[None][test] Add MNNVL to unittest and speed up weight creation #7071
Conversation
Signed-off-by: Tailing Yuan <[email protected]>
📝 WalkthroughWalkthroughModify integration test matrices to add a fused-MoE FP4 all-to-all test variant and update a unit test to allocate three fused-MoE weight tensors on CUDA; no production code or public APIs changed. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Suggested reviewers
Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. 📜 Recent review detailsConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro 💡 Knowledge Base configuration:
You can enable these sources in your CodeRabbit configuration. 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
⏰ 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)
✨ Finishing Touches🧪 Generate unit tests
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. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
yuantailing
changed the title
|
/bot run |
![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 (1)
tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (1)
1-1: Missing NVIDIA copyright headerThis source file lacks the required NVIDIA copyright header per repo guidelines. Please prepend the current-year header.
Apply at file top:
+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved.
🧹 Nitpick comments (1)
tests/unittest/_torch/modules/test_fused_moe.py (1)
214-222: Optional: be explicit about device index to avoid surprisesYou set the current device via
torch.cuda.set_device(mapping.rank), sodevice="cuda"works. For clarity and future-proofing, consider targeting the device explicitly.Apply:
- w1_weight = torch.empty((INTERMEDIATE_SIZE, HIDDEN_SIZE), - dtype=dtype, - device="cuda") + w1_weight = torch.empty((INTERMEDIATE_SIZE, HIDDEN_SIZE), + dtype=dtype, + device=f"cuda:{mapping.rank}") - w2_weight = torch.empty((HIDDEN_SIZE, INTERMEDIATE_SIZE), - dtype=dtype, - device="cuda") + w2_weight = torch.empty((HIDDEN_SIZE, INTERMEDIATE_SIZE), + dtype=dtype, + device=f"cuda:{mapping.rank}") - w3_weight = torch.empty((INTERMEDIATE_SIZE, HIDDEN_SIZE), - dtype=dtype, - device="cuda") + w3_weight = torch.empty((INTERMEDIATE_SIZE, HIDDEN_SIZE), + dtype=dtype, + device=f"cuda:{mapping.rank}")
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (4)
tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py(1 hunks)tests/integration/test_lists/test-db/l0_dgx_b200.yml(1 hunks)tests/integration/test_lists/test-db/l0_dgx_h100.yml(1 hunks)tests/unittest/_torch/modules/test_fused_moe.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tests/unittest/_torch/modules/test_fused_moe.pytensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tests/unittest/_torch/modules/test_fused_moe.pytensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py
🧠 Learnings (1)
📚 Learning: 2025-07-28T17:06:08.621Z
Learnt from: moraxu
PR: NVIDIA/TensorRT-LLM#6303
File: tests/integration/test_lists/qa/examples_test_list.txt:494-494
Timestamp: 2025-07-28T17:06:08.621Z
Learning: In TensorRT-LLM testing, it's common to have both CLI flow tests (test_cli_flow.py) and PyTorch API tests (test_llm_api_pytorch.py) for the same model. These serve different purposes: CLI flow tests validate the traditional command-line workflow, while PyTorch API tests validate the newer LLM API backend. Both are legitimate and should coexist.
Applied to files:
tests/integration/test_lists/test-db/l0_dgx_b200.yml
🔇 Additional comments (4)
tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (1)
467-472: Good guard: avoid AttributeError when no load balancer is presentAdding
self.layer_load_balancerto theis_last_callguard prevents a call toget_local_statistic_tensor()when no load balancer exists. This fixes a crash in MNNVL path when EPLB is disabled.tests/integration/test_lists/test-db/l0_dgx_b200.yml (1)
18-18: Add MNNVL FP4 MoE all-to-all test on B200: looks goodThe new test entry matches the existing parametrized test id
MNNVLand is correctly scoped to B200 hardware.tests/unittest/_torch/modules/test_fused_moe.py (1)
214-222: Allocate weights directly on GPU to speed test setup: nice improvementConstructing w1/w2/w3 on CUDA avoids extra H2D copies and speeds up tests.
tests/integration/test_lists/test-db/l0_dgx_h100.yml (1)
76-76: Verify MNNVL availability on H100 CI runnersThe new integration test on H100 forces use of
AlltoallMethodType.MNNVL, butWideEPMoE.__init__will unconditionally callMnnvlMemory.initialize()when
alltoall_method_type == MNNVL. If the H100 CI image doesn’t have the MNNVL SDK built and its shared library onLD_LIBRARY_PATH, this will fail at module init.Please ensure one of the following before landing:
- The H100 CI environment includes a compiled MNNVL library and correct runtime paths.
