Skip to content

V1 - dont look for bucket we know don't exists #1606

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
adobrzyn merged 2 commits into from
Jul 16, 2025
Merged

V1 - dont look for bucket we know don't exists #1606

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6e3d05d
fix
adobrzyn Jul 16, 2025
97a55ca
Change Nones to exception
adobrzyn Jul 16, 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
13 changes: 11 additions & 2 deletions vllm/v1/worker/hpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -129,6 +129,10 @@ class DecodeInputData:
logits_indices: Optional[torch.Tensor] = None


class BucketingFailedException(Exception):
pass


def bool_helper(value):
value = value.lower()
return value in ("y", "yes", "t", "true", "on", "1")
Expand Down Expand Up @@ -966,6 +970,8 @@ def _bucketize_merged_prompt(self, seq_lens, num_blocks):

def _bucketize_2d_prompt(self, seq_lens, num_blocks):
bs = len(seq_lens)
if bs > self.max_prefill_batch_size:
raise BucketingFailedException
seq = max(seq_lens)
num_blocks = max(num_blocks) if len(num_blocks) > 0 else 0
bs, seq, num_blocks = self.bucketing_manager.find_prompt_bucket(
Expand All @@ -981,8 +987,11 @@ def _get_prompt_bucketing_fn(self):
def _can_merge_prefill_contents(self, lhs, rhs):
combined_num_tokens = lhs.get_num_tokens() + rhs.get_num_tokens()
bucketing_fn = self._get_prompt_bucketing_fn()
target_bs, target_seq, target_blocks = bucketing_fn(
combined_num_tokens, [])
try:
target_bs, target_seq, target_blocks = bucketing_fn