Skip to content

[V1][Minor] Print KV cache size in token counts #13596

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 1 commit into from
Feb 20, 2025
Merged

[V1][Minor] Print KV cache size in token counts #13596

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
10 changes: 6 additions & 4 deletions vllm/v1/core/kv_cache_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -519,11 +519,13 @@ def _get_kv_cache_config_uniform_type(vllm_config: VllmConfig,
"num_gpu_blocks_override=%d", num_blocks, num_gpu_blocks_override)
num_blocks = num_gpu_blocks_override

logger.info("# GPU blocks: %d", num_blocks)
max_concurrency = (num_blocks * vllm_config.cache_config.block_size /
vllm_config.model_config.max_model_len)
num_tokens = num_blocks * vllm_config.cache_config.block_size
Copy link
Collaborator

@comaniac comaniac Feb 20, 2025

Choose a reason for hiding this comment

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

This may not be accurate with hybrid of cache.

Copy link
Collaborator Author

@WoosukKwon WoosukKwon Feb 20, 2025

Choose a reason for hiding this comment

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

Ah yes, but I think num_blocks is already inaccurate? We should reconsider this for hybrid memory models.

num_tokens_str = f"{num_tokens:,}"
logger.info("GPU KV cache size: %s tokens", num_tokens_str)
max_model_len_str = f"{vllm_config.model_config.max_model_len:,}"
max_concurrency = num_tokens / vllm_config.model_config.max_model_len
logger.info("Maximum concurrency for %s tokens per request: %.2fx",
Comment on lines +524 to 527
Copy link
Member

@mgoin mgoin Feb 20, 2025

Choose a reason for hiding this comment

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

We could consider putting these lines together in one log

Copy link
Collaborator Author

@WoosukKwon WoosukKwon Feb 20, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

I'd personally prefer them on separate lines because the first line is more important and needs to be checked every time, while the second line is less critical.

vllm_config.model_config.max_model_len, max_concurrency)
max_model_len_str, max_concurrency)

per_layer_size = page_size * num_blocks

Expand Down