[OpenBLAS_jll] Update to new build with BFloat16 kernels #53059
Merged
Conversation
This file contains 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
giordano
added
building
|
We also have a new SVE patch in Yggdrasil: https://github.com/JuliaPackaging/Yggdrasil/blob/master/O/OpenBLAS/OpenBLAS%400.3.26/bundled/patches/90-darwin-sve.patch that I don't see included here. It is needed for the Apple aarch64 BFloat enabled builds. |
The new build also includes patches to * improve multithreading * fix builds of BFloat16 kernels on AVX512BF16
…alues This would have caught an old upstream bug in OpenBLAS, which was later fixed in v0.3.14.
|
Oops, added, thanks! |
ViralBShah
approved these changes
Jan 26, 2024
|
Once this is in the nightly, how do I make BLAS calls? Is there a bgemm out there. I'm pretty sure that @timothyaldendavis cares too |
|
Yeah you can just |
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Outdated suggestions cannot be applied.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.
This PR also
deps/patches/neoverse-generic-kernels.patch) not needed anymore for an old bug fixed upstream in OpenBLAS. This results in ~5x speedup in the computation ofBLAS.nrm2(and henceLinearAlgebra.normfor vectors longer thanLinearAlgebra.NRM2_CUTOFF(== 32) elements) when the neoversen1 kernels are used, e.g. by default on all Apple Silicon CPUsCorresponding PR in Yggdrasil: JuliaPackaging/Yggdrasil#7202. CC: @imciner2.