model : fix k-pool indexer softmax gridDim.y overflow - #11
Open
AIalliAI wants to merge 2 commits into
Open
Conversation
Reshape the k-pool gate logits before softmax so n_new does not map to gridDim.y, which is capped at 65535 on CUDA. Fixes prompt processing abort at n_kv >= 262144 reported in ggml-org#27773. Assisted-by: Cursor
Assisted-by: Cursor
Author
|
Cross-ref: ggml-org#27773, ggml-org#27752, ggml-org#27754, ggml-org#27917, #9. |
This was referenced Sep 11, 2026
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Overview
Reshape k-pool gate logits before
ggml_soft_maxson_newdoes not map togridDim.y, which is capped at 65535 on CUDA. During k-pool indexer prefill atn_kv >= 262144, the gate softmax is shaped[kpool, head_dim, n_new]and trips the launch atkpool = 4.Reshape to
[kpool, head_dim * n_new]before softmax, reshape back after. Same pattern as the comment inqwen4exp.cpp.Stacked on ggml-org#27773.
Additional information
Related: ggml-org#27754, unslothai#214, ggml-org#27901, ggml-org#28175. Separate from #9 (tensor naming).
Requirements