Fix KleidiAI FP16 #3769
Closed
Fix KleidiAI FP16 #3769
Conversation
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
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
force-pushed
the
export-D70606808
branch
2 times, most recently
from
3cccf41 to
426d983
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 5, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 5, 2025
Summary: Pull Request resolved: pytorch#3769 X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 Differential Revision: D70606808
Nicoshev
force-pushed
the
export-D70606808
branch
2 times, most recently
from
9df54cc to
bf5b9aa
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 5, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 5, 2025
Summary: Pull Request resolved: pytorch#3769 X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 Differential Revision: D70606808
Nicoshev
force-pushed
the
export-D70606808
branch
2 times, most recently
from
eb22db6 to
2400c83
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 10, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
force-pushed
the
export-D70606808
branch
from
2400c83 to
7d3cdf6
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 10, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 10, 2025
Summary: Pull Request resolved: pytorch#3769 X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
Nicoshev
force-pushed
the
export-D70606808
branch
2 times, most recently
from
c2d560a to
eb3c685
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 11, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 11, 2025
Summary: Pull Request resolved: pytorch#3769 X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
Nicoshev
force-pushed
the
export-D70606808
branch
from
eb3c685 to
111bf81
Compare
Nicoshev
added a commit
to Nicoshev/FBGEMM
that referenced
this pull request
Mar 11, 2025
Summary: X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
Nicoshev
force-pushed
the
export-D70606808
branch
from
111bf81 to
27d6cbe
Compare
Summary: Pull Request resolved: pytorch#3769 X-link: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808
|
This pull request was exported from Phabricator. Differential Revision: D70606808 |
Nicoshev
force-pushed
the
export-D70606808
branch
from
27d6cbe to
fdc8224
Compare
|
This pull request has been merged in d17faaa. |
liligwu
pushed a commit
to ROCm/FBGEMM
that referenced
this pull request
Mar 14, 2025
Summary: Pull Request resolved: pytorch#3769 X-link: https://github.com/facebookresearch/FBGEMM/pull/849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808 fbshipit-source-id: b5ada2519eee65e826fa5647967a23db30911f23
q10
pushed a commit
to q10/FBGEMM
that referenced
this pull request
Apr 10, 2025
Summary: X-link: pytorch#3769 Pull Request resolved: facebookresearch/FBGEMM#849 FP32Test was failing when KleidiAI was enabled. It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32 PackedGemmMatrixFP16 constructors were moved to a .cc file, compiled with the rest of fbgemm. This ensures the KleidiAI flag is set when compiling such code. Previously, consumers of the library would include FbgemmPackMatrixB.h from their .cpp files. Because the KleidiAI flag was not set when compiling their code, the intended case for the float16 was never run. Additionally, we have ingested a change in KleidiAI's inline assembly, which handles nan beta values as 0. Reviewed By: psaab Differential Revision: D70606808 fbshipit-source-id: b5ada2519eee65e826fa5647967a23db30911f23
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Summary:
FP32Test was failing when KleidiAI was enabled.
It turns out FPCommon.h and PackMatrixB.h were not acconditioned to handle using kleidAI for FP16 and not for FP32
Differential Revision: D70606808