Allow all RDNA2 archs to use sdot4 intrinsic#8629
Merged
JohannesGaessler merged 1 commit intoggml-org:masterfrom Jul 23, 2024
Merged
Allow all RDNA2 archs to use sdot4 intrinsic#8629JohannesGaessler merged 1 commit intoggml-org:masterfrom
JohannesGaessler merged 1 commit intoggml-org:masterfrom
Conversation
The check gating the use of `__builtin_amdgc_sdot4` specifically checks for gfx1030. This causes a severe perf regression for anything gfx103? that's not gfx1030 and not using `HSA_OVERRIDE_GFX_VERSION` (if you've built ROCm to support it). We already have a generic RDNA2 define, let's use it.
Contributor
Author
|
Original discussion (including repro) is at lamikr/rocm_sdk_builder#114 (comment) . That incorrectly stated that there was no arch-specific logic since I missed it on first read. |
JohannesGaessler
approved these changes
Jul 23, 2024
Contributor
JohannesGaessler
left a comment
JohannesGaessler
left a comment
There was a problem hiding this comment.
I am not particularly knowledgeable when it comes to ROCm but this looks correct to me.
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this pull request
Jul 27, 2024
The check gating the use of `__builtin_amdgc_sdot4` specifically checks for gfx1030. This causes a severe perf regression for anything gfx103? that's not gfx1030 and not using `HSA_OVERRIDE_GFX_VERSION` (if you've built ROCm to support it). We already have a generic RDNA2 define, let's use it.
Seunghhon
pushed a commit
to Seunghhon/llama.cpp
that referenced
this pull request
Apr 26, 2026
The check gating the use of `__builtin_amdgc_sdot4` specifically checks for gfx1030. This causes a severe perf regression for anything gfx103? that's not gfx1030 and not using `HSA_OVERRIDE_GFX_VERSION` (if you've built ROCm to support it). We already have a generic RDNA2 define, let's use it.
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
2 participants
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.
The check gating the use of
__builtin_amdgc_sdot4specifically checks for gfx1030. This causes a severe perf regression for anything gfx103? that's not gfx1030 and not usingHSA_OVERRIDE_GFX_VERSION(if you've built ROCm to support it). We already have a generic RDNA2 define, let's use it.With this change my custom ROCm build that includes gfx1036 support (and uses the gfx1030 kernels) performs identically with or without
HSA_OVERRIDE_GFX_VERSION.