[CI] Update sleef submodule to v3.8 #151955
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/151955
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 4 PendingAs of commit 697b3f0 with merge base 2f74cff ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
Ok, we can't update to 3.9, as this moves base compiler compatibility to C++20 |
malfet
changed the title
malfet
changed the title
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
|
@pytorchbot merge -f "All builds passed, remaining tests seems irrelevant" |
|
The merge job was canceled or timed out. This most often happen if two merge requests were issued for the same PR, or if merge job was waiting for more than 6 hours for tests to finish. In later case, please do not hesitate to reissue the merge command |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
As reported in #149292, according to manual, `vfmsq_f32` implements `c - a * b` rather than `a * b - c`, so it's call must be prefixed with `vnegq_f32` Also, adjust the tests to use OpMath for FMA computation to avoid accuracy error accumulation due to non-fused multiply-and-add over lower precision dtypes Note that `Vectorized::fmsub` is not currently instantiated anywhere, so it could safely remain broken TODO: - Enable C++ testing on MacOS and/or aarch64 platforms (right now Mac tests are build without C++ tests) Fixes #149292 Pull Request resolved: #152075 Approved by: https://github.com/swolchok ghstack dependencies: #151955
As reported in pytorch#149292, according to manual, `vfmsq_f32` implements `c - a * b` rather than `a * b - c`, so it's call must be prefixed with `vnegq_f32` Also, adjust the tests to use OpMath for FMA computation to avoid accuracy error accumulation due to non-fused multiply-and-add over lower precision dtypes Note that `Vectorized::fmsub` is not currently instantiated anywhere, so it could safely remain broken TODO: - Enable C++ testing on MacOS and/or aarch64 platforms (right now Mac tests are build without C++ tests) Fixes pytorch#149292 Pull Request resolved: pytorch#152075 Approved by: https://github.com/swolchok ghstack dependencies: pytorch#151955
As reported in #149292, according to manual, `vfmsq_f32` implements `c - a * b` rather than `a * b - c`, so it's call must be prefixed with `vnegq_f32` Also, adjust the tests to use OpMath for FMA computation to avoid accuracy error accumulation due to non-fused multiply-and-add over lower precision dtypes Note that `Vectorized::fmsub` is not currently instantiated anywhere, so it could safely remain broken TODO: - Enable C++ testing on MacOS and/or aarch64 platforms (right now Mac tests are build without C++ tests) Fixes #149292 Pull Request resolved: #152075 Approved by: https://github.com/swolchok ghstack dependencies: #151955 (cherry picked from commit 2ea8653)
[vec128] Fix fmsub NEON defintion (#152075) As reported in #149292, according to manual, `vfmsq_f32` implements `c - a * b` rather than `a * b - c`, so it's call must be prefixed with `vnegq_f32` Also, adjust the tests to use OpMath for FMA computation to avoid accuracy error accumulation due to non-fused multiply-and-add over lower precision dtypes Note that `Vectorized::fmsub` is not currently instantiated anywhere, so it could safely remain broken TODO: - Enable C++ testing on MacOS and/or aarch64 platforms (right now Mac tests are build without C++ tests) Fixes #149292 Pull Request resolved: #152075 Approved by: https://github.com/swolchok ghstack dependencies: #151955 (cherry picked from commit 2ea8653) Co-authored-by: Nikita Shulga <nshulga@meta.com>
5 participants
Stack from ghstack (oldest at bottom):
Should help with RISC-V cross-compilation.
3.9.0 migration is blocked by sleef project switching to C++20