[https://nvbugs/5634220][fix] Add developer guide back and fix some i…

[nv-guomingz]

…nvalid urls.

Summary by CodeRabbit

• Documentation

  • Updated multiple documentation links to point to specific version references (0.21.0) for TensorRT LLM performance data
  • Corrected Dynamo documentation backend references across guides

• Chores

  • Extended example configuration with new CUDA graph and MOE backend settings

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 the stage-list parameter 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.md
and the scripts/test_to_stage_mapping.py helper.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip 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-pipeline

Reuse 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.

[nv-guomingz]

/bot run

[coderabbitai]

📝 Walkthrough

Walkthrough

Multiple documentation files were updated with corrected hyperlinks—primarily changing TensorRT LLM documentation references from generic latest paths to specific versioned paths (0.21.0) and updating Dynamo backend links. A configuration file was extended with new sections for CUDA graph and MOE settings. One formatting adjustment added a blank line to a restructured text file.

Changes

Cohort / File(s)	Summary
TensorRT LLM Documentation Links (Versioned to 0.21.0) docs/source/blogs/H100vsA100.md, docs/source/blogs/H200launch.md, examples/sample_weight_stripping/README.md	Updated TensorRT LLM documentation URLs from generic latest reference to specific versioned path (0.21.0) across multiple blog and example files
Dynamo Backend Link Updates docs/source/blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.md, docs/source/features/disagg-serving.md	Updated Dynamo documentation links to point to specific backend README paths instead of general trtllm.html pages
Configuration Extensions examples/llm-api/extra-llm-api-config.yml	Added two new top-level configuration blocks: cuda_graph_config (with enable_padding and max_batch_size fields) and moe_config (with backend field)
Documentation URL Updates examples/models/core/multimodal/README.md	Updated NeVA documentation hyperlink to reflect new URL path structure
Structural Formatting docs/source/index.rst	Added blank line before Developer Guide toctree section

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

• URL updates are repetitive and low-risk across multiple files
• Configuration additions in YAML file are straightforward key-value expansions with no complex logic
• Formatting change is trivial

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The PR description is largely incomplete - it contains only the template with placeholder text (@coderabbitai summary) and no actual explanation of the issue, solution, or test coverage.	Complete the Description and Test Coverage sections with details about the bug being fixed and relevant tests. Provide a clear explanation of the changes made.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is partially related to the changeset - it mentions adding developer guide back and fixing invalid URLs, which aligns with the changes updating documentation links to versioned paths and adding a blank line to index.rst.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

📜 Recent review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between cacb8a8 and 685b562.

📒 Files selected for processing (8)

• docs/source/blogs/H100vsA100.md (1 hunks)
• docs/source/blogs/H200launch.md (1 hunks)
• docs/source/blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.md (1 hunks)
• docs/source/features/disagg-serving.md (1 hunks)
• docs/source/index.rst (1 hunks)
• examples/llm-api/extra-llm-api-config.yml (1 hunks)
• examples/models/core/multimodal/README.md (1 hunks)
• examples/sample_weight_stripping/README.md (1 hunks)

🧰 Additional context used

🧠 Learnings (13)

📓 Common learnings

Learnt from: venkywonka
Repo: NVIDIA/TensorRT-LLM PR: 6029
File: .github/pull_request_template.md:45-53
Timestamp: 2025-08-27T17:50:13.264Z
Learning: For PR templates in TensorRT-LLM, avoid suggesting changes that would increase developer overhead, such as converting plain bullets to mandatory checkboxes. The team prefers guidance-style bullets that don't require explicit interaction to reduce friction in the PR creation process.

Learnt from: farshadghodsian
Repo: NVIDIA/TensorRT-LLM PR: 7101
File: docs/source/blogs/tech_blog/blog9_Deploying_GPT_OSS_on_TRTLLM.md:36-36
Timestamp: 2025-08-21T00:16:56.457Z
Learning: TensorRT-LLM container release tags in documentation should only reference published NGC container images. The README badge version may be ahead of the actual published container versions.

📚 Learning: 2025-08-21T00:16:56.457Z

Learnt from: farshadghodsian
Repo: NVIDIA/TensorRT-LLM PR: 7101
File: docs/source/blogs/tech_blog/blog9_Deploying_GPT_OSS_on_TRTLLM.md:36-36
Timestamp: 2025-08-21T00:16:56.457Z
Learning: TensorRT-LLM container release tags in documentation should only reference published NGC container images. The README badge version may be ahead of the actual published container versions.

