Cherry pick/pr11143 shared expert fusion v0#16270
Closed
leslie-fang25 wants to merge 6 commits into
Closed
Conversation
Signed-off-by: leslief <leslief@nvidia.com>
Signed-off-by: leslief <leslief@nvidia.com>
Enhance the check to enable this fusion path Signed-off-by: leslief <leslief@nvidia.com>
Signed-off-by: leslief <leslief@nvidia.com>
…te_dsl to avoid IMA Signed-off-by: leslief <leslief@nvidia.com>
Add a --nsys flag that brackets the measured MoE forward with cudaProfilerStart/Stop (after warmup) and wraps only the forward with a 'bench_moe.measured' NVTX range, so an external 'nsys profile -c cudaProfilerApi --capture-range-end stop' run captures just the measured forward (no autotune / warmup / graph capture). - timing/nsys.py: guarded _NsysProfiler (cudaProfilerStart/Stop) + measured_range NVTX context manager (reuses tensorrt_llm._utils.nvtx_range; null ctx when off). - eager & cuda_graph timing paths: NVTX wraps only the forward (excludes the L2-flush memset, matching the CUDA-event latency window); cudaProfilerStart is placed after warmup; for CUDA graph the NVTX range is captured into the graph. - worker: nsys and CUPTI kernel breakdown are mutually exclusive (both use CUPTI); --nsys disables the kernel breakdown with a warning. - report an nsys_capture marker in instrumentation / dashboard columns. - run_nsys.sh: single-GPU and multi-GPU (nsys after mpirun, per-rank -o) recipes. --nsys off is a no-op: latency measurement and existing outputs are unchanged. Signed-off-by: leslief <leslief@nvidia.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
@coderabbitai summary
Description
Test Coverage
PR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
If PR introduces API changes, an appropriate PR label is added - either
api-compatibleorapi-breaking. Forapi-breaking, includeBREAKINGin the PR title.Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
GitHub Bot Help
To see a list of available CI bot commands, please comment
/bot help.