[TRTLLM-13696][test] Part2.3: Migrate CPU only tests - auto deploy - #17314
Merged
tongyuantongyu merged 2 commits intoAug 11, 2026
Conversation
Member
Author
|
/bot run --stage-list "CPU-Generic-x86-1, CPU-Generic-arm-1" --disable-fail-fast |
Collaborator
|
PR_Github #64046 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #64046 [ run ] completed with state |
Member
Author
|
/bot run --disable-fail-fast |
Collaborator
|
PR_Github #64083 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #64083 [ run ] completed with state
|
tongyuantongyu
marked this pull request as ready for review
August 6, 2026 02:20
tongyuantongyu
requested review from
QiJune,
StanleySun639,
govind-ramnarayan,
mzweilz,
niukuo and
schetlur-nv
August 6, 2026 02:20
Contributor
WalkthroughAutoDeploy CPU coverage now includes integration, unit, model, compile, transformation, utility, multigpu, and standalone suites. Shim tests use shared runtime mocks. Quantization tests use conditional bias construction. ChangesCPU test coverage
Estimated code review effort: 3 (Moderate) | ~20 minutes Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py (1)
33-56: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winAdd return annotations to the changed fixtures and context managers.
tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py#L33-L56: annotate_mock_autodeploy_runtime_dependenciesas an iterator yieldingNone.tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py#L136-L160: annotate both context managers with their yielded mock types.tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py#L117-L120: annotatetest_config_paramswith a precise iterator/configuration type.As per coding guidelines, “Annotate every function, use
Nonefor procedures.”🤖 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/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py` around lines 33 - 56, Annotate all specified test fixtures and context managers with return types: in tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py lines 33-56, update _mock_autodeploy_runtime_dependencies to return an iterator yielding None; in lines 136-160, annotate both context managers with the precise mock types they yield; and in tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py lines 117-120, annotate test_config_params with its precise iterator/configuration type. Use None for procedures and preserve the existing fixture behavior.Source: Coding guidelines
🤖 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/unittest/auto_deploy/singlegpu/shim/test_llm_config.py`:
- Around line 117-138: Update the test_config_params fixture to patch the
get_device_count symbol actually used by LlmArgs, from
tensorrt_llm.llmapi.utils, or remove the patch if device detection should remain
real. Ensure the fixture’s device-count setup aligns with the gpus_per_node=1
scenario so the value under test is controlled correctly.
---
Nitpick comments:
In `@tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py`:
- Around line 33-56: Annotate all specified test fixtures and context managers
with return types: in
tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py lines
33-56, update _mock_autodeploy_runtime_dependencies to return an iterator
yielding None; in lines 136-160, annotate both context managers with the precise
mock types they yield; and in
tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py lines 117-120,
annotate test_config_params with its precise iterator/configuration type. Use
None for procedures and preserve the existing fixture behavior.
