Repository navigation
gfx1250: use hipblaslt device alpha for NVFP4 GEMM - #766
Open
matthiasdiener wants to merge 7 commits into
Open
matthiasdiener wants to merge 7 commits into
matthiasdiener wants to merge 7 commits into
Conversation
1 task done
matthiasdiener
changed the title
matthiasdiener
marked this pull request as ready for review
October 20, 2026 20:47
matthiasdiener
requested review from
Micky774,
ipanfilo,
wangye805 and
wenchenvincent
as code owners
October 20, 2026 20:47
ipanfilo
reviewed
Oct 27, 2026
| const float fp8_max = te_fp8_fnuz() ? 240.0f : 448.0f; | ||
| const float factor_inv = 1.0f / (fp4_max * fp4_max * fp8_max * fp8_max); | ||
| // 4over6 tensors carry an E4M3 bound of 256, not 448. | ||
| const float fp8_max_A = te_fp8_fnuz() ? 240.0f : static_cast<float>(inputA.nvfp4_e4m3_max); |
Collaborator
There was a problem hiding this comment.
NVFP4 is not supported on gfx942 so do not need to check for fnuz
| list(APPEND transformer_engine_LINKER_LIBS hiprtc) | ||
| find_package(hipblaslt) | ||
| list(APPEND transformer_engine_LINKER_LIBS roc::hipblaslt) | ||
| # Device scalar alpha (ROCm/rocm-libraries#13482) is not tied to a hipBLASLt version. |
Collaborator
There was a problem hiding this comment.
Not tied yet? If it is not part of ROCm release it only works as developer feature
Contributor
Author
There was a problem hiding this comment.
Rephrased it to saying that it is not in a hipblaslt release yet in 190401a. In this case, the code from this PR will fall back to the previous implementation.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
Makes use of ROCm/rocm-libraries#13482, but falls back transparently if device alpha support is not available.
Performance results: https://github.com/AMD-ROCm-Internal/frameworks-internal/issues/389#issuecomment-6299513129
Type of change
Changes
Please list the changes introduced in this PR:
Checklist: