vulkan/cuda: fix topk_moe with exp_probs_b #18071
Open
+117
−35
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
jeffbolznv
requested review from
0cc4m,
JohannesGaessler and
am17an
and removed request for
0cc4m and
ggerganov
jeffbolznv
commented
Dec 15, 2025
| } | ||
| } | ||
|
|
||
| bool ggml_cuda_should_use_topk_moe(const ggml_tensor * softmax, const ggml_tensor * weights, const ggml_tensor * clamp) { |
Collaborator
Author
There was a problem hiding this comment.
I also noticed that clamp is always nullptr. I didn't attempt to fix this.
I updated test_topk_moe to more closely match llm_graph_context::build_moe_ffn and added coverage for exp_probs_b and some other missing combinations. This exposed a bug in both CUDA and Vulkan backends where they were assuming the input to argsort and the input to get_rows are the same. I'd like to optimize this graph in another change, but for now just get it functional. CUDA also had a bug where it got n_experts from the wrong place, leading to GGML_ASSERT failures in some of the new tests.
jeffbolznv
force-pushed
the
topk_moe_tests
branch
from
21e0267 to
e262bf9
Compare
github-actions
bot
added
testing
am17an
approved these changes
Dec 16, 2025
Collaborator
am17an
left a comment
am17an
left a comment
There was a problem hiding this comment.
CUDA changes look fine. Thanks for fixing!
ggerganov
approved these changes
Dec 16, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
I updated test_topk_moe to more closely match llm_graph_context::build_moe_ffn and added coverage for exp_probs_b and some other missing combinations. This exposed a bug in both CUDA and Vulkan backends where they were assuming the input to argsort and the input to get_rows are the same. I'd like to optimize this graph in another change, but for now just get it functional.
CUDA also had a bug where it got n_experts from the wrong place, leading to GGML_ASSERT failures in some of the new tests.