[https://nvbugs/6287561][fix] Add get_sm_version() < 90 check at the top of run_MTP() in… - #15343
[https://nvbugs/6287561][fix] Add get_sm_version() < 90 check at the top of run_MTP() in…#15343tensorrt-cicd wants to merge 1 commit into
get_sm_version() < 90 check at the top of run_MTP() in…#15343Conversation
📝 WalkthroughWalkthroughThis change adds an SM-version check to the MTP speculative decoding example so it skips execution on GPUs below SM 90 with a warning, and it removes the integration-test waiver for that MTP example. ChangesMTP speculative decoding gating
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
examples/llm-api/llm_speculative_decoding.py (1)
1-1:⚠️ Potential issue | 🟠 Major | ⚡ Quick winAdd the required NVIDIA copyright header to this modified Python file.
Line 1 starts directly with content, but modified TensorRT-LLM source files are required to include the NVIDIA copyright/license header block.
As per coding guidelines, “All TensorRT-LLM Open Source Software code should contain an NVIDIA copyright header... update year on modified files.”🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@examples/llm-api/llm_speculative_decoding.py` at line 1, The file llm_speculative_decoding.py is missing the required NVIDIA copyright/license header block that must appear at the beginning of all modified TensorRT-LLM source files. Add the NVIDIA copyright and license header as a comment block at the very start of the file, before the existing content (### :title Speculative Decoding), ensuring it includes the copyright notice, appropriate year, and license information formatted according to TensorRT-LLM coding guidelines for Python files.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@examples/llm-api/llm_speculative_decoding.py`:
- Line 1: The file llm_speculative_decoding.py is missing the required NVIDIA
copyright/license header block that must appear at the beginning of all modified
TensorRT-LLM source files. Add the NVIDIA copyright and license header as a
comment block at the very start of the file, before the existing content (###
:title Speculative Decoding), ensuring it includes the copyright notice,
appropriate year, and license information formatted according to TensorRT-LLM
coding guidelines for Python files.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 13b1200d-5400-484e-a25f-d1516cf1400c
📒 Files selected for processing (2)
examples/llm-api/llm_speculative_decoding.pytests/integration/test_lists/waives.txt
💤 Files with no reviewable changes (1)
- tests/integration/test_lists/waives.txt
7e9ff03 to
cfc91d6
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #54392 [ run ] triggered by Bot. Commit: |
|
PR_Github #54392 [ run ] completed with state
|
cfc91d6 to
3931432
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #54576 [ run ] triggered by Bot. Commit: |
|
PR_Github #54576 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #54699 [ run ] triggered by Bot. Commit: |
|
PR_Github #54699 [ run ] completed with state
|
|
/bot run |
|
PR_Github #54797 [ run ] triggered by Bot. Commit: |
|
PR_Github #54797 [ run ] completed with state
|
3931432 to
7d06c73
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #55437 [ run ] triggered by Bot. Commit: |
|
PR_Github #55437 [ run ] completed with state
|
7d06c73 to
0a8e67e
Compare
0a8e67e to
03a2bd1
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #58192 [ run ] triggered by Bot. Commit: |
|
PR_Github #58192 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61163 [ run ] triggered by Bot. Commit: |
|
PR_Github #61163 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61205 [ run ] triggered by Bot. Commit: |
|
PR_Github #61205 [ run ] completed with state
|
fbaf25e to
da74fec
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #61264 [ run ] triggered by Bot. Commit: |
|
PR_Github #61264 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61774 [ run ] triggered by Bot. Commit: |
|
PR_Github #61774 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61806 [ run ] triggered by Bot. Commit: |
|
PR_Github #61806 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61842 [ run ] triggered by Bot. Commit: |
|
PR_Github #61842 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61863 [ run ] triggered by Bot. Commit: |
|
PR_Github #61863 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #61890 [ run ] triggered by Bot. Commit: |
|
PR_Github #61890 [ run ] completed with state
|
Signed-off-by: qgai <qgai@nvidia.com>
f97b508 to
41eab75
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #62154 [ run ] triggered by Bot. Commit: |
|
PR_Github #62154 [ run ] completed with state
|
8 participants
Summary
get_sm_version() < 90check at the top ofrun_MTP()inexamples/llm-api/llm_speculative_decoding.pythat emitslogger.warning()and returns cleanly; the subprocess returns 0, parent pytest's subprocess.wait() returns immediately, and Hopper+ behavior is unchanged. Remove the now-redundantfull:A100X/...mtpwaiver entry.Test plan
Links
Summary by CodeRabbit
Release Notes
New Features
Tests