Skip to content

[Log] Optimize kv cache memory log from Bytes to GiB #25204

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
merged 3 commits into from
Sep 23, 2025
Merged

[Log] Optimize kv cache memory log from Bytes to GiB #25204

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1d07a7f
optimize from bytes to GiB
yewentao256 Sep 18, 2025
d8f2687
update
yewentao256 Sep 18, 2025
fa3c783
debug info
yewentao256 Sep 23, 2025
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
12 changes: 7 additions & 5 deletions vllm/v1/worker/gpu_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -383,13 +383,15 @@ def compile_or_warm_up_model(self) -> None:
f"for non-torch memory, and {GiB(cuda_graph_memory_bytes)} "
f"GiB for CUDAGraph memory. Replace gpu_memory_utilization "
f"config with `--kv-cache-memory="
f"{kv_cache_memory_bytes_to_requested_limit}` to fit into "
f"requested memory, or `--kv-cache-memory="
f"{kv_cache_memory_bytes_to_gpu_limit}` to fully "
f"{kv_cache_memory_bytes_to_requested_limit}` "
f"({GiB(kv_cache_memory_bytes_to_requested_limit)} GiB) to fit "
f"into requested memory, or `--kv-cache-memory="
f"{kv_cache_memory_bytes_to_gpu_limit}` "
f"({GiB(kv_cache_memory_bytes_to_gpu_limit)} GiB) to fully "
Comment on lines +388 to +390
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think using GiB is valid for the cli argument

Copy link
Member Author

@yewentao256 yewentao256 Sep 19, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mgoin Thanks! I have updated the code, now it would be something like --kv-cache-memory=8511484 (xxGiB)

f"utilize gpu memory. Current kv cache memory in use is "
f"{int(self.available_kv_cache_memory_bytes)} bytes.")
f"{GiB(self.available_kv_cache_memory_bytes)} GiB.")

logger.info(msg)
logger.debug(msg)

# Warm up sampler and preallocate memory buffer for logits and other
# sampling related tensors of max possible shape to avoid memory
Expand Down