Skip to content

<Scheduler> Eliminate the constraint that prevents chunked prefill from working with padding-aware scheduling #2145

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

Closed
wants to merge 1 commit into from
Closed

<Scheduler> Eliminate the constraint that prevents chunked prefill from working with padding-aware scheduling #2145

Show file tree
Hide file tree
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
3 changes: 0 additions & 3 deletions vllm/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1686,9 +1686,6 @@ def _verify_args(self) -> None:
and not self.use_padding_aware_scheduling:
raise ValueError("max_num_prefill_seqs can be only "
"used with padding-aware-scheduling. ")
if self.use_padding_aware_scheduling and self.chunked_prefill_enabled:
raise ValueError("Padding-aware scheduling currently "
"does not work with chunked prefill ")

@property
def is_multi_step(self) -> bool:
Expand Down
15 changes: 11 additions & 4 deletions vllm/core/scheduler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1399,10 +1399,17 @@ def _schedule_chunked_prefill(self) -> SchedulerOutputs:
inter token latency because decodes requests don't need to be blocked
by prefill requests.
"""
budget = SchedulingBudget(
token_budget=self.scheduler_config.max_num_batched_tokens,
max_num_seqs=self.scheduler_config.max_num_seqs,
)
if self.scheduler_config.use_padding_aware_scheduling:
budget = PaddingAwareSchedulingBudget(
token_budget=self.scheduler_config.max_num_batched_tokens,
max_num_seqs=self.scheduler_config.max_num_seqs,
max_num_prefill_seqs=self.scheduler_config.max_num_prefill_seqs
)
else:
budget = SchedulingBudget(
token_budget=self.scheduler_config.max_num_batched_tokens,
max_num_seqs=self.scheduler_config.max_num_seqs,
)
curr_loras: Set[int] = set()

prefills = SchedulerPrefillOutputs.create_empty()
Expand Down
Loading