Applied to files:

• docs/source/features/disagg-serving.md
• docs/source/blogs/H200launch.md
• docs/source/blogs/H100vsA100.md
• docs/source/blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.md

📚 Learning: 2025-08-06T13:58:07.506Z

Learnt from: galagam
Repo: NVIDIA/TensorRT-LLM PR: 6487
File: tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py:1-12
Timestamp: 2025-08-06T13:58:07.506Z
Learning: In TensorRT-LLM, test files (files under tests/ directories) do not require NVIDIA copyright headers, unlike production source code files. Test files typically start directly with imports, docstrings, or code.

Applied to files:

• docs/source/blogs/H200launch.md
• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-08-26T09:37:10.463Z

Learnt from: jiaganc
Repo: NVIDIA/TensorRT-LLM PR: 7031
File: tensorrt_llm/bench/dataclasses/configuration.py:90-104
Timestamp: 2025-08-26T09:37:10.463Z
Learning: In TensorRT-LLM, the `get_pytorch_perf_config()` method returns `self.pytorch_config` which can contain default `cuda_graph_config` values, so `llm_args` may already have this config before the extra options processing.

Applied to files:

• examples/llm-api/extra-llm-api-config.yml

📚 Learning: 2025-08-26T09:37:10.463Z

Learnt from: jiaganc
Repo: NVIDIA/TensorRT-LLM PR: 7031
File: tensorrt_llm/bench/dataclasses/configuration.py:90-104
Timestamp: 2025-08-26T09:37:10.463Z
Learning: In TensorRT-LLM's bench configuration, the `get_pytorch_perf_config()` method returns `self.pytorch_config` which is a Dict[str, Any] that can contain default values including `cuda_graph_config`, making the fallback `llm_args["cuda_graph_config"]` safe to use.

Applied to files:

• examples/llm-api/extra-llm-api-config.yml

📚 Learning: 2025-08-08T04:10:19.038Z

Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 6728
File: cpp/tensorrt_llm/plugins/mixtureOfExperts/mixtureOfExpertsPlugin.cpp:966-966
Timestamp: 2025-08-08T04:10:19.038Z
Learning: TensorRT plugins currently don't support padding functionality, and TensorRT is not getting new features (in maintenance mode). This means that duplicating parameters like mExpertHiddenSize in function calls, even with TODO comments, can be acceptable as pragmatic solutions within these constraints.

Applied to files:

• examples/llm-api/extra-llm-api-config.yml

📚 Learning: 2025-08-11T20:09:24.389Z

Learnt from: achartier
Repo: NVIDIA/TensorRT-LLM PR: 6763
File: tests/integration/defs/triton_server/conftest.py:16-22
Timestamp: 2025-08-11T20:09:24.389Z
Learning: In the TensorRT-LLM test infrastructure, the team prefers simple, direct solutions (like hard-coding directory traversal counts) over more complex but robust approaches when dealing with stable directory structures. They accept the maintenance cost of updating tests if the layout changes.

Applied to files:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-08-27T17:50:13.264Z

Learnt from: venkywonka
Repo: NVIDIA/TensorRT-LLM PR: 6029
File: .github/pull_request_template.md:45-53
Timestamp: 2025-08-27T17:50:13.264Z
Learning: For PR templates in TensorRT-LLM, avoid suggesting changes that would increase developer overhead, such as converting plain bullets to mandatory checkboxes. The team prefers guidance-style bullets that don't require explicit interaction to reduce friction in the PR creation process.

Applied to files:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-09-09T09:40:45.658Z

Learnt from: fredricz-20070104
Repo: NVIDIA/TensorRT-LLM PR: 7645
File: tests/integration/test_lists/qa/llm_function_core.txt:648-648
Timestamp: 2025-09-09T09:40:45.658Z
Learning: In TensorRT-LLM test lists, it's common and intentional for the same test to appear in multiple test list files when they serve different purposes (e.g., llm_function_core.txt for comprehensive core functionality testing and llm_function_core_sanity.txt for quick sanity checks). This duplication allows tests to be run in different testing contexts.

Applied to files:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-07-28T17:06:08.621Z

Learnt from: moraxu
Repo: NVIDIA/TensorRT-LLM PR: 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:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-09-23T15:13:48.819Z

Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/kernels/nccl_device/multimem.h:20-30
Timestamp: 2025-09-23T15:13:48.819Z
Learning: TRT-LLM targets modern CUDA toolkits that support FP8 datatypes, so cuda_fp8.h can be included unconditionally without version guards in TRT-LLM code.