🪄 Autofix
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: 9b1812d7-2483-45dc-bea1-7d1ea29c1480
📒 Files selected for processing (79)
tests/integration/test_lists/test-db/l0_cpu.ymltests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_gpt_oss_modeling.pytests/unittest/auto_deploy/multigpu/custom_ops/test_ad_dist_strategies.pytests/unittest/auto_deploy/multigpu/transformations/library/test_tp_sharding.pytests/unittest/auto_deploy/singlegpu/compile/test_captured_graph.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_runner.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_utils.pytests/unittest/auto_deploy/singlegpu/custom_ops/attention/test_torch_attention_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_torch_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_triton_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/moe/test_mxfp4_moe_layout.pytests/unittest/auto_deploy/singlegpu/custom_ops/quantization/test_quant.pytests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_gptq_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_resource_handlers.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_update_kv_cache.pytests/unittest/auto_deploy/singlegpu/mlir/test_decompose.pytests/unittest/auto_deploy/singlegpu/mlir/test_dialect.pytests/unittest/auto_deploy/singlegpu/mlir/test_elementwise_fusion_e2e.pytests/unittest/auto_deploy/singlegpu/mlir/test_fx_mlir_roundtrip.pytests/unittest/auto_deploy/singlegpu/mlir/test_string_codegen.pytests/unittest/auto_deploy/singlegpu/mlir/test_subgraph_discovery.pytests/unittest/auto_deploy/singlegpu/models/test_cohere_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_deepseek_v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_eagle.pytests/unittest/auto_deploy/singlegpu/models/test_exaone_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma3n_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_lite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm_moe_dsa_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_granite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_hf.pytests/unittest/auto_deploy/singlegpu/models/test_hunyuan_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_internlm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_kimi_k2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_minimax_m2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_olmo3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_openelm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_5_moe.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_next_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_seed_oss_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_skywork_r1v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_smollm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_starcoder2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_step3p7_modeling.pytests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.pytests/unittest/auto_deploy/singlegpu/shim/test_llm_config.pytests/unittest/auto_deploy/singlegpu/test_graph_canonicalize.pytests/unittest/auto_deploy/singlegpu/test_hf_export_info.pytests/unittest/auto_deploy/singlegpu/test_mistral_small_4_tokenizer_bridge.pytests/unittest/auto_deploy/singlegpu/test_pattern_matcher.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_bmm_moe_hooks.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_silu_mul.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_swiglu.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_gemm_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_kvcache_vswa_metadata.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_mrope_delta_cache.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quant_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quantization.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_shared_kv_attention.pytests/unittest/auto_deploy/singlegpu/transformations/test_export.pytests/unittest/auto_deploy/singlegpu/transformations/test_pipeline_cache.pytests/unittest/auto_deploy/singlegpu/utils/test_config.pytests/unittest/auto_deploy/singlegpu/utils/test_create_derived_custom_op.pytests/unittest/auto_deploy/singlegpu/utils/test_dist_config.pytests/unittest/auto_deploy/singlegpu/utils/test_example_configs.pytests/unittest/auto_deploy/singlegpu/utils/test_multi_stream_utils.pytests/unittest/auto_deploy/singlegpu/utils/test_node_utils_sharding.pytests/unittest/auto_deploy/singlegpu/utils/test_quantization_utils.pytests/unittest/auto_deploy/standalone/test_standalone_import.py
tongyuantongyu
force-pushed
the
ytong/host-only-test-ad
branch
from
August 6, 2026 02:52
4588a83 to
1ad436e
Compare
Contributor
|
Note GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer. |
Contributor
There was a problem hiding this comment.
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/unittest/auto_deploy/singlegpu/custom_ops/quantization/test_quant.py`:
- Line 153: Rename the local variable input to input_tensor in the affected
test, and update both operator calls to use input_tensor so the built-in input
is no longer shadowed.
🪄 Autofix
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: 821ebf12-1a09-443a-a191-d8880304ac4f
📒 Files selected for processing (79)
