coffee/DeepInfer
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepInfer.git synced 2026-10-28 14:57:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

bench_neuron: one untimed request a cell (context-bucket graphs compile lazily)

Browse Source 
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
whn09 2026-10-23 21:16:45 +08:00
parent 50483a06df
commit 32e08d78cd
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
examples/neuron/bench_neuron.py
View File

@ -122,6 +122,8 @@ def main(dp_size: int, dp_rank: int) -> None:
_run_one(llm, _random_prompt(lens[0], 10_000 + w + 7919 * dp_rank, vocab) if busy else None, params)
results = []
for length in lens:
# One untimed request a cell: a cell's first request pays for its own context bucket's lazy compile.
_run_one(llm, _random_prompt(length, 20_000 + length + 7919 * dp_rank, vocab) if busy else None, params)
cell = []
for r in range(args.repeats):
got = _run_one(llm, _random_prompt(length, 1000 * length + r + 7919 * dp_rank, vocab) if busy else None, params)