[nvbug/5361223] doc: Update Llama4 deployment guide: update config & note concurrency #6222
Conversation
Signed-off-by: raayandhar <[email protected]>
WalkthroughThe documentation for the Llama4 Maverick Eagle guide was updated. The Docker run command example reordered parameters, added the Changes
Estimated code review effort1 (~2 minutes) Poem
📜 Recent review detailsConfiguration used: .coderabbit.yaml 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
⏰ 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)
🪧 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
|
|
cc: @pcastonguay |
![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 (3)
docs/source/blogs/tech_blog/blog6_Llama4_maverick_eagle_guide.md (3)
71-77: Inlineecho -eYAML is brittle – switch to a HEREDOC for clarityPacking ~15 YAML lines into a single
echo -ecommand is hard to read, easy to break, and hides indentation errors that will make the server silently fall back to defaults.
Consider a HEREDOC so users can copy-paste / tweak the config without fighting shell-escaping:- docker.io/<username>/tensorrt_llm:main sh \ - -c "echo -e 'disable_overlap_scheduler: false\nenable_autotuner: false\nenable_attention_dp: false\nenable_min_latency: true\ncuda_graph_config:\n max_batch_size: 8\nspeculative_config:\n decoding_type: Eagle\n max_draft_len: 3\n speculative_model_dir: /config/models/eagle\n eagle3_one_model: true\nkv_cache_config:\n enable_block_reuse: false' > c.yaml && \ + docker.io/<username>/tensorrt_llm:main sh -c '\ +cat > /tmp/c.yaml <<EOF +disable_overlap_scheduler: false +enable_autotuner: false +enable_attention_dp: false +enable_min_latency: true +cuda_graph_config: + max_batch_size: 8 +speculative_config: + decoding_type: Eagle + max_draft_len: 3 + speculative_model_dir: /config/models/eagle + eagle3_one_model: true +kv_cache_config: + enable_block_reuse: false +EOF +TRT_LLM_DISABLE_LOAD_WEIGHTS_IN_PARALLEL=True \ +trtllm-serve /config/models/maverick --host 0.0.0.0 --port 8000 \ + --backend pytorch --tp_size 8 --ep_size 1 \ + --trust_remote_code --extra_llm_api_options /tmp/c.yaml \ + --kv_cache_free_gpu_memory_fraction 0.75'This keeps the guide readable and reduces copy-paste errors.
71-71: Double-negative flagdisable_overlap_scheduler: falsemay confuse readersBecause the key is phrased as a “disable” switch, setting it to
falseactually means “enabled”. A brief parenthetical comment (e.g., “set to false → overlap scheduler ON”) will save newcomers a head-scratch.
144-147: Provide actionable knobs for high-concurrency tuningThe note correctly warns about TPS degradation, but readers will be left wondering what to tweak. Add one-sentence guidance such as:
“For multi-user workloads, increase
max_batch_size, enableenable_overlap_scheduler, and consider settingenable_min_latency: falseto trade a small latency increase for substantially higher aggregate TPS.”This makes the section immediately useful.
|
@jiahanc since you created this file, could you review the changes? Thanks. |
docs/source/blogs/tech_blog/blog6_Llama4_maverick_eagle_guide.md
Outdated
Show resolved
Hide resolved
Signed-off-by: raayandhar <[email protected]>
|
/bot skip --comment "Doc update" |
|
PR_Github #12596 [ skip ] triggered by Bot |
|
PR_Github #12596 [ skip ] completed with state |
…note concurrency (NVIDIA#6222) Signed-off-by: raayandhar <[email protected]> Signed-off-by: Yang Li <[email protected]>
…note concurrency (NVIDIA#6222) Signed-off-by: raayandhar <[email protected]> Signed-off-by: Shreyas Misra <[email protected]>
…note concurrency (NVIDIA#6222) Signed-off-by: raayandhar <[email protected]> Signed-off-by: Ransiki Zhang <[email protected]>
Summary by CodeRabbit
[nvbug/5361223] doc: Update Llama4 deployment guide: update config & note concurrency
Description
Small update to the serving config and added a small note on how increasing concurrency degrades TPS/user performance rapidly.
Test Coverage
Tested locally with TRTLLM, cutoff at commit ae28b3a (jul. 17).
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.