tests/integration/test_lists/test-db/l0_cpu.ymltests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_gpt_oss_modeling.pytests/unittest/auto_deploy/multigpu/custom_ops/test_ad_dist_strategies.pytests/unittest/auto_deploy/multigpu/transformations/library/test_tp_sharding.pytests/unittest/auto_deploy/singlegpu/compile/test_captured_graph.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_runner.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_utils.pytests/unittest/auto_deploy/singlegpu/custom_ops/attention/test_torch_attention_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_torch_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_triton_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/moe/test_mxfp4_moe_layout.pytests/unittest/auto_deploy/singlegpu/custom_ops/quantization/test_quant.pytests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_gptq_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_resource_handlers.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_update_kv_cache.pytests/unittest/auto_deploy/singlegpu/mlir/test_decompose.pytests/unittest/auto_deploy/singlegpu/mlir/test_dialect.pytests/unittest/auto_deploy/singlegpu/mlir/test_elementwise_fusion_e2e.pytests/unittest/auto_deploy/singlegpu/mlir/test_fx_mlir_roundtrip.pytests/unittest/auto_deploy/singlegpu/mlir/test_string_codegen.pytests/unittest/auto_deploy/singlegpu/mlir/test_subgraph_discovery.pytests/unittest/auto_deploy/singlegpu/models/test_cohere_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_deepseek_v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_eagle.pytests/unittest/auto_deploy/singlegpu/models/test_exaone_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma3n_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_lite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm_moe_dsa_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_granite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_hf.pytests/unittest/auto_deploy/singlegpu/models/test_hunyuan_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_internlm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_kimi_k2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_minimax_m2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_olmo3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_openelm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_5_moe.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_next_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_seed_oss_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_skywork_r1v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_smollm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_starcoder2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_step3p7_modeling.pytests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.pytests/unittest/auto_deploy/singlegpu/shim/test_llm_config.pytests/unittest/auto_deploy/singlegpu/test_graph_canonicalize.pytests/unittest/auto_deploy/singlegpu/test_hf_export_info.pytests/unittest/auto_deploy/singlegpu/test_mistral_small_4_tokenizer_bridge.pytests/unittest/auto_deploy/singlegpu/test_pattern_matcher.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_bmm_moe_hooks.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_silu_mul.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_swiglu.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_gemm_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_kvcache_vswa_metadata.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_mrope_delta_cache.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quant_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quantization.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_shared_kv_attention.pytests/unittest/auto_deploy/singlegpu/transformations/test_export.pytests/unittest/auto_deploy/singlegpu/transformations/test_pipeline_cache.pytests/unittest/auto_deploy/singlegpu/utils/test_config.pytests/unittest/auto_deploy/singlegpu/utils/test_create_derived_custom_op.pytests/unittest/auto_deploy/singlegpu/utils/test_dist_config.pytests/unittest/auto_deploy/singlegpu/utils/test_example_configs.pytests/unittest/auto_deploy/singlegpu/utils/test_multi_stream_utils.pytests/unittest/auto_deploy/singlegpu/utils/test_node_utils_sharding.pytests/unittest/auto_deploy/singlegpu/utils/test_quantization_utils.pytests/unittest/auto_deploy/standalone/test_standalone_import.py
🚧 Files skipped from review as they are similar to previous changes (77)
- tests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_gpt_oss_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_gemm_fusion.py
- tests/unittest/auto_deploy/singlegpu/models/test_deepseek_v2_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/test_pipeline_cache.py
- tests/unittest/auto_deploy/singlegpu/transformations/test_export.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_triton_mla_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_step3p7_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_swiglu.py
- tests/unittest/auto_deploy/singlegpu/models/test_skywork_r1v2_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/attention/test_torch_attention_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_lite_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_llama3_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm_moe_dsa_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_kimi_k2_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_resource_handlers.py
- tests/unittest/auto_deploy/singlegpu/utils/test_dist_config.py
- tests/unittest/auto_deploy/multigpu/custom_ops/test_ad_dist_strategies.py
- tests/unittest/auto_deploy/singlegpu/utils/test_config.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_string_codegen.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/moe/test_mxfp4_moe_layout.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_torch_mla_op.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_mrope_delta_cache.py
- tests/unittest/auto_deploy/singlegpu/compile/test_piecewise_runner.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_fx_mlir_roundtrip.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_5_moe.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_quantization.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_gptq_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_next_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_llama4_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_kvcache_vswa_metadata.py
