[None][chore] Add Dynamo configs to TRTLLM CI - Agg

[brb-nv]

Description

This MR adds Dynamo configs to TRTLLM CI to catch issues early. This MR has agg configs.

Test Coverage

N/A

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

• Update tava architecture diagram if there is a significant design change in PR.

• 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

To see a list of available CI bot commands, please comment /bot help.

Summary by CodeRabbit

• New Features

  • Added performance sanity test support for DGX H200 (8 GPUs), GB200 multi-GPU/multi-node, and B200 platforms.
  • Introduced test configurations for new AI models: Qwen3 32B FP8, Qwen3 235B A22B FP8, DeepSeek-V3.2 FP4, K25 Thinking FP4, and GPT-OSS-120B FP4.

• Tests

  • Extended model path dictionary with Qwen3 32B FP8 support.
  • Added multiple new performance sanity test entries across test suites for model/hardware variant combinations.

[brb-nv]

Question for infra reviewer: This is adding a new stage. Is there anything else we need to do?

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR adds support for new performance sanity test configurations and models. Changes include a Jenkins stage definition for DGX H200 GPUs, registration of the Qwen3 32B FP8 model, new test entries across multiple test lists, and performance sanity YAML configurations for model deployments (Qwen3, GPT-OSS-120B, DeepSeek-V3.2, K25-thinking) on Blackwell and Hopper GPUs with specific tensor parallelism and throughput constraints.

Changes

Cohort / File(s)	Summary
CI Pipeline Configuration jenkins/L0_Test.groovy	Added a new Slurm perf-sanity post-merge stage entry for DGX H200 with 8 GPUs, mapping to platform dgx-h200-x8 and YAML config l0_dgx_h200_perf_sanity.
Model Registry tests/integration/defs/perf/test_perf_sanity.py	Extended MODEL_PATH_DICT with new model key qwen3_32b_fp8 mapped to Hugging Face model path "Qwen3/Qwen3-32B-FP8".
Test List Configurations tests/integration/test_lists/test-db/l0_*.yml	Added new perf sanity test entries across multiple test suites: l0_b200_multi_gpus_perf_sanity.yml (1 entry for Qwen3 variant), l0_gb200_multi_gpus_perf_sanity.yml (2 entries for GPT-OSS), l0_gb200_multi_nodes_perf_sanity_node2_gpu8.yml (1 entry for DeepSeek multi-node), and created new l0_dgx_h200_perf_sanity.yml with 2 performance sanity tests for K25-thinking and Qwen3 variants.
Perf Sanity YAML Configs tests/scripts/perf-sanity/aggregated/dynamo_*.yaml	Added 6 new performance sanity configuration files defining model metadata, hardware layout, server configs (tensor/expert/pipeline parallelism, batch sizes, throughput limits), TRT-LLM attention backends, KV cache settings, and client workload definitions for: Qwen3 235B/32B FP8 (H200), GPT-OSS-120B FP4 (B200), DeepSeek-V3.2 FP4 (GB200 multi-node), and K25-thinking FP4 (B200).

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~22 minutes

Possibly related PRs

• [None][infra] Add K2.5 Perf Tests into CI #12931: Modifies the same CI test mapping file (jenkins/L0_Test.groovy) and test model registry (tests/integration/defs/perf/test_perf_sanity.py), adding new stage entries and model path mappings.

Suggested reviewers

• mlefeb01
• EmmaQiaoCh
• dc3671
• ruodil
• ZhanruiSunCh

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: adding Dynamo configs to TRTLLM CI, specifically aggregated configs.
Description check	✅ Passed	The PR description follows the template structure with Description and Test Coverage sections filled in, though Test Coverage is marked N/A and the description explanation is minimal.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

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/test_perf_sanity.py (1)

1-1: ⚠️ Potential issue | 🟠 Major

Update the copyright year in the modified Python source header.

Line 1 still says 2022-2025 even though this file is modified in 2026.

