Skip to content

[None][fix] Cap TLLM_BENCHMARK_REQ_QUEUES_SIZE to avoid fill-loop hang #13065

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
reasonsolo merged 2 commits into from
Apr 21, 2026
Merged

[None][fix] Cap TLLM_BENCHMARK_REQ_QUEUES_SIZE to avoid fill-loop hang #13065

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5dcd828
[None][fix] Cap TLLM_BENCHMARK_REQ_QUEUES_SIZE to avoid fill-loop hang
reasonsolo Apr 15, 2026
9cce9cf
[None][fix] Fix type bug and formatting in queue size cap logic
reasonsolo Apr 15, 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
18 changes: 17 additions & 1 deletion examples/disaggregated/slurm/benchmark/submit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -228,9 +228,25 @@ def build_worker_environment(worker_config, env_config, role, benchmark_mode,
'TRTLLM_DISABLE_KV_CACHE_TRANSFER_OVERLAP',
'TRTLLM_DISABLE_KV_CACHE_TRANSFER_OVERLAP=1')
if role == "GEN":
gen_config = worker_config.get('gen', {})
concurrency_int = int(concurrency)
max_batch_size = int(
gen_config.get('max_batch_size', concurrency_int))
enable_attention_dp = gen_config.get('enable_attention_dp', False)
tp_size = int(gen_config.get('tensor_parallel_size', 1))
max_capacity = ((max_batch_size * tp_size)
if enable_attention_dp else max_batch_size)
queue_size = min(max_capacity, concurrency_int)
if queue_size < concurrency_int:
print(f"[WARNING] TLLM_BENCHMARK_REQ_QUEUES_SIZE capped to "
f"{queue_size} (max_batch_size={max_batch_size} x "
f"tp_size={tp_size} with "
f"attention_dp={enable_attention_dp}) "
f"which is less than concurrency={concurrency}. "
f"Fill loop would hang if set to {concurrency}.")
upsert_env_config(env_config, 'gen_worker_env_var',
'TLLM_BENCHMARK_REQ_QUEUES_SIZE',
f'TLLM_BENCHMARK_REQ_QUEUES_SIZE={concurrency}')
f'TLLM_BENCHMARK_REQ_QUEUES_SIZE={queue_size}')

# 2. Add profiling env vars to env_config (conditional)
if nsys_on:
Expand Down
Loading