- tests/unittest/auto_deploy/singlegpu/compile/test_piecewise_utils.py
- tests/unittest/auto_deploy/singlegpu/models/test_seed_oss_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_internlm3_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_openelm_modeling.py
- tests/unittest/auto_deploy/singlegpu/test_mistral_small_4_tokenizer_bridge.py
- tests/unittest/auto_deploy/singlegpu/utils/test_node_utils_sharding.py
- tests/unittest/auto_deploy/singlegpu/models/test_mistral_modeling.py
- tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py
- tests/unittest/auto_deploy/singlegpu/utils/test_example_configs.py
- tests/unittest/auto_deploy/singlegpu/utils/test_multi_stream_utils.py
- tests/unittest/auto_deploy/singlegpu/models/test_smollm3_modeling.py
- tests/unittest/auto_deploy/standalone/test_standalone_import.py
- tests/unittest/auto_deploy/singlegpu/test_graph_canonicalize.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_shared_kv_attention.py
- tests/unittest/auto_deploy/singlegpu/models/test_starcoder2_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_minimax_m2_modeling.py
- tests/integration/test_lists/test-db/l0_cpu.yml
- tests/unittest/auto_deploy/singlegpu/utils/test_quantization_utils.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_decompose.py
- tests/unittest/auto_deploy/singlegpu/test_pattern_matcher.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_silu_mul.py
- tests/unittest/auto_deploy/singlegpu/models/test_eagle.py
- tests/unittest/auto_deploy/singlegpu/models/test_olmo3_modeling.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_dialect.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_elementwise_fusion_e2e.py
- tests/unittest/auto_deploy/singlegpu/utils/test_create_derived_custom_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_subgraph_discovery.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_bmm_moe_hooks.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma2_modeling.py
- tests/unittest/auto_deploy/multigpu/transformations/library/test_tp_sharding.py
- tests/unittest/auto_deploy/singlegpu/compile/test_captured_graph.py
- tests/unittest/auto_deploy/singlegpu/models/test_mistral3_modeling.py
- tests/unittest/auto_deploy/singlegpu/test_hf_export_info.py
- tests/unittest/auto_deploy/singlegpu/models/test_hf.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_update_kv_cache.py
- tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py
- tests/unittest/auto_deploy/singlegpu/models/test_hunyuan_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma3n_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_exaone_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_granite_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_cohere_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma4_modeling.py
Member
Author
|
/bot run --disable-fail-fast |
Collaborator
|
PR_Github #64261 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #64261 [ run ] completed with state
|
brnguyen2
reviewed
Aug 6, 2026
brnguyen2
left a comment
brnguyen2
left a comment
Collaborator
There was a problem hiding this comment.
The mechanism here is worth being explicit about, because it makes this more than a tagging change: jenkins/L0_Test.groovy:1455 runs CPU stages with -m cpu_only and every GPU stage with -m "not cpu_only". Marking a test moves it off GPU hardware rather than adding a CPU run. That's the intent for device-agnostic tests, but it means a module-level pytestmark = pytest.mark.cpu_only over a file that still contains GPU-gated tests silently deletes those from CI — the CPU stage skips them via their own skipif(not torch.cuda.is_available()), and the GPU stages never select them. See the inline comment on test_multi_stream_utils.py; that one is an explicitly documented regression guard.
Worth auditing the other ~20 files where you used module-level pytestmark rather than per-test marks, for the same pattern. The rest looks careful — I checked every newly-marked file for collection-time CUDA (module imports, @parametrize arguments evaluated at import) and the only real case was the one you already fixed in test_quant.py; get_device_counts() degrades to skipif marks so test_tp_sharding.py collects fine without GPUs.
The test_create_ad_executor.py refactor into _mock_ad_engine_build / _mock_py_executor_creation is a clear improvement over the repeated five-deep with stacks.
Signed-off-by: Yuan Tong <13075180+tongyuantongyu@users.noreply.github.com>
Signed-off-by: Yuan Tong <13075180+tongyuantongyu@users.noreply.github.com>
tongyuantongyu
force-pushed
the
ytong/host-only-test-ad
branch
from
August 7, 2026 04:17
1ad436e to
8727bf0
Compare
Contributor
|
Note GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer. |
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (1)
tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py (1)
28-32: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winAnnotate the new fixture function.
patch_device_countis a generator fixture without a return annotation. AddGenerator[None, None, None]or an equivalent precise type to satisfy the guideline that every function must be annotated.Proposed fix
-def patch_device_count(): +def patch_device_count() -> Generator[None, None, 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/unittest/auto_deploy/singlegpu/shim/test_llm_config.py` around lines 28 - 32, Add a precise generator return annotation, such as Generator[None, None, None], to the patch_device_count pytest fixture while preserving its existing patch and yield behavior.Source: Coding guidelines
🤖 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.
