[None][chore] Cherry-pick from (#7598) Make low_precision_combine as a llm arg #7898
Conversation
zongfeijing
requested review from
Naveassaf,
Superjomn,
kaiyux,
laikhtewari and
yuxianq
|
/bot run |
|
PR_Github #19516 [ run ] triggered by Bot |
📝 WalkthroughWalkthroughAdds a new boolean flag use_low_precision_moe_combine, exposed via CLI, threaded through MoeConfig → PyTorchConfig → model engine/load, and consumed in fused MoE to replace the previous environment-variable gate. Defaults to False and remains NVFP4-gated. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User
participant CLI as CLI Parser (quickstart_advanced.py)
participant LLMArgs as MoeConfig (llmapi/llm_args.py)
participant PTConf as PyTorchConfig
participant Engine as PyTorchModelEngine
participant Loader as _load_model(...)
participant MoE as Fused MoE (fused_moe_wide_ep.py)
User->>CLI: --use_low_precision_moe_combine
CLI->>LLMArgs: MoeConfig(use_low_precision_moe_combine)
LLMArgs->>PTConf: get_pytorch_backend_config(..., use_low_precision_moe_combine)
PTConf->>Engine: construct with backend config
Engine->>Loader: _load_model(..., use_low_precision_moe_combine=...)
note over Loader,MoE: Model initialization
Loader->>MoE: Build/Configure modules with model_config
alt enable_alltoall && NVFP4 supported
MoE->>MoE: If model_config.use_low_precision_moe_combine then use low-precision combine
else
MoE->>MoE: Use standard precision combine
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 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
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (5)
tensorrt_llm/_torch/model_config.py (1)
1-1: Add NVIDIA Apache-2.0 header (2025).Required by repo guidelines for all .py files.
+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License.tensorrt_llm/_torch/pyexecutor/config.py (1)
1-1: Add NVIDIA Apache-2.0 header (2025).+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License.tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (1)
1-1: Add NVIDIA Apache-2.0 header (2025).+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License.examples/llm-api/quickstart_advanced.py (1)
1-1: Add NVIDIA Apache-2.0 header (2025).+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License.tensorrt_llm/llmapi/llm_args.py (1)
1-1: Add NVIDIA Apache-2.0 header (2025).+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# http://www.apache.org/licenses/LICENSE-2.0 +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License.
🧹 Nitpick comments (6)
tensorrt_llm/_torch/model_config.py (1)
14-20: Duplicate logger import; keep only one.Two different imports alias the same symbol; drop the redundant one.
-from tensorrt_llm import logger from tensorrt_llm._torch.pyexecutor.config_utils import is_nemotron_hybrid @@ -from tensorrt_llm.logger import logger +from tensorrt_llm.logger import loggertensorrt_llm/_torch/pyexecutor/config.py (1)
58-58: Field addition: LGTM; consider documenting constraints.Add a brief comment noting “only effective with NVFP4 and alltoall-enabled paths,” mirroring the CLI/help text.
- use_low_precision_moe_combine: bool = False + # Use low-precision combine in MoE (effective only with NVFP4 + alltoall) + use_low_precision_moe_combine: bool = Falsetensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (2)
192-193: Gate is correct; add user-facing warning + info_once when ineffective/effective.Warn if the flag is set but NVFP4 isn’t active, and log when it’s enabled.
- self.use_low_precision_combine = model_config.use_low_precision_moe_combine and qm.has_nvfp4( - ) + requested_lp = model_config.use_low_precision_moe_combine + has_nvfp4 = qm.has_nvfp4() + if requested_lp and not has_nvfp4: + logger.warning( + "use_low_precision_moe_combine requested but NVFP4 quantization is not enabled; ignoring." + ) + self.use_low_precision_combine = requested_lp and has_nvfp4 + if self.use_low_precision_combine: + logger.info_once( + "Enabling low-precision MoE combine (NVFP4).", + key="use_low_precision_moe_combine_enabled")
192-193: Optional: keep env-var fallback during transition.If backward compatibility is desired, OR the flag with the legacy env var (then remove in a future major).
- requested_lp = model_config.use_low_precision_moe_combine + requested_lp = model_config.use_low_precision_moe_combine or ( + os.environ.get("TRTLLM_MOE_USE_LOW_PRECISION_COMBINE", "0") == "1" + )examples/llm-api/quickstart_advanced.py (1)
76-81: CLI flag: LGTM; clarify help text.Make constraints explicit to reduce confusion.
- help='Use low precision combine in MoE (only for NVFP4 quantization)') + help='Use low-precision combine in MoE (requires NVFP4 quantization and MoE alltoall path).')tensorrt_llm/llmapi/llm_args.py (1)
194-199: MoeConfig field: LGTM; consider marking as beta for telemetry warnings.If still evolving, set status="beta" so warn_on_unstable_feature_usage surfaces a warning.
- use_low_precision_moe_combine: bool = Field( + use_low_precision_moe_combine: bool = Field( default=False, - description= + description= "Use low precision combine in MoE operations (only for NVFP4 quantization). When enabled, uses lower precision for combining expert outputs to improve performance." - ) + , status="beta")
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
examples/llm-api/quickstart_advanced.py(2 hunks)tensorrt_llm/_torch/model_config.py(1 hunks)tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py(1 hunks)tensorrt_llm/_torch/pyexecutor/config.py(1 hunks)tensorrt_llm/_torch/pyexecutor/model_engine.py(1 hunks)tensorrt_llm/llmapi/llm_args.py(2 hunks)
🧰 Additional context used
📓 Path-based instructions (3)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Use only spaces, no tabs; indent with 4 spaces.