- Guard or skip the H100 test entry when
not MnnvlMemory.supports_mnnvl().Locations to review:
- tests/integration/test_lists/test-db/l0_dgx_h100.yml: line 76
- tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py: lines 200–207 (
MnnvlMemory.initialize(), workspace setup)- tests/unittest/_torch/modules/test_fused_moe.py: parameterize blocks at lines 187–189 and 297–299
|
PR_Github #15862 [ run ] triggered by Bot |
|
PR_Github #15862 [ run ] completed with state |
|
/bot run --stage-list "DGX_B200-4_GPUs-PyTorch-Post-Merge-1" |
|
PR_Github #16114 [ run ] triggered by Bot |
|
PR_Github #16114 [ run ] completed with state |
Signed-off-by: Tailing Yuan <[email protected]>
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (3)
tests/unittest/_torch/modules/test_fused_moe.py (3)
242-256: Free temporary GPU weights after loading to both models to reduce peak VRAMNow that weights live on GPU, keeping the
weightsdict until function exit adds an extra ≈1.7 GB per rank (72 experts × 3 matrices × 1536×2560 bf16) on top of two model parameter copies. Delete the dict after bothload_weightscalls to lower peak usage and reduce OOM risk in multi-rank runs.Apply this minimal change:
alltoall_model.to("cuda") alltoall_model.load_weights([weights]) ref_model = CutlassFusedMoE( num_experts=NUM_EXPERTS, routing_method=routing_method, hidden_size=HIDDEN_SIZE, intermediate_size=INTERMEDIATE_SIZE, dtype=dtype, reduce_results=True, model_config=ModelConfig(mapping=mapping, max_num_tokens=MAX_NUM_TOKENS), ) ref_model.to("cuda") ref_model.load_weights([weights]) + # Free temporary GPU weight tensors to lower peak memory. + del weights
209-211: Optional: seed CUDA RNG explicitly for clarity/determinism
torch.manual_seed(mapping.rank)typically seeds CUDA as well, but making the CUDA seed explicit keeps this test consistent with others in the file that call both and can help avoid flakiness after switching to on-GPU initialization.torch.cuda.set_device(mapping.rank) - torch.manual_seed(mapping.rank) + torch.manual_seed(mapping.rank) + # Optional: seed CUDA RNG explicitly for clarity. + torch.cuda.manual_seed(mapping.rank)
1-1: Add NVIDIA copyright header (2025) per repo guidelinesTests are also source files; please prepend the standard NVIDIA header.
+# Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. 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.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (3)
tests/integration/test_lists/test-db/l0_dgx_b200.yml(1 hunks)tests/integration/test_lists/test-db/l0_dgx_h100.yml(1 hunks)tests/unittest/_torch/modules/test_fused_moe.py(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- tests/integration/test_lists/test-db/l0_dgx_b200.yml
- tests/integration/test_lists/test-db/l0_dgx_h100.yml
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tests/unittest/_torch/modules/test_fused_moe.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tests/unittest/_torch/modules/test_fused_moe.py
⏰ 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)
tests/unittest/_torch/modules/test_fused_moe.py (1)
215-222: Allocate expert weights directly on CUDA — LGTMCreating w1/w2/w3 on device="cuda" avoids CPU→GPU staging copies and aligns with the PR’s goal to speed up weight creation. The initializer works on-GPU and respects the current device set earlier.
|
/bot run --extra-stage "DGX_B200-4_GPUs-PyTorch-Post-Merge-1" |
|
PR_Github #16325 [ run ] triggered by Bot |
|
PR_Github #16325 [ run ] completed with state |
Signed-off-by: Tailing Yuan <[email protected]>
|
/bot run |
|
PR_Github #16366 [ run ] triggered by Bot |
|
PR_Github #16366 [ run ] completed with state |
|
/bot run |
|
PR_Github #16396 [ run ] triggered by Bot |
|
PR_Github #16396 [ run ] completed with state |
Summary by CodeRabbit
Description
Test Coverage
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.