Skip to content

[None][test] Fix qwen3.5_9b avg_seq_len and restore H100 coverage for three Qwen3.5 perf cases #17422

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
yufeiwu-nv merged 4 commits into from
Aug 10, 2026
Merged

[None][test] Fix qwen3.5_9b avg_seq_len and restore H100 coverage for three Qwen3.5 perf cases #17422

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
994ca23
[None][test] Set avg_seq_len for the qwen3.5_9b 500/2000 perf case
yufeiwu-nv Aug 7, 2026
84d92a8
[None][test] Restore H100 coverage for two Qwen3.5 perf cases
yufeiwu-nv Aug 7, 2026
ff67109
[None][test] Restore H100 coverage for qwen3.5_122b_a10b 128,128
yufeiwu-nv Aug 7, 2026
cc686e2
Merge branch 'main' into fp4
yufeiwu-nv Aug 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
11 changes: 11 additions & 0 deletions tests/integration/defs/perf/pytorch_model_config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -324,6 +324,17 @@ def get_model_yaml_config(model_label: str,
},
}
},
# Qwen3.5-9B hybrid GDN: V2 KV/SSM pool split needs the real seq len.
{
'patterns': [
'qwen3.5_9b-bench-pytorch-bfloat16-maxbs:512-maxnt:2048-input_output_len:500,2000',
Comment thread
coderabbitai[bot] marked this conversation as resolved.
],
'config': {
'kv_cache_config': {
'avg_seq_len': 2500,
},
}
},
# MiniMax-M3 MXFP8 block-sparse MoE: sparse backend, no KV reuse, trust_remote_code, capped max_seq_len to avoid the 1M-default CUDA-graph OOM.
{
'patterns': ['minimax_m3_mxfp8'],
Expand Down
7 changes: 3 additions & 4 deletions tests/integration/test_lists/qa/llm_perf_core.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,6 +106,8 @@ llm_perf_core:
- perf/test_perf.py::test_perf[qwen3.5_35b_a3b_fp8-bench-pytorch-float8-maxbs:1-input_output_len:8000,1000-reqs:10-con:1] #min_latency
- perf/test_perf.py::test_perf[qwen3.5_35b_a3b_fp8-bench-pytorch-float8-maxbs:512-input_output_len:8000,1000-con:256] #max_throughput
#qwen3.5_122b_a10b (MoE BF16 234G, 4-GPU)
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:128,128-ep:4-tp:4-gpus:4]
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:500,2000-ep:4-tp:4-gpus:4]
Comment thread
yufeiwu-nv marked this conversation as resolved.
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:2000,500-ep:4-tp:4-gpus:4]
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:8000,1000-ep:4-tp:4-gpus:4]
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:1000,2000-ep:4-tp:4-gpus:4]
Expand Down Expand Up @@ -279,7 +281,6 @@ llm_perf_core:
- perf/test_perf.py::test_perf[minimax_m3_fp4-bench-pytorch-float4-maxbs:1-input_output_len:1024,1024-reqs:10-con:1-ep:8-tp:8-gpus:8] #min_latency
- perf/test_perf.py::test_perf[minimax_m3_fp4-bench-pytorch-float4-maxbs:256-maxnt:8192-input_output_len:1024,1024-con:256-ep:8-tp:8-gpus:8] TIMEOUT(120) #max_throughput
- perf/test_perf.py::test_perf[deepseek_r1_0528_fp4-bench-pytorch-float4-maxbs:384-maxnt:1536-input_output_len:1000,2000-reqs:10000-con:3072-ep:8-tp:8-gpus:8] TIMEOUT(120) #max throughput test
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-input_output_len:1000,2000-ep:8-tp:8-gpus:8]


# 8: B300 test cases
Expand Down Expand Up @@ -311,6 +312,7 @@ llm_perf_core:
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-input_output_len:500,2000-ep:8-tp:8-gpus:8]
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-input_output_len:2000,500-ep:8-tp:8-gpus:8]
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-input_output_len:8000,1000-ep:8-tp:8-gpus:8]
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-input_output_len:1000,2000-ep:8-tp:8-gpus:8]
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-maxbs:1-input_output_len:8000,1000-reqs:10-con:1-ep:8-tp:8-gpus:8] #min_latency
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp8-bench-pytorch-float8-maxbs:512-input_output_len:8000,1000-con:512-ep:8-tp:8-gpus:8] #max_throughput

Expand All @@ -326,9 +328,6 @@ llm_perf_core:
gpu_memory:
gt: 90000
tests:
#qwen3.5_122b_a10b (MoE BF16 234G, 4-GPU): OOM on 80G GPUs, needs >90G
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:128,128-ep:4-tp:4-gpus:4]
- perf/test_perf.py::test_perf[qwen3.5_122b_a10b-bench-pytorch-bfloat16-input_output_len:500,2000-ep:4-tp:4-gpus:4]
#qwen3.5_397b_a17b_fp4 (MoE FP4 234G, 8-GPU ep=8)
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp4-bench-pytorch-float4-input_output_len:128,128-ep:8-tp:8-gpus:8]
- perf/test_perf.py::test_perf[qwen3.5_397b_a17b_fp4-bench-pytorch-float4-input_output_len:500,2000-ep:8-tp:8-gpus:8]
Expand Down
Loading