✅ Proposed fix

-# SPDX-FileCopyrightText: Copyright (c) 2022-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-FileCopyrightText: Copyright (c) 2022-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.

As per coding guidelines, "All TensorRT-LLM source files must contain an NVIDIA copyright header with the year of latest meaningful modification."

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/integration/defs/perf/test_perf_sanity.py` at line 1, Update the source
header line that currently reads "2022-2025" to include the latest modification
year (e.g., change to "2022-2026") so the file header reflects the 2026
modification; locate the top-of-file copyright header in test_perf_sanity.py and
replace the year range accordingly.

🧹 Nitpick comments (1)

tests/scripts/perf-sanity/aggregated/dynamo_qwen3_235b_a22b_fp8_hopper.yaml (1)

15-15: Harden remote-code execution in CI configs.

trust_remote_code: true is understandable for some model stacks, but it weakens CI safety. Please document why it is required here and prefer pinned/immutable model revisions when remote code is enabled.

Also applies to: 47-47

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/scripts/perf-sanity/aggregated/dynamo_qwen3_235b_a22b_fp8_hopper.yaml`
at line 15, The CI config enables unsafe remote-code execution via the YAML key
trust_remote_code: true; update this to either trust_remote_code: false or, if
remote code is required for this model, add an inline comment and a pinned
immutable model revision (e.g., set model_revision to a specific commit/sha or
exact model version) and document in the file why remote code is necessary for
the model stack so reviewers can audit the change; ensure the same change is
applied where trust_remote_code appears elsewhere in this file.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Outside diff comments:
In `@tests/integration/defs/perf/test_perf_sanity.py`:
- Line 1: Update the source header line that currently reads "2022-2025" to
include the latest modification year (e.g., change to "2022-2026") so the file
header reflects the 2026 modification; locate the top-of-file copyright header
in test_perf_sanity.py and replace the year range accordingly.

---

Nitpick comments:
In `@tests/scripts/perf-sanity/aggregated/dynamo_qwen3_235b_a22b_fp8_hopper.yaml`:
- Line 15: The CI config enables unsafe remote-code execution via the YAML key
trust_remote_code: true; update this to either trust_remote_code: false or, if
remote code is required for this model, add an inline comment and a pinned
immutable model revision (e.g., set model_revision to a specific commit/sha or
exact model version) and document in the file why remote code is necessary for
the model stack so reviewers can audit the change; ensure the same change is
applied where trust_remote_code appears elsewhere in this file.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 51dc1d5b-dcac-462b-a4c0-524e8ccaedca

📥 Commits

Reviewing files that changed from the base of the PR and between 813d877 and 12e9200.

📒 Files selected for processing (11)

• jenkins/L0_Test.groovy
• tests/integration/defs/perf/test_perf_sanity.py
• tests/integration/test_lists/test-db/l0_b200_multi_gpus_perf_sanity.yml
• tests/integration/test_lists/test-db/l0_dgx_h200_perf_sanity.yml
• tests/integration/test_lists/test-db/l0_gb200_multi_gpus_perf_sanity.yml
• tests/integration/test_lists/test-db/l0_gb200_multi_nodes_perf_sanity_node2_gpu8.yml
• tests/scripts/perf-sanity/aggregated/dynamo_deepseek_v32_fp4_2_nodes_grace_blackwell.yaml
• tests/scripts/perf-sanity/aggregated/dynamo_gpt_oss_120b_fp4_blackwell.yaml
• tests/scripts/perf-sanity/aggregated/dynamo_k25_thinking_fp4_blackwell.yaml
• tests/scripts/perf-sanity/aggregated/dynamo_qwen3_235b_a22b_fp8_hopper.yaml
• tests/scripts/perf-sanity/aggregated/dynamo_qwen3_32b_fp8_hopper.yaml

[brb-nv]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #44078 [ run ] triggered by Bot. Commit: afecdd0 Link to invocation