[SYCL][CUDA] Swapped bf16 builtins with inline asm. #10695
Merged
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
Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
jchlanda
approved these changes
Aug 7, 2023
cperkinsintel
approved these changes
Aug 8, 2023
ldrumm
reviewed
Aug 8, 2023
Improved formating. Signed-off-by: JackAKirk <jack.kirk@codeplay.com>
ldrumm
approved these changes
Aug 8, 2023
|
@JackAKirk, please, merge tip of the |
The issue is not related to the patch and reported in #10460. |
steffenlarsen
pushed a commit
that referenced
this pull request
Sep 5, 2023
This reverts commit bfc1956. With workaround landed in #10695, we should be able to reland the community commit llvm/llvm-project@250f2bb now. This revert the revert to reland it.
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Outdated suggestions cannot be applied.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.
A recent pulldown had to revert this commit llvm/llvm-project@250f2bb
The reason for this is that SYCL runtime uses
uint16_tas the storage type forbfloat16and this commit uses__bf16instead.Eventually we plan to fully support the
__bf16type in SYCL runtime: see this discussion: #10457.But until
__bf16support is added it is easy for us to instead use inline asm operating onuint16_tin the nvptx backend.This PR makes this change. When this PR is merged llvm/llvm-project@250f2bb can be safely pulled down.