Draft: [AD] Enabled test_trtllm_bench_backend_comparison on L0 of B200 and A30 #7002
Conversation
Signed-off-by: Eran Geva <[email protected]>
Signed-off-by: Eran Geva <[email protected]>
Signed-off-by: Eran Geva <[email protected]>
📝 WalkthroughWalkthroughExpands PyTorch singlegpu test coverage by removing -k exclusions in two integration test lists. Updates singlegpu TRT-LLM bench test to adjust memory metric calculations, rename a variable, align expected memory ranges, and relax backend tolerance default from 0.2 to 0.3. Changes
Sequence Diagram(s)sequenceDiagram
participant Test as test_trtllm_bench_backend_comparison
participant Bench as TRT-LLM Bench Runner
participant Logs as Runtime Logs
participant Parser as Metrics Parser
participant Validator as Memory/Perf Validator
Test->>Bench: Launch with config (layers, batch, tolerances)
Bench-->>Logs: Emit current_cache_size, free_mem_pre/post, new_cache_size
Test->>Parser: Parse logs
Parser-->>Validator: Metrics (pre/post free mem, cache sizes)
Validator-->>Test: Check ranges, memory_reduction>0, cache size match (±1%), perf tolerances
Test-->>Test: Assert pass/fail
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 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. ✨ 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
|
MrGeva
changed the title
|
/bot run |
|
PR_Github #15622 [ run ] triggered by Bot |
![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 (6)
tests/integration/test_lists/test-db/l0_a30.yml (1)
21-21: Narrow L0 test scope to avoid running all 111 singlegpu testsDropping the existing
-kfilter causes L0 on A30 to execute all 111 tests undersinglegpu, significantly increasing runtime. If the goal is only to runtest_trtllm_bench_backend_comparison, please update the YAML entry accordingly:Affected file:
- tests/integration/test_lists/test-db/l0_a30.yml (line 21)
Suggested change:
- - unittest/_torch/auto_deploy/unit/singlegpu + - unittest/_torch/auto_deploy/unit/singlegpu -k "test_trtllm_bench_backend_comparison"tests/integration/test_lists/test-db/l0_b200.yml (1)
73-73: Filter to the specific L0 bench comparison test
Running the baresinglegpufolder kicks off all 112 tests undertests/unittest/_torch/auto_deploy/unit/singlegpu. To limit L0 to only the backend comparison test, update line 73 oftests/integration/test_lists/test-db/l0_b200.yml:- - unittest/_torch/auto_deploy/unit/singlegpu + - unittest/_torch/auto_deploy/unit/singlegpu -k "test_trtllm_bench_backend_comparison"Verification results:
- Total singlegpu tests: 112
test_trtllm_bench_backend_comparisonexists attests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py:604tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py (4)
1-6: Add NVIDIA copyright headerPer repo guidelines, prepend the NVIDIA copyright header (current year) to all Python files.
+# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + import json import re import subprocess import tempfile from pathlib import Path
282-282: Hard-coded extra_consumption_mb increased to 2700; consider parameterizing by GPU or deriving from runtimeThe unexplained 2.7GB “extra consumption” is carried as a constant. This will likely differ across A30 vs. B200 and over time. Either:
- make it a per-GPU map (A30/B100/B200), or
- infer it from a small warmup measurement, then set the expected ranges accordingly.
If you prefer a minimal step, add a TODO with the GPU-specific follow-up and wire a per-GPU override via torch.cuda.get_device_name().
- extra_consumption_mb = 2700 + # TODO(egeva): consider varying by GPU (A30/B200) or deriving from a warmup probe. + extra_consumption_mb = 2700
585-596: Minor: avoid shadowing “llm_root” import with fixture nameThe parameter llm_root shadows the earlier import from utils.cpp_paths (hence the noqa: F811). If feasible, drop the import (pytest fixture doesn’t need it) or rename the param to llm_root_path to avoid shadowing. Not blocking.
-def test_trtllm_bench(llm_root): # noqa: F811 +def test_trtllm_bench(llm_root): # noqa: F811 # Consider renaming param to llm_root_path
403-414: Remove unusedrequire_metricsparameter & unreachable warn‐only branchThe
validate_and_extract_kv_cache_metricshelper (defined at tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py:381) declares arequire_metricsflag (default True), but all call sites (e.g. line 513) pass or rely on the default ofTrue. The “warn‐only” path (require_metrics=False) is therefore never hit.• Tests never invoke
require_metrics=False, making the else‐branch unreachable.
• To clean up, either:
– Remove therequire_metricsparameter and its conditional, always failing when metrics are missing; or
– If warn‐only behavior is genuinely needed, update callers to passFalseand have the function return gracefully instead of asserting.
📜 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 (3)
tests/integration/test_lists/test-db/l0_a30.yml(1 hunks)tests/integration/test_lists/test-db/l0_b200.yml(1 hunks)tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py(5 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/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.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/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py
🧠 Learnings (2)
📓 Common learnings
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.
📚 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/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py
🔇 Additional comments (4)
tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py (4)
293-294: Post-forward free memory range aligned to pre-range; rely entirely on the >0 reduction checkSetting expected_free_mem_post_range equal to pre-range is okay given the explicit memory_reduction > 0 assertion added below. Just be aware that this relaxes the previous lower-bound tightening and may let borderline cases pass. If you want a mild guardrail, consider shrinking post upper bound by a small delta (e.g., 64–128MB) to catch no-op resizes.
Do you want me to propose a small delta-based post-range to catch no-op cases while keeping the check stable on A30/B200?
347-354: Good: explicit memory reduction assertion clarifies the intentChecking that free_mem_pre > free_mem_post removes ambiguity and complements the broader range checks.
447-458: Backend tolerance relaxed to 0.3: OK for L0 stability, but keep an eye on regressionsIncreasing backend_relative_tolerance from 0.2 to 0.3 is reasonable to reduce flakes between autodeploy and PyTorch on mixed hardware, especially at L0. If this becomes too permissive, consider lowering back after a few green runs.
Would you like me to open a follow-up to collect a week of tolerance telemetry and tighten it if stable?
605-624: Docstring extension is clear and helpfulGood documentation of the metrics parsed and the validation steps, including the rationale behind extra_consumption_mb. This will aid future investigation.
|
PR_Github #15622 [ run ] completed with state |
2 participants
depends (and based) on this PR: #6997
Summary by CodeRabbit
Tests
Documentation
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.
Details
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.