fix cp inference#4619
Merged
Merged
Conversation
Contributor
There was a problem hiding this comment.
Pull request overview
This PR addresses a context-parallel (CP) inference correctness issue in TurboMind’s LLaMA unified attention path by explicitly initializing the CP partial (M, L) reduction buffer to avoid stale values (e.g., when CTAs early-exit for finished sequences in async mode).
Changes:
- Add a CUDA helper (
invokeFillNegInfML) to initialize(M, L)pairs to(-inf, 0)before attention/reduction. - Call this initializer during
UnifiedAttentionLayer::Run(BatchOp::kPrepare, ...)when CP is enabled. - Document the intent of the initializer in
cp_utils.h.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| src/turbomind/models/llama/unified_attention_layer.cc | Calls the new CP partial buffer initializer during batch prepare. |
| src/turbomind/kernels/attention/cp_utils.h | Declares invokeFillNegInfML with a new explanatory comment. |
| src/turbomind/kernels/attention/cp_utils.cu | Implements the CUDA kernel + launcher to fill (M, L) pairs with (-inf, 0). |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+29
to
+38
| void invokeFillNegInfML(float* data, size_t n_pairs, cudaStream_t stream) | ||
| { | ||
| if (n_pairs == 0) { | ||
| return; | ||
| } | ||
| constexpr int block = 256; | ||
| const size_t n_quads = n_pairs >> 1; | ||
| const size_t grid = (n_quads + block - 1) / block; | ||
| FillNegInfMLKernel<<<grid, block, 0, stream>>>(reinterpret_cast<float4*>(data), n_quads); | ||
| TM_CUDA_CHECK(cudaGetLastError()); |
Comment on lines
+23
to
+26
| // Fill an array of (M, L) pairs with (-inf, 0). Used to initialize this rank's | ||
| // slot in `partial_ML` before attention, so that reduce treats slots left | ||
| // untouched by early-exiting CTAs (e.g. finished sequences in async mode) as | ||
| // no-contribution rather than reading stale data from previous batches. |
Comment on lines
228
to
+232
| auto& d = data_.at(phase); | ||
| Clear(tmp_attn_.slice(0, d->decode.n + d->prefill.q_sum)); | ||
| Clear(split_cnt_); | ||
| if (engine_param_.attn_cp_size > 1) { | ||
| invokeFillNegInfML(partial_ML_.data(), partial_ML_.size() / 2, core::Context::stream().handle()); |
lvhan028
approved these changes
May 28, 2026
lzhangzz
approved these changes
May 28, 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
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.
Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.
Motivation
Please describe the motivation of this PR and the goal you want to achieve through this PR.
Modification
Please briefly describe what modification is made in this PR.
BC-breaking (Optional)
Does the modification introduce changes that break the backward-compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.
Use cases (Optional)
If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.
Checklist