Nitpick comments:
In `@tests/unittest/auto_deploy/singlegpu/shim/test_llm_config.py`:
- Around line 28-32: Add a precise generator return annotation, such as
Generator[None, None, None], to the patch_device_count pytest fixture while
preserving its existing patch and yield behavior.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 3d97f9c5-d584-4592-9603-707da3e38acb
📒 Files selected for processing (79)
tests/integration/test_lists/test-db/l0_cpu.ymltests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_gpt_oss_modeling.pytests/unittest/auto_deploy/multigpu/custom_ops/test_ad_dist_strategies.pytests/unittest/auto_deploy/multigpu/transformations/library/test_tp_sharding.pytests/unittest/auto_deploy/singlegpu/compile/test_captured_graph.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_runner.pytests/unittest/auto_deploy/singlegpu/compile/test_piecewise_utils.pytests/unittest/auto_deploy/singlegpu/custom_ops/attention/test_torch_attention_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_torch_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_triton_mla_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/moe/test_mxfp4_moe_layout.pytests/unittest/auto_deploy/singlegpu/custom_ops/quantization/test_quant.pytests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_gptq_op.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_resource_handlers.pytests/unittest/auto_deploy/singlegpu/custom_ops/test_update_kv_cache.pytests/unittest/auto_deploy/singlegpu/mlir/test_decompose.pytests/unittest/auto_deploy/singlegpu/mlir/test_dialect.pytests/unittest/auto_deploy/singlegpu/mlir/test_elementwise_fusion_e2e.pytests/unittest/auto_deploy/singlegpu/mlir/test_fx_mlir_roundtrip.pytests/unittest/auto_deploy/singlegpu/mlir/test_string_codegen.pytests/unittest/auto_deploy/singlegpu/mlir/test_subgraph_discovery.pytests/unittest/auto_deploy/singlegpu/models/test_cohere_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_deepseek_v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_eagle.pytests/unittest/auto_deploy/singlegpu/models/test_exaone_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma3n_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_gemma_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_lite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_glm_moe_dsa_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_granite_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_hf.pytests/unittest/auto_deploy/singlegpu/models/test_hunyuan_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_internlm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_kimi_k2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_llama4_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_minimax_m2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_mistral_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_olmo3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_openelm_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_5_moe.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_moe_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_qwen3_next_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_seed_oss_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_skywork_r1v2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_smollm3_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_starcoder2_modeling.pytests/unittest/auto_deploy/singlegpu/models/test_step3p7_modeling.pytests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.pytests/unittest/auto_deploy/singlegpu/shim/test_llm_config.pytests/unittest/auto_deploy/singlegpu/test_graph_canonicalize.pytests/unittest/auto_deploy/singlegpu/test_hf_export_info.pytests/unittest/auto_deploy/singlegpu/test_mistral_small_4_tokenizer_bridge.pytests/unittest/auto_deploy/singlegpu/test_pattern_matcher.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_bmm_moe_hooks.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_silu_mul.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_swiglu.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_gemm_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_kvcache_vswa_metadata.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_mrope_delta_cache.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quant_fusion.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_quantization.pytests/unittest/auto_deploy/singlegpu/transformations/library/test_shared_kv_attention.pytests/unittest/auto_deploy/singlegpu/transformations/test_export.pytests/unittest/auto_deploy/singlegpu/transformations/test_pipeline_cache.pytests/unittest/auto_deploy/singlegpu/utils/test_config.pytests/unittest/auto_deploy/singlegpu/utils/test_create_derived_custom_op.pytests/unittest/auto_deploy/singlegpu/utils/test_dist_config.pytests/unittest/auto_deploy/singlegpu/utils/test_example_configs.pytests/unittest/auto_deploy/singlegpu/utils/test_multi_stream_utils.pytests/unittest/auto_deploy/singlegpu/utils/test_node_utils_sharding.pytests/unittest/auto_deploy/singlegpu/utils/test_quantization_utils.pytests/unittest/auto_deploy/standalone/test_standalone_import.py
🚧 Files skipped from review as they are similar to previous changes (76)
