[[TRTLLM-13948][feat] Clean up DeepSeek tests using CPP Transceiver v1 by asfiyab-nvidia · Pull Request #17090 · NVIDIA/TensorRT-LLM

asfiyab-nvidia · 2026-07-30T23:46:14Z

Description

This PR removes the DeepSeek tests that use the CPP Transceiver V1. Summary of changes:

Test Changes in `test_disaggregated_serving.py`

`TestDeepSeekV32Exp`

Removed: test_auto_dtype - redundant with test_kv_cache_v2_nixl_python (besides TP settings)

`TestDeepSeekV3Lite`

Removed tests:

test_nixl_backend - redundant with test_kv_cache_v2_nixl_python
test_auto_dtype - shares redundancies with test_kv_cache_v2_nixl_python and test_guided_decoding

Retained tests:

test_gen_only_sync - retained to test sync KV cache transfer
test_auto_dtype_with_helix - retained for Python transceiver execution path
test_guided_decoding - retained for testing guided decoding backends and ctx+gen spec dec format

The tests that were removed were also deleted from corresponding test lists

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)
If PR introduces API changes, an appropriate PR label is added - either api-compatible or api-breaking. For api-breaking, include BREAKING in the PR title.
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.

Dev Engineer Review

Removed redundant DeepSeek CPP Transceiver V1 coverage and updated retained tests to use the Python transceiver with NIXL.
Retained synchronous KV-cache transfer, Helix, and guided-decoding coverage.
Updated QA, CI, and waiver lists to remove deleted or obsolete test cases.
No apparent configuration or formatting inconsistencies were identified.

QA Engineer Review

Modified test_disaggregated_serving.py: removed test_auto_dtype from TestDeepSeekV32Exp, removed test_nixl_backend and test_auto_dtype from TestDeepSeekV3Lite, and updated test_gen_only_sync.
Updated qa/llm_function_core.txt, qa/llm_function_rtx6k.txt, test-db/l0_dgx_b200.yml, l0_dgx_b300.yml, l0_dgx_h100.yml, l0_dgx_h200.yml, and waives.txt to remove corresponding entries.
Retained tests remain represented across the relevant CI and QA lists.
Verdict: needs follow-up pending CBTS coverage data.

Signed-off-by: Asfiya Baig <asfiyab@nvidia.com>

coderabbitai · 2026-07-30T23:50:21Z

Walkthrough

Updated DeepSeek disaggregated-serving accuracy tests to use Python/NIXL configurations, removed obsolete auto-dtype cases, expanded KV-cache evaluations, and synchronized QA, platform test lists, and waiver entries.

Changes

Disaggregated-serving test coverage

Layer / File(s)	Summary
Python and NIXL runtime configuration `tests/integration/defs/accuracy/test_disaggregated_serving.py`	Synchronous generation-only coverage is Python-only; Helix and guided-decoding transfers use NIXL with the Python runtime.
Accuracy coverage adjustments `tests/integration/defs/accuracy/test_disaggregated_serving.py`	Removed obsolete DeepSeek auto-dtype tests and expanded selected KV-cache evaluations to MMLU plus GSM8K.
Configured test inventory updates `tests/integration/test_lists/...`, `tests/integration/test_lists/waives.txt`	Updated QA and platform test lists and removed obsolete auto-dtype waiver entries.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

NVIDIA/TensorRT-LLM#16908: Updates DeepSeek transceiver defaults related to the Python/NIXL configuration changes.

Suggested reviewers: tburt-nv

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title concisely summarizes the primary change—cleaning up DeepSeek tests that use CPP Transceiver V1—and includes the relevant ticket and feature type.
Description check	✅ Passed	The description clearly explains the removed and retained tests and rationale, though its required Test Coverage section is empty and the checklist remains unchecked.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

Fix failing CI checks

🧪 Generate unit tests (beta)

Create PR with unit tests

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

coderabbitai

Actionable comments posted: 1

🤖 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.

Inline comments:
In `@tests/integration/defs/accuracy/test_disaggregated_serving.py`:
- Line 1056: Update the test_gen_only_sync method with the required return type
annotation, using the appropriate annotation for its test method behavior.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 69392dbb-874d-41f6-9c44-e37bd47fcebe

📥 Commits

Reviewing files that changed from the base of the PR and between 5c5ef98 and 36f4732.

