added comments and new define (USING_INTEL_SSE, USING_ARM_NEON and USING_CPP) in LogOpCPU_tests which should help with understand what is going on. Checking for SSE2 and ARM Neon only when OCIO_USE_SSE is ON. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…SE_SSE (they are sync up). Will replace OCIO_USE_SSE eventually. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Reverted defined changes in LogOpCPU_tests Updated ocio.bat with OCIO_USE_SIMD Minors comments changes Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
documentation tweak Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Removing Findsse2neon as it is not needed Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…d on Apple and might interfere with optimization if not careful. Updated CheckSupportSSE2.cmake and adding some checks in SSE.h to support universal build. Added comments and new define (USING_INTEL_SSE, USING_ARM_NEON and USING_CPP) in LogOpCPU_tests which should help with understand what is going on. Checking for SSE2 and ARM Neon only when OCIO_USE_SSE is ON. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…SE_SSE (they are sync up). Will replace OCIO_USE_SSE eventually. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Reverted defined changes in LogOpCPU_tests Updated ocio.bat with OCIO_USE_SIMD Minors comments changes Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Removing Findsse2neon as it is not needed Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Removing ocio_check_dependency_version.cmake as it is not needed anymore. Changed how we detect if the installed OpenEXR and OpenImageIO are compatible with OCIO. It should be more robust now. Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
…/github.com/autodesk-forks/OpenColorIO into adsk_contrib/add-support-for-neon-intrinsic Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
| set(is_OpenEXR_VERSION_valid FALSE) | ||
| # Check for compatibility between OpenEXR and OpenImageIO. | ||
| # Will set is_OpenEXR_VERSION_valid to TRUE if valid. | ||
| include(CheckForOpenEXRCompatibility) |
Contributor
Author
There was a problem hiding this comment.
Changed how we check the version of the OpenEXR used by OpenImageIO. This should be more robust and won't create any unwanted target or variables.
| @@ -1,38 +0,0 @@ | |||
| # SPDX-License-Identifier: BSD-3-Clause | |||
Contributor
Author
There was a problem hiding this comment.
removed in favour of CheckForOpenEXRCompatibility.cmake
Contributor
Author
|
The branch is now rebased and conflicts have been resolved. |
remia
reviewed
Mar 26, 2023
Collaborator
remia
left a comment
remia
left a comment
There was a problem hiding this comment.
Looks good to me @cedrik-fuoco-adsk, thank you for adding more comments to the SSE header. Have to say Rosetta is surprisingly good at handling the SSE x86_64 binary. I guess if we want to further optimise we would have to start more in depth profiling but the improvements we get seem already quite large as is.
…rge issue) Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
Signed-off-by: Cédrik Fuoco <cedrik.fuoco@autodesk.com>
remia
approved these changes
Apr 3, 2023
|
Hi all, I'd love to see these changes merged -- any remaining blockers? Thank you for your work on this, @cedrik-fuoco-adsk! |
This was referenced Jun 19, 2023
michdolan
previously approved these changes
Aug 28, 2023
Collaborator
|
Closing since #1828 replaces this PR |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR is about Add support for ARM Neon intrinsics #1753.
Note that at least another commit will be needed once Adsk contrib - Add support for minimum and recommended versions for dependencies is merged into the main branch.
Some performance results with a MacBook M1 (using a custom heavy_transform.clf):