Files:
tensorrt_llm/_torch/model_config.pytensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.pytensorrt_llm/_torch/pyexecutor/config.pytensorrt_llm/llmapi/llm_args.pyexamples/llm-api/quickstart_advanced.pytensorrt_llm/_torch/pyexecutor/model_engine.py
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+.
Indent Python code with 4 spaces; do not use tabs.
Maintain module namespace when importing; prefer 'from package.subpackage import foo' then 'foo.SomeClass()' instead of importing the class directly.
Python filenames should be snake_case (e.g., some_file.py).
Python classes use PascalCase names.
Functions and methods use snake_case names.
Local variables use snake_case; prefix 'k' for variables that start with a number (e.g., k_99th_percentile).
Global variables use upper SNAKE_CASE prefixed with 'G' (e.g., G_MY_GLOBAL).
Constants use upper SNAKE_CASE (e.g., MY_CONSTANT).
Avoid shadowing variables from an outer scope.
Initialize all externally visible members of a class in the constructor.
Prefer docstrings for interfaces that may be used outside a file; comments for in-function or file-local interfaces.
Use Google-style docstrings for classes and functions (Sphinx-parsable).
Document attributes and variables inline so they render under the class/function docstring.
Avoid reflection when a simpler, explicit approach suffices (e.g., avoid dict(**locals()) patterns).
In try/except, catch the most specific exceptions possible.
For duck-typing try/except, keep the try body minimal and use else for the main logic.
Files:
tensorrt_llm/_torch/model_config.pytensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.pytensorrt_llm/_torch/pyexecutor/config.pytensorrt_llm/llmapi/llm_args.pyexamples/llm-api/quickstart_advanced.pytensorrt_llm/_torch/pyexecutor/model_engine.py
**/*.{cpp,cxx,cc,h,hpp,hh,hxx,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Prepend the NVIDIA Apache-2.0 copyright header with current year to the top of all source files (e.g., .cpp, .h, .cu, .py).
Files:
tensorrt_llm/_torch/model_config.pytensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.pytensorrt_llm/_torch/pyexecutor/config.pytensorrt_llm/llmapi/llm_args.pyexamples/llm-api/quickstart_advanced.pytensorrt_llm/_torch/pyexecutor/model_engine.py
🧠 Learnings (1)
📚 Learning: 2025-08-26T06:07:02.166Z
Learnt from: shaharmor98
PR: NVIDIA/TensorRT-LLM#7231
File: tensorrt_llm/_torch/pyexecutor/_util.py:504-509
Timestamp: 2025-08-26T06:07:02.166Z
Learning: In tensorrt_llm/_torch/pyexecutor/_util.py, when calling model_engine.set_lora_model_config(), pass model_binding_config.mlp_hidden_size directly without multiplying by mapping.tp_size, as the mlp_hidden_size from get_bindings_model_config() is already the per-TP rank value needed for LoRA weight packaging.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
🧬 Code graph analysis (2)
tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (2)
tensorrt_llm/_torch/modules/linear.py (1)
has_nvfp4(1903-1906)tensorrt_llm/_torch/modules/fused_moe/interface.py (1)
has_nvfp4(296-299)
examples/llm-api/quickstart_advanced.py (1)
tensorrt_llm/llmapi/llm_args.py (1)
MoeConfig(168-202)
⏰ 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 (4)
examples/llm-api/quickstart_advanced.py (1)
236-236: Propagation into MoeConfig: LGTM.tensorrt_llm/llmapi/llm_args.py (1)
2511-2513: Pass-through to PyTorchConfig: LGTM.tensorrt_llm/_torch/model_config.py (1)
136-137: New config knob: LGTM; make sure it actually plumbs through to fused MoE.Default False, NVFP4-only is fine. Please confirm the flag reaches ModelConfig at load time and is consumed by fused_moe_wide_ep.
tensorrt_llm/_torch/pyexecutor/model_engine.py (1)
314-315: LGTM: flag correctly threaded; verify downstream consumption to avoid TypeError.Passing use_low_precision_moe_combine via _load_model’s **kwargs is correct. Please confirm:
- PyTorchConfig defines use_low_precision_moe_combine.
- checkpoint_loader.load_config accepts this kwarg (either explicitly or via **kwargs).
- The flag is actually consumed in the MoE path.
Run:
zongfeijing
changed the title
|
PR_Github #19516 [ run ] completed with state |
Signed-off-by: Zongfei Jing <20381269+zongfeijing@users.noreply.github.com>
zongfeijing
force-pushed
the
user/zongfeij/cherry-pick-7598
branch
from
514f60a to
96f3914
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #19629 [ run ] triggered by Bot |
|
PR_Github #19629 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #19662 [ run ] triggered by Bot |
|
PR_Github #19662 [ run ] completed with state |
Signed-off-by: Zongfei Jing <20381269+zongfeijing@users.noreply.github.com>
|
/bot run --disable-fail-fast |
|
PR_Github #19734 [ run ] triggered by Bot |
|
/bot run --disable-fail-fast |
|
PR_Github #19734 [ run ] completed with state |
|
PR_Github #19815 [ run ] triggered by Bot |
|
PR_Github #19815 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #19858 [ run ] triggered by Bot |
|
PR_Github #19858 [ run ] completed with state |
Signed-off-by: Zongfei Jing <20381269+zongfeijing@users.noreply.github.com>
|
/bot run --disable-fail-fast |
|
PR_Github #20161 [ run ] triggered by Bot |
|
PR_Github #20161 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #20170 [ run ] triggered by Bot |
|
PR_Github #20170 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #20179 [ run ] triggered by Bot |
|
PR_Github #20179 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #20190 [ run ] triggered by Bot |
|
PR_Github #20190 [ run ] completed with state |
6 participants
Summary by CodeRabbit
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)
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.