📒 Files selected for processing (8)

tests/integration/defs/accuracy/test_disaggregated_serving.py
tests/integration/test_lists/qa/llm_function_core.txt
tests/integration/test_lists/qa/llm_function_rtx6k.txt
tests/integration/test_lists/test-db/l0_dgx_b200.yml
tests/integration/test_lists/test-db/l0_dgx_b300.yml
tests/integration/test_lists/test-db/l0_dgx_h100.yml
tests/integration/test_lists/test-db/l0_dgx_h200.yml
tests/integration/test_lists/waives.txt

💤 Files with no reviewable changes (7)

tests/integration/test_lists/test-db/l0_dgx_b300.yml
tests/integration/test_lists/qa/llm_function_rtx6k.txt
tests/integration/test_lists/test-db/l0_dgx_h100.yml
tests/integration/test_lists/test-db/l0_dgx_h200.yml
tests/integration/test_lists/test-db/l0_dgx_b200.yml
tests/integration/test_lists/waives.txt
tests/integration/test_lists/qa/llm_function_core.txt

coderabbitai · 2026-07-30T23:50:24Z

-                             ids=["python", "cpp"])
-    def test_gen_only_sync(self, transceiver_runtime):
-        """Test gen-only synchronous KV transfer with each NIXL runtime.
+    def test_gen_only_sync(self):


📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Add the required return annotation.

-def test_gen_only_sync(self): +def test_gen_only_sync(self) -> None:

As per coding guidelines, every Python function must be annotated.

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

def test_gen_only_sync(self):

def test_gen_only_sync(self) -> None:

🤖 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 `@tests/integration/defs/accuracy/test_disaggregated_serving.py` at line 1056, Update the test_gen_only_sync method with the required return type annotation, using the appropriate annotation for its test method behavior.

Source: Coding guidelines

chienchunhung

test_gen_only_sync is no longer parametrized, so pytest no longer generates the [python] node ID. Please update this test-list entry to ...::test_gen_only_sync; otherwise the QA selection references a nonexistent test and the retained synchronous-transfer coverage will not run.

nv-xtf · 2026-07-31T02:41:30Z

@@ -1255,7 +1175,8 @@ def test_auto_dtype_with_helix(self, comms_medium, cuda_graph_config,
            "enable_chunked_prefill": False,
            "cuda_graph_config": None,
            "cache_transceiver_config": {
-                "backend": "UCX",
+                "backend": "NIXL",


For test_auto_dtype_with_helix — could you double-check whether the Python transceiver actually supports cp_size > 1? From what I can see, KvCacheTransceiverV2._check_compatible rejects cp_size != 1 (tensorrt_llm/_torch/disaggregation/transceiver.py), and Helix relies on context parallelism on the gen side, so this test likely can't be migrated and has to stay on the C++ transceiver.

[[TRTLLM-13948][feat] Clean up DeepSeek tests using CPP Transceiver v1

36f4732

Signed-off-by: Asfiya Baig <asfiyab@nvidia.com>

asfiyab-nvidia requested review from a team as code owners July 30, 2026 23:46

asfiyab-nvidia requested review from BowenFu, chienchunhung, crazydemo, joyang-nv, longlee0622, mzweilz, niukuo, nv-xtf, ruodil, schetlur-nv and yingguo-trt July 30, 2026 23:46

github-actions Bot assigned asfiyab-nvidia Jul 30, 2026

coderabbitai Bot reviewed Jul 30, 2026

View reviewed changes

asfiyab-nvidia marked this pull request as draft July 30, 2026 23:51

chienchunhung reviewed Jul 31, 2026

View reviewed changes

yingguo-trt approved these changes Jul 31, 2026

View reviewed changes

nv-xtf requested changes Jul 31, 2026

View reviewed changes

This was referenced Jul 31, 2026

[None][fix] reserve draft KV at disagg transition for gen-only-dspark case #16990

Merged

[https://nvbugs/6427411][test] Re-enable PP regression tests #17117

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[[TRTLLM-13948][feat] Clean up DeepSeek tests using CPP Transceiver v1 - #17090

[[TRTLLM-13948][feat] Clean up DeepSeek tests using CPP Transceiver v1#17090
asfiyab-nvidia wants to merge 1 commit into
NVIDIA:mainfrom
asfiyab-nvidia:asfiyab/ds-remove-cpp-transceiver-tests

asfiyab-nvidia commented Jul 30, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Jul 30, 2026

Uh oh!

coderabbitai Bot left a comment

Uh oh!

coderabbitai Bot Jul 30, 2026

Uh oh!

chienchunhung left a comment

Uh oh!

nv-xtf Jul 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

	def test_gen_only_sync(self):
	def test_gen_only_sync(self) -> None:

Uh oh!

Conversation

asfiyab-nvidia commented Jul 30, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Test Changes in test_disaggregated_serving.py

TestDeepSeekV32Exp

TestDeepSeekV3Lite

Test Coverage

PR Checklist

GitHub Bot Help

Dev Engineer Review

QA Engineer Review

Uh oh!

coderabbitai Bot commented Jul 30, 2026

Walkthrough

Changes

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

coderabbitai Bot Jul 30, 2026

Choose a reason for hiding this comment

Uh oh!

chienchunhung left a comment

Choose a reason for hiding this comment

Uh oh!

nv-xtf Jul 31, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

asfiyab-nvidia commented Jul 30, 2026 •

edited by coderabbitai Bot

Loading

Test Changes in `test_disaggregated_serving.py`

`TestDeepSeekV32Exp`

`TestDeepSeekV3Lite`