- tests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_triton_mla_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_mistral_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_step3p7_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_deepseek_v2_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_granite_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_lite_modeling.py
- tests/unittest/auto_deploy/singlegpu/utils/test_dist_config.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_gemm_fusion.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_5_moe.py
- tests/unittest/auto_deploy/singlegpu/models/test_cohere_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_llama4_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_kimi_k2_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_starcoder2_modeling.py
- tests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_gpt_oss_modeling.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_string_codegen.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/moe/test_mxfp4_moe_layout.py
- tests/unittest/auto_deploy/singlegpu/models/test_qwen3_next_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_update_kv_cache.py
- tests/unittest/auto_deploy/singlegpu/models/test_openelm_modeling.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_fx_mlir_roundtrip.py
- tests/unittest/auto_deploy/singlegpu/test_hf_export_info.py
- tests/unittest/auto_deploy/multigpu/transformations/library/test_tp_sharding.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_subgraph_discovery.py
- tests/unittest/auto_deploy/singlegpu/compile/test_piecewise_utils.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_kvcache_vswa_metadata.py
- tests/unittest/auto_deploy/singlegpu/models/test_mistral3_modeling.py
- tests/unittest/auto_deploy/singlegpu/utils/test_multi_stream_utils.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_silu_mul.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/mla/test_torch_mla_op.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_gptq_op.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.py
- tests/unittest/auto_deploy/standalone/test_standalone_import.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm_moe_dsa_modeling.py
- tests/unittest/auto_deploy/singlegpu/test_mistral_small_4_tokenizer_bridge.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_dialect.py
- tests/unittest/auto_deploy/singlegpu/transformations/test_pipeline_cache.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma2_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_mrope_delta_cache.py
- tests/unittest/auto_deploy/multigpu/custom_ops/test_ad_dist_strategies.py
- tests/unittest/auto_deploy/singlegpu/models/test_exaone_modeling.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/attention/test_torch_attention_op.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_fuse_swiglu.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma4_modeling.py
- tests/unittest/auto_deploy/singlegpu/utils/test_node_utils_sharding.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_quantization.py
- tests/unittest/auto_deploy/singlegpu/custom_ops/test_resource_handlers.py
- tests/unittest/auto_deploy/singlegpu/utils/test_quantization_utils.py
- tests/unittest/auto_deploy/singlegpu/compile/test_piecewise_runner.py
- tests/unittest/auto_deploy/singlegpu/utils/test_create_derived_custom_op.py
- tests/unittest/auto_deploy/singlegpu/models/test_eagle.py
- tests/unittest/auto_deploy/singlegpu/test_graph_canonicalize.py
- tests/unittest/auto_deploy/singlegpu/models/test_minimax_m2_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_seed_oss_modeling.py
- tests/unittest/auto_deploy/singlegpu/utils/test_config.py
- tests/unittest/auto_deploy/singlegpu/utils/test_example_configs.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_decompose.py
- tests/unittest/auto_deploy/singlegpu/models/test_llama3_modeling.py
- tests/integration/test_lists/test-db/l0_cpu.yml
- tests/unittest/auto_deploy/singlegpu/models/test_internlm3_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_skywork_r1v2_modeling.py
- tests/unittest/auto_deploy/singlegpu/shim/test_create_ad_executor.py
- tests/unittest/auto_deploy/singlegpu/models/test_smollm3_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_decilm_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/test_export.py
- tests/unittest/auto_deploy/singlegpu/models/test_olmo3_modeling.py
- tests/unittest/auto_deploy/singlegpu/compile/test_captured_graph.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_quant_fusion.py
- tests/unittest/auto_deploy/singlegpu/models/test_hunyuan_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_bmm_moe_hooks.py
- tests/unittest/auto_deploy/singlegpu/models/test_gemma3n_modeling.py
- tests/unittest/auto_deploy/singlegpu/transformations/library/test_shared_kv_attention.py
- tests/unittest/auto_deploy/singlegpu/test_pattern_matcher.py
- tests/unittest/auto_deploy/singlegpu/mlir/test_elementwise_fusion_e2e.py
StanleySun639
approved these changes
Aug 7, 2026
brnguyen2
approved these changes
Aug 7, 2026
brnguyen2
left a comment
brnguyen2
left a comment
Collaborator
There was a problem hiding this comment.
