Skip to content

[V1] Simplify GPUModelRunner._update_states check #13265

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 14, 2025
Merged

[V1] Simplify GPUModelRunner._update_states check #13265

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
6 changes: 4 additions & 2 deletions vllm/v1/worker/gpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -347,6 +347,8 @@ def _update_states(self, scheduler_output: "SchedulerOutput") -> bool:
self.input_batch.block_table.append_row(req_index, start_index,
req_data.new_block_ids)

batch_changed = len(removed_req_indices) > 0 or len(req_ids_to_add) > 0

# Add the new or resumed requests to the persistent batch.
# The smaller empty indices are filled first.
removed_req_indices = sorted(removed_req_indices, reverse=True)
Expand All @@ -363,8 +365,8 @@ def _update_states(self, scheduler_output: "SchedulerOutput") -> bool:
# Condense the batched states if there are empty indices.
if removed_req_indices:
self.input_batch.condense(removed_req_indices)
return (len(unscheduled_req_ids) > 0 or len(req_ids_to_add) > 0
or len(scheduler_output.finished_req_ids) > 0)

return batch_changed

def _prepare_inputs(self, scheduler_output: "SchedulerOutput"):
total_num_scheduled_tokens = scheduler_output.total_num_scheduled_tokens
Expand Down