Applied to files:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-08-20T07:43:36.447Z

Learnt from: ChristinaZ
Repo: NVIDIA/TensorRT-LLM PR: 7068
File: cpp/tensorrt_llm/kernels/moeTopKFuncs.cuh:169-172
Timestamp: 2025-08-20T07:43:36.447Z
Learning: In TensorRT-LLM MOE kernels, when processing up to 128 experts across 32 threads, each thread handles at most 4 experts (N < 5 constraint), where N represents candidates per thread rather than total system capacity.

Applied to files:

• docs/source/blogs/H100vsA100.md

📚 Learning: 2025-08-19T12:45:11.997Z

Learnt from: amitz-nv
Repo: NVIDIA/TensorRT-LLM PR: 7033
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:0-0
Timestamp: 2025-08-19T12:45:11.997Z
Learning: In tensorrt_llm/_torch/pyexecutor/model_engine.py, DoRA (Delta Orthogonal Rank Adaptation) functionality was removed from the PyTorch flow to eliminate issues with inverted DoRA detection logic. The original is_dora condition was checking if scaling_vec_pointer == 0, which was potentially incorrect.

Applied to files:

• docs/source/blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.md

🪛 LanguageTool

examples/sample_weight_stripping/README.md

[style] ~244-~244: Try using a synonym here to elevate your writing.
Context: ...hitecture/checkpoint.html). Since these make up the vast majority of weights, the prune...

(CONSTITUTE_COMPRISE)

⏰ 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 (8)

docs/source/index.rst (1)

77-77: Formatting consistency improvement.

The blank line addition aligns with the spacing pattern used throughout the rest of the file between toctree sections, enhancing readability without affecting functionality.

examples/models/core/multimodal/README.md (1)

904-904: Update NeVA documentation link to official NeMo framework documentation.

The link now correctly points to the NVIDIA NeMo framework's official documentation. Using "latest" here is appropriate since this references external framework documentation rather than versioned product documentation.

docs/source/blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.md (1)

127-127: Improve Dynamo documentation link specificity.

The link now points directly to the TensorRT-LLM backend integration guide within Dynamo's documentation, providing more precise guidance for users implementing disaggregated serving with Dynamo.

docs/source/features/disagg-serving.md (1)

189-189: Align Dynamo documentation link with backend-specific integration guide.

Consistent with the related update in blog5, this link now points to the backend-specific TensorRT-LLM integration documentation within Dynamo's documentation, improving guidance for implementation.

examples/sample_weight_stripping/README.md (1)

244-244: Verify versioned documentation path (0.21.0) consistency.

This link is pinned to version 0.21.0, consistent with other TensorRT LLM documentation updates in this PR. Please confirm this version is currently published and appropriate for the audience.

Minor note: The static analysis tool flagged the phrase "make up the vast majority" as potentially redundant with "comprise." The current phrasing is fine, but if you refactor, consider the more direct: "Since these comprise the vast majority of weights, the pruner will decrease checkpoint size up to 99%."

examples/llm-api/extra-llm-api-config.yml (1)

1-5: Verify new configuration blocks are schema-valid and appropriately documented.

The configuration file introduces two new top-level sections (cuda_graph_config and moe_config). Please confirm:

1. These keys are recognized and supported by the LLM API backend configuration schema
2. The values (e.g., enable_padding: True, max_batch_size: 16, backend: trtllm) are valid and appropriate defaults for example configuration

If these are new features being introduced, consider adding inline comments to document their purpose:

# Enable padding for CUDA graphs with a max batch size of 16
cuda_graph_config:
    enable_padding: True
    max_batch_size: 16

# Configure Mixture of Experts to use TensorRT-LLM backend
moe_config:
    backend: trtllm

docs/source/blogs/H100vsA100.md (1)

31-31: Align with versioned documentation strategy.

This update pins the TensorRT LLM Performance Documentation to version 0.21.0, consistent with the same link in docs/source/blogs/H200launch.md. This ensures readers access version-consistent performance data.

Please confirm that version 0.21.0 is a currently published and stable documentation version, given this pin is used in multiple places in the PR.

docs/source/blogs/H200launch.md (1)

24-24: Let me try a different web search approach to verify the version:

Based on the verification, TensorRT-LLM Release 0.21.0 is an officially published version with documented key features including Gemma3 VLM support and large-scale EP support. Additionally, the version is actively used in current Triton NGC containers (version 25.08), confirming it is a stable and current version.

The versioned documentation link is appropriate and correctly pinned.

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[tensorrt-cicd]

PR_Github #23510 [ run ] triggered by Bot. Commit: 685b562

[tensorrt-cicd]

PR_Github #23510 [ run ] completed with state SUCCESS. Commit: 685b562
/LLM/release-1.1/L0_MergeRequest_PR pipeline #410 completed with status: 'SUCCESS'