[None][doc] blog: Scaling Expert Parallelism in TensorRT-LLM (Part 2: Performance Status and Optimization) #6547
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Fix typos Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Add online EPLB optimization part for large EP blog 2 Signed-off-by: Dongxu Yang <dongxuy@nvidia.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> update moe aux kernels Signed-off-by: Enwei Zhu <21126786+syuoni@users.noreply.github.com> scale img Signed-off-by: Enwei Zhu <21126786+syuoni@users.noreply.github.com> some updates on doc Signed-off-by: Dongxu Yang <dongxuy@nvidia.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Update image Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> update perf graphs Signed-off-by: Xianjie <5410381+qiaoxj07@users.noreply.github.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> allgather optimization Signed-off-by: Fred Wei <fredw@nvidia.com> update perf graphs Signed-off-by: Xianjie <5410381+qiaoxj07@users.noreply.github.com> Update e2e performance description Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Add host overhead description Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Polish Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> add deatails for allgather opt Signed-off-by: Fred Wei <fredw@nvidia.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Add Acknowledgements Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> Update Signed-off-by: Kaiyu Xie <26294424+kaiyux@users.noreply.github.com> add png for moe communication kernel Signed-off-by: Fred Wei <fredw@nvidia.com>
Contributor
📝 WalkthroughWalkthroughA new technical blog post detailing performance optimizations for large-scale Expert Parallelism in TensorRT-LLM was added. Additionally, an existing blog post was updated to reference that reproduction scripts are now available in the main branch, replacing previous instructions to use a feature branch. Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant Blog4 (Existing)
participant Blog8 (New)
User->>Blog4: Reads updated reproduction steps (main branch)
User->>Blog8: Reads new optimization and performance details
Blog8->>TensorRT-LLM: Describes kernel/comm optimizations, EPLB, MTP, host overhead reduction
Blog8->>User: Provides benchmarks, scripts, and future work
Estimated code review effort🎯 2 (Simple) | ⏱️ ~8 minutes Suggested labels
Suggested reviewers
Note ⚡️ Unit Test Generation is now available in beta!Learn more here, or try it out under "Finishing Touches" below. ✨ 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. Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
Contributor
There was a problem hiding this comment.
Actionable comments posted: 0
🔭 Outside diff range comments (1)
docs/source/blogs/tech_blog/blog4_Scaling_Expert_Parallelism_in_TensorRT-LLM.md (1)
518-523: Broken reference contradicts the new “main-branch” statementLine 518 states that reproduction scripts are now in main, yet the hyperlink on Line 522 still points to the historical
feat/large-epbranch.
Please update the URL so users are not redirected to a stale branch.-Please, refer to the [EP Load Balancer example](https://github.com/NVIDIA/TensorRT-LLM/tree/feat/large-ep/examples/ep_load_balancer) ... +Please, refer to the [EP Load Balancer example](https://github.com/NVIDIA/TensorRT-LLM/tree/main/examples/ep_load_balancer) ...
🧹 Nitpick comments (2)
docs/source/blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.md (2)
44-47: Add alt-text for all embedded imagesMarkdown-lint (MD045) flags every
<img>tag because thealtattribute is missing.
Alt-text is required for accessibility and helps when the image cannot be loaded.Example fix for Figure 1 (repeat for all images):
- <img src="../media/tech_blog8_kernel_breakdown.png" width="1000"> + <img src="../media/tech_blog8_kernel_breakdown.png" width="1000" alt="Kernel-time breakdown across EP sizes">Also applies to: 60-63, 72-75, 88-93, 280-307
254-260: NVTX environment-variable snippet renders as a prompt, not a plain blockThe leading
>in front of$makes the snippet look like shell output rather than commands to copy-paste.
For consistency with earlier code blocks, drop the prompt symbols inside fenced code blocks:->$ cat /sys/kernel/mm/transparent_hugepage/enabled -... +$ cat /sys/kernel/mm/transparent_hugepage/enabled +...
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (8)
docs/source/blogs/media/tech_blog8_communication_kernel.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_kernel_breakdown.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_moe_aux_kernels1.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_moe_aux_kernels2.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_perf-1k-1k-dep.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_perf-4k-1k-dep.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_perf-8k-1k-dep.pngis excluded by!**/*.pngdocs/source/blogs/media/tech_blog8_perf-8k-1k-e2e-mtp.pngis excluded by!**/*.png
📒 Files selected for processing (2)
docs/source/blogs/tech_blog/blog4_Scaling_Expert_Parallelism_in_TensorRT-LLM.md(2 hunks)docs/source/blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.md(1 hunks)
🧰 Additional context used
🧠 Learnings (3)
📓 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: in tensorrt-llm testing, it's common to have both cli flow tests (test_cli_flow.py) and pytorch api ...
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:
docs/source/blogs/tech_blog/blog4_Scaling_Expert_Parallelism_in_TensorRT-LLM.mddocs/source/blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.md
📚 Learning: in the tensorrt-llm waive list merging system, removed lines are always located at the end of the me...
Learnt from: yiqingy0
PR: NVIDIA/TensorRT-LLM#5198
File: jenkins/mergeWaiveList.py:0-0
Timestamp: 2025-07-22T08:33:49.109Z
Learning: In the TensorRT-LLM waive list merging system, removed lines are always located at the end of the merge waive lists, which is why the mergeWaiveList.py script uses reverse traversal - it's an optimization for this specific domain constraint.
Applied to files:
docs/source/blogs/tech_blog/blog4_Scaling_Expert_Parallelism_in_TensorRT-LLM.md
🪛 LanguageTool
docs/source/blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.md
[style] ~50-~50: Consider shortening this phrase to strengthen your wording.
Context: ...P without EPLB.
(MAKE_IMPROVEMENTS)
🪛 markdownlint-cli2 (0.17.2)
docs/source/blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.md
5-5: Emphasis used instead of a heading
(MD036, no-emphasis-as-heading)
45-45: Images should have alternate text (alt text)
(MD045, no-alt-text)
60-60: Images should have alternate text (alt text)
(MD045, no-alt-text)
73-73: Images should have alternate text (alt text)
(MD045, no-alt-text)
90-90: Images should have alternate text (alt text)
(MD045, no-alt-text)
282-282: Images should have alternate text (alt text)
(MD045, no-alt-text)
289-289: Images should have alternate text (alt text)
(MD045, no-alt-text)
296-296: Images should have alternate text (alt text)
(MD045, no-alt-text)
305-305: Images should have alternate text (alt text)
(MD045, no-alt-text)
⏰ 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
Member
Author
|
/bot skip --comment "doc changes" |
Collaborator
|
PR_Github #13790 [ skip ] triggered by Bot |
juney-nvidia
approved these changes
Aug 1, 2025
Collaborator
|
PR_Github #13790 [ skip ] completed with state |
lancelly
pushed a commit
to lancelly/TensorRT-LLM
that referenced
this pull request
Aug 6, 2025
… Performance Status and Optimization) (NVIDIA#6547) Signed-off-by: Kaiyu XIe <26294424+kaiyux@users.noreply.github.com> Signed-off-by: Lanyu Liao <lancelly@users.noreply.github.com>
jain-ria
pushed a commit
to jain-ria/TensorRT-LLM
that referenced
this pull request
Aug 7, 2025
… Performance Status and Optimization) (NVIDIA#6547) Signed-off-by: Kaiyu XIe <26294424+kaiyux@users.noreply.github.com>
This was referenced Aug 19, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
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.