Approving — the comments below are optional touch-ups, not blockers.
Revisit — all three prior threads check out against the current code, not just the resolution status:
test_multi_stream_utils.py— module-levelpytestmarkreplaced by a class-level marker onTestDisableMultiStreamonly.TestRecordStreamDuringCudaGraphCapture(line 108) is unmarked, so GPU stages still select it. Fixed correctly.test_hf.py— markers are now per-test;test_hf_load_state_dict_with_device(line 44) is deliberately left unmarked, so itstorch.cuda.is_available()branch still runs on GPU stages. Fixed correctly.test_llm_config.py— theget_device_countpatch moved out of the data fixture into a separatepatch_device_countfixture requested only bytest_config_flow. No leak into later tests. Fixed correctly.
I also re-checked the general hazard I raised last round, and want to record why it's not a problem here: tests/unittest/conftest.py:228 (pytest_ignore_collect) skips importing any test file whose source doesn't literally contain pytest.mark.cpu_only when the mark expression is cpu_only. That's what makes adding whole directories to l0_cpu.yml safe — test_flashinfer_attention_op.py, which evaluates torch.cuda.get_device_capability(0) in a decorator, is never imported on a 0-GPU node. It also means any file with at least one marked test is imported in full on the CPU stage, so module-level imports there must be CPU-safe (test_mxfp4_moe_layout.py imports triton_kernels at module scope) — CI evidence says that holds today.
Every remaining module-level pytestmark = pytest.mark.cpu_only (piecewise_utils, mxfp4_moe_layout, gptq_op, update_kv_cache, the three mlir files, create_ad_executor, llm_config) contains no GPU-gated test, so nothing else is silently dropped.
One non-blocking note inline.
Member
Author
|
/bot run --disable-fail-fast |
Collaborator
|
PR_Github #64934 [ run ] triggered by Bot. Commit: |
govind-ramnarayan
approved these changes
Aug 10, 2026
govind-ramnarayan
left a comment
govind-ramnarayan
left a comment
Collaborator
There was a problem hiding this comment.
LGTM as long as it is green on the new CI stage (I guess if a migrated test requires a GPU it will not be?)
Collaborator
|
PR_Github #64934 [ run ] completed with state |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 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.
Dev Engineer Review
pytest.mark.cpu_only.tests/integration/test_lists/test-db/l0_cpu.yml.CPU-Generic-x86-1andCPU-Generic-arm-1.UNSTABLE. Multi-GPU tests require theci: full pre-merge approvedlabel and follow-up investigation.QA Engineer Review
test_quant.py,test_create_ad_executor.py,test_llm_config.py, andtest_resource_handlers.py._torch/auto_deploy,auto_deploy/multigpu,auto_deploy/singlegpu, andauto_deploy/standalonetotests/integration/test_lists/test-db/l0_cpu.yml.UNSTABLEand another run was requested after failure investigation.Description
Move AutoDeploy test cases don't need GPU to CPU stage.
Test Coverage
The CPU stage.
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-compatibleorapi-breaking. Forapi-breaking, includeBREAKINGin 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.