Skip to content

[https://nvbugs/6221055][fix] added four GB300 rows to agg_unit_mem_df.csv #14785

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
QiJune merged 6 commits into from
Jul 10, 2026
Merged

[https://nvbugs/6221055][fix] added four GB300 rows to agg_unit_mem_df.csv #14785

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
83cb292
[nvbugs/6221055][fix] Add GB300 rows to agg_unit_mem_df.csv
tensorrt-cicd May 30, 2026
1e720cb
Merge branch 'main' into repair-bot-bug6221055
QiJune Jul 10, 2026
f8b9ca9
unwaive thop/parallel
QiJune Jul 10, 2026
c89e103
[CI] Print GPU name for unittest parallel config
QiJune Jul 10, 2026
8636cf7
[CI] Add GB300 unittest parallel config
QiJune Jul 10, 2026
ebf9a26
[CI] Remove unused B300 unittest parallel config
QiJune Jul 10, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions tests/integration/defs/agg_unit_mem_df.csv
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,3 +138,7 @@ unittest/_torch/misc,NVIDIA H100,4,
unittest/_torch/thop/parallel,NVIDIA H100,16,
unittest/_torch/thop/parallel_hw_agnostic,NVIDIA H100,16,
unittest/_torch/flashinfer/test_trtllm_flashinfer_symbol_collision.py,NVIDIA H100 PCIe,1,
unittest/_torch/attention,NVIDIA GB300,4,
unittest/_torch/thop/parallel,NVIDIA GB300,16,
unittest/_torch/thop/serial,NVIDIA GB300,4,
unittest/_torch/executor,NVIDIA GB300,4,
11 changes: 9 additions & 2 deletions tests/integration/defs/test_unittests.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,13 +84,19 @@ def test_unittests_v2(llm_root, llm_venv, case: str, output_dir, request):

# This dataframe is not manually edited. Infra team will regularly generate this dataframe based on test execution results.
# If you need to override this policy, please use postprocess code as below.
agg_unit_mem_df = pd.read_csv(
f'{test_root}/integration/defs/agg_unit_mem_df.csv')
agg_unit_mem_path = f'{test_root}/integration/defs/agg_unit_mem_df.csv'
print(f'Loading unittest parallel config from: {agg_unit_mem_path}')
agg_unit_mem_df = pd.read_csv(agg_unit_mem_path)
gpu_handle = pynvml.nvmlDeviceGetHandleByIndex(0)
gpu_name = pynvml.nvmlDeviceGetName(gpu_handle)
if isinstance(gpu_name, bytes):
gpu_name = gpu_name.decode()
print(f'GPU name from NVML (index 0): {gpu_name!r}')
print(f'GPU names available in parallel config: '
f'{sorted(agg_unit_mem_df["gpu"].dropna().unique().tolist())}')
agg_unit_mem_df = agg_unit_mem_df[agg_unit_mem_df['gpu'] == gpu_name]
print(f'Matched parallel config rows for GPU {gpu_name!r}: '
f'{len(agg_unit_mem_df)}')
print(agg_unit_mem_df)

parallel_dict = {}
Expand All @@ -101,6 +107,7 @@ def test_unittests_v2(llm_root, llm_venv, case: str, output_dir, request):
print(parallel_dict)

cur_key = (gpu_name, case)
print(f'Parallel config lookup key: {cur_key!r}')
if cur_key in parallel_dict:
num_workers = parallel_dict[cur_key]
num_workers = min(num_workers, 8)
Expand Down
1 change: 0 additions & 1 deletion tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,7 +260,6 @@ full:GB300/accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_fp8_moe_d
full:GB300/accuracy/test_llm_api_pytorch_multimodal.py::TestExaone4_5_33B::test_auto_dtype[forced_chunked_prefill] SKIP (https://nvbugs/6422318)
full:GB300/accuracy/test_llm_api_pytorch_multimodal.py::TestExaone4_5_33B::test_auto_dtype[full_budget] SKIP (https://nvbugs/6422318)
full:GB300/test_e2e.py::test_qwen_e2e_cpprunner_large_new_tokens[DeepSeek-R1-Distill-Qwen-1.5B-DeepSeek-R1-Distill-Qwen-1.5B] SKIP (https://nvbugs/6414760)
full:GB300/unittest/_torch/thop/parallel SKIP (https://nvbugs/6422443)
full:GH200/examples/test_nemotron.py::test_llm_nemotron_3_8b_1gpu[bfloat16-fp8] SKIP (arm is not supported)
full:GH200/examples/test_qwen2audio.py::test_llm_qwen2audio_single_gpu[qwen2_audio_7b_instruct] SKIP (arm is not supported)
full:H100/accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype[mtp_nextn=2-overlap_scheduler=False] SKIP (https://nvbugs/6313072)
Expand Down
Loading