[None][test] add deepseek r1/v3 model with chunked prefill cases #7124
Conversation
📝 WalkthroughWalkthroughAdds pattern-based model config entries to enable attention data-parallel and chunked prefill for Deepseek variants, removes a duplicate pattern, and appends chunked-prefill performance tests to cluster and full QA test lists. No public function signatures were changed. Changes
Sequence Diagram(s)sequenceDiagram
participant Runner as Perf Runner
participant Config as get_model_yaml_config
participant Patterns as pattern_config
Runner->>Config: get_model_yaml_config(model_label)
Config->>Patterns: iterate pattern_config entries
alt model_label contains "deepseek_r1"
Note right of Config #DDEBF7: set enable_attention_dp = True
end
alt model_label matches chunked prefill group
Note right of Config #F6F8E9: set enable_attention_dp = True<br/>set enable_chunked_prefill = True
end
Config-->>Runner: return base_config with flags applied
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
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
tests/integration/defs/perf/pytorch_model_config.py (1)
32-34: Python 3.8 compatibility: avoid PEP 585 built-in generics (list[str])Repo targets Python 3.8+, but list[str] requires Python 3.9+. Use typing.List/Optional for compatibility.
@@ -def get_model_yaml_config(model_label: str, - lora_dirs: list[str] = None) -> dict: +from typing import List, Optional + +def get_model_yaml_config(model_label: str, + lora_dirs: Optional[List[str]] = None) -> dict:
🧹 Nitpick comments (1)
tests/integration/test_lists/qa/llm_perf_cluster.yml (1)
20-22: Optional: Align 1-GPU chunked prefill with existing timeoutsI checked and confirmed that this heavy 1-GPU chunked prefill test currently has no timeout, while similar multi-GPU cases do use
TIMEOUT(120). Adding the same timeout here can help avoid flakiness on slower nodes without impacting fast runners.• Location:
- tests/integration/test_lists/qa/llm_perf_cluster.yml line 21
• Proposed change:
- - perf/test_perf.py::test_perf[deepseek_v3_lite_nvfp4-bench-pytorch-float4-maxbs:512-maxnt:2048-kv_frac:0.85-input_output_len:5000,500-reqs:200] + - perf/test_perf.py::test_perf[deepseek_v3_lite_nvfp4-bench-pytorch-float4-maxbs:512-maxnt:2048-kv_frac:0.85-input_output_len:5000,500-reqs:200] TIMEOUT(120)This mirrors the multi-GPU patterns and should be safe to apply.
📜 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/defs/perf/pytorch_model_config.py(2 hunks)tests/integration/test_lists/qa/llm_perf_cluster.yml(3 hunks)tests/integration/test_lists/qa/llm_perf_full.yml(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/integration/defs/perf/pytorch_model_config.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/integration/defs/perf/pytorch_model_config.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 (3)
tests/integration/defs/perf/pytorch_model_config.py (1)
56-62: LGTM: default enable_attention_dp for all DeepSeek R1 labelsThe broad substring pattern 'deepseek_r1' is appropriate and ensures attention DP is enabled across R1 variants before more specific rules apply.
tests/integration/test_lists/qa/llm_perf_cluster.yml (2)
86-88: LGTM: chunked prefill nvfp4 patterns for 4-GPU block align with config and include TIMEOUTNames include kv_frac:0.85 and should match the chunked prefill pattern rules after the config change. Timeout is consistent with similar heavy cases.
146-148: LGTM: chunked prefill nvfp4 patterns for 8-GPU block align with config and include TIMEOUTThese should correctly trigger enable_chunked_prefill and are scoped with ep/tp/gpus=8. Timeout looks appropriate.
ruodil
changed the title
ruodil
force-pushed
the
user/ruodil/new_feature
branch
from
142aa93 to
8d20e84
Compare
Signed-off-by: ruodil <200874449+ruodil@users.noreply.github.com>
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
tests/integration/test_lists/qa/llm_perf_cluster.yml (1)
20-21: Add missing 1-GPU 256-maxbs chunked prefill case for deepseek_v3_lite_nvfp4We currently have only the 512-maxbs chunked-prefill entry in tests/integration/test_lists/qa/llm_perf_cluster.yml (around lines 20–21) but no matching 256-maxbs case. Chunked prefill is confirmed enabled for deepseek_v3_lite_nvfp4 in tests/integration/defs/perf/pytorch_model_config.py, so this test will be exercised as intended.
• tests/integration/test_lists/qa/llm_perf_cluster.yml (lines 20–21): insert the 256-maxbs entry directly below the existing 512-maxbs case
• tests/integration/defs/perf/pytorch_model_config.py: no changes needed—enable_chunked_prefill: Trueis already set for deepseek_v3_lite_nvfp4Apply this diff:
# for chunked prefill cases - perf/test_perf.py::test_perf[deepseek_v3_lite_nvfp4-bench-pytorch-float4-maxbs:512-maxnt:2048-kv_frac:0.85-input_output_len:5000,500-reqs:200] + - perf/test_perf.py::test_perf[deepseek_v3_lite_nvfp4-bench-pytorch-float4-maxbs:256-maxnt:1024-kv_frac:0.85-input_output_len:2000,2000-reqs:200]
📜 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/defs/perf/pytorch_model_config.py(2 hunks)tests/integration/test_lists/qa/llm_perf_cluster.yml(3 hunks)tests/integration/test_lists/qa/llm_perf_full.yml(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- tests/integration/defs/perf/pytorch_model_config.py
- tests/integration/test_lists/qa/llm_perf_full.yml
⏰ 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/integration/test_lists/qa/llm_perf_cluster.yml (1)
86-88: Confirmed: 4-GPU chunked-prefill tests wire upmaxntandenable_chunked_prefillcorrectlyI verified that tests/integration/defs/perf/pytorch_model_config.py includes
'enable_chunked_prefill': True,and that the performance harness propagates
max_num_tokens(parsed from themaxnt:label) into the LLM invocation. In the runtime (tensorrt_llm/llmapi/llm.py), whenenable_chunked_prefill=Truethe early‐rejection checks are bypassed andenable_chunked_contextis set, triggering chunked prefill rather than an error. No test-side changes are needed.
|
@jmydurant please kindly review as well |
Signed-off-by: ruodil <200874449+ruodil@users.noreply.github.com>
|
/bot reuse-pipeline |
|
/bot skip --comment "skip test as just adding cases" |
|
PR_Github #18917 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #18917 [ reuse-pipeline ] completed with state |
|
PR_Github #18919 [ skip ] triggered by Bot |
|
PR_Github #18919 [ skip ] completed with state |
|
/bot skip --comment "skip test as just adding cases" |
|
PR_Github #19248 [ skip ] triggered by Bot |
|
PR_Github #19248 [ skip ] completed with state |
…DIA#7124) Signed-off-by: ruodil <200874449+ruodil@users.noreply.github.com>
…DIA#7124) Signed-off-by: ruodil <200874449+ruodil@users.noreply.github.com>
…DIA#7124) Signed-off-by: ruodil <200874449+ruodil@users.noreply.github.com>
4 participants
Summary by CodeRabbit
New Features
Tests
Refactor
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.