[None][chore] AutoDeploy: neat disablement of transforms in pipeline #7736
Conversation
lucaslie
requested review from
galagam and
nzmora-nvidia
and removed request for
nvchenghaoz
|
/bot run |
|
PR_Github #18644 [ run ] triggered by Bot |
📝 WalkthroughWalkthroughThe default auto-deploy config updates post_load_fusion entries to explicitly include enabled: false for fuse_gemms, fuse_fp4_gemms, fuse_fp8_gemms, and fuse_moe. Previously commented-out blocks are now concrete entries. fuse_rmsnorm gains inline TODO comments without behavior change. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant App as App
participant Config as default.yaml
participant Loader as Config Loader
participant Fusion as PostLoadFusionManager
App->>Loader: Initialize
Loader->>Config: Read settings
Config-->>Loader: post_load_fusion entries
Loader->>Fusion: Configure transforms
loop For each transform
Fusion->>Fusion: Check enabled flag
alt enabled == true
Fusion->>App: Apply transform
else enabled == false
Fusion-->>App: Skip transform
end
end
note over Fusion,App: Explicitly disabled by default: fuse_gemms, fuse_fp4_gemms, fuse_fp8_gemms, fuse_moe
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Pre-merge checks and finishing touches✅ Passed checks (3 passed)
✨ Finishing touches🧪 Generate unit tests
Tip 👮 Agentic pre-merge checks are now available in preview!Pro plan users can now enable pre-merge checks in their settings to enforce checklists before merging PRs.
Please see the documentation for more information. Example: reviews:
pre_merge_checks:
custom_checks:
- name: "Undocumented Breaking Changes"
mode: "warning"
instructions: |
Pass/fail criteria: All breaking changes to public APIs, CLI flags, environment variables, configuration keys, database schemas, or HTTP/GraphQL endpoints must be documented in the "Breaking Change" section of the PR description and in CHANGELOG.md. Exclude purely internal or private changes (e.g., code not exported from package entry points or explicitly marked as internal).Please share your feedback with us on this Discord post. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (3)
tensorrt_llm/_torch/auto_deploy/config/default.yaml (3)
80-88: Consider a single feature flag to re-enable GEMM fusions for perf testing.Disabling three GEMM fusion variants is clear, but it’s convenient to flip them on together for triage/experiments (e.g., a top-level
--enable_gemm_fusionsor an env var that overridesenabled:false).Example override pattern (if your loader supports it):
# Pseudocode: apply late overrides when env/CLI flag is set -if overrides.enable_gemm_fusions: - for k in ["fuse_gemms", "fuse_fp4_gemms", "fuse_fp8_gemms"]: - cfg.transforms[k].enabled = True
97-97: Double-check interactions with earlier MOE quantization passes.
quantize_fp8_moe/quantize_nvfp4_moeearlier in the pipeline may rely on or benefit fromfuse_moe. Disablingfuse_moecan be perf-only, but verify it doesn’t affect graph correctness or expected kernels.Please run the MOE model e2e pipeline (small config) and confirm graph equivalence and no perf regressions beyond the known OOM avoidance goal.
103-104: Make backend selection a first-class knob now (defaulting to flashinfer).Since
fuse_rmsnormis kept enabled, add abackendoverride in the same config section (and plumb it through) so CI/perf triage can switch between backends without editing YAML.Proposed minimal addition (if supported by the loader):
fuse_rmsnorm: # TODO (lucaslie): add backend selection as part of configurable inference optimizers # check if we can fuse rmsnorm stage: post_load_fusion - backend: flashinfer + backend: flashinfer # default + # override_backend_env: TLLM_FUSE_RMSNORM_BACKEND # optional: env override
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
tensorrt_llm/_torch/auto_deploy/config/default.yaml(1 hunks)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Pre-commit Check
🔇 Additional comments (1)
tensorrt_llm/_torch/auto_deploy/config/default.yaml (1)
80-88: ```shell
#!/bin/bash
set -euo pipefail
echo "PWD: $(pwd)"
echo "Top-level files:"
ls -la || trueecho; echo "1) Find auto_deploy paths and default.yaml files"
rg -n --hidden -S 'auto_deploy' || true
rg -n --hidden -S 'default.yaml' || trueecho; echo "2) Show tensorrt_llm/_torch/auto_deploy/config/default.yaml if present"
FILE="tensorrt_llm/_torch/auto_deploy/config/default.yaml"
if [ -f "$FILE" ]; then
echo "---- $FILE ----"
sed -n '1,240p' "$FILE" || true
else
echo "$FILE not found"
fiecho; echo "3) Search for uses of the 'enabled' flag under tensorrt_llm"
rg -n -C3 --hidden -S '\bget(["'"'"']enabled["'"'"'])|\benabled\b|config.get(["'"'"']enabled["'"'"'])' tensorrt_llm || trueecho; echo "4) Search for transform iteration / scheduling patterns"
rg -n -C3 --hidden -S '\btransforms\b|\bpasses\b|for\s+\w+\s+in\s+.*transforms' tensorrt_llm || trueecho; echo "5) Search for specific fuse transform names"
rg -n -C3 --hidden -S 'fuse_gemms|fuse_fp4_gemms|fuse_fp8_gemms|post_load_fusion' tensorrt_llm || trueecho; echo "6) Search for YAML/config loading code that may map YAML to runtime objects"
rg -n -C3 --hidden -S 'yaml.safe_load|yaml.load|OmegaConf|from_yaml|load_yaml|omegaconf' tensorrt_llm || trueecho; echo "7) As a fallback, search for 'enabled' across the whole repo (may be noisy)"
rg -n --hidden -S '\benabled\b' || trueecho; echo "Done."
</blockquote></details> </blockquote></details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
lucaslie
force-pushed
the
ll/neat_transform_disabling
branch
from
39144d9 to
0901d83
Compare
|
/bot run |
|
PR_Github #18648 [ run ] triggered by Bot |
|
PR_Github #18644 [ run ] completed with state |
|
PR_Github #18648 [ run ] completed with state |
|
/bot run |
|
PR_Github #18662 [ run ] triggered by Bot |
|
PR_Github #18662 [ run ] completed with state |
|
/bot run |
|
PR_Github #18665 [ run ] triggered by Bot |
|
PR_Github #18665 [ run ] completed with state |
|
/bot run |
Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com>
lucaslie
force-pushed
the
ll/neat_transform_disabling
branch
from
0901d83 to
9874ecd
Compare
|
/bot run |
|
PR_Github #18669 [ run ] triggered by Bot |
|
PR_Github #18670 [ run ] triggered by Bot |
|
PR_Github #18669 [ run ] completed with state |
|
PR_Github #18670 [ run ] completed with state |
|
/bot run |
|
/bot run |
|
PR_Github #18789 [ run ] triggered by Bot |
|
PR_Github #18789 [ run ] completed with state |
|
/bot skip --comment "Unrelated failures, minor changes" |
|
PR_Github #18808 [ skip ] triggered by Bot |
|
PR_Github #18808 [ skip ] completed with state |
…VIDIA#7736) Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com>
…VIDIA#7736) Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com>
7 participants
Summary by CodeRabbit
Chores
Documentation
Description
By not commenting out a transform in
default.yamland instead use the keywordenabled: falsethe pipeline can more easily be reconfigured via an extra yaml or the appropriate CLI flag. For example, to enablefuse_moefor a single experiment you can now just run: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)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
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
/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...Provide a user friendly way for developers to interact with a Jenkins server.
Run
/bot [-h|--help]to print this help message.See details below for each supported subcommand.
run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--disable-fail-fast(OPTIONAL) : Disable fail fast on build/tests/infra failures.--skip-test(OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.--stage-list "A10-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.--gpu-type "A30, H100_PCIe"(OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.--only-multi-gpu-test(OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.--disable-multi-gpu-test(OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.--add-multi-gpu-test(OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.--post-merge(OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip testing for latest commit on pull request.
--comment "Reason for skipping build/test"is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.reuse-pipeline
reuse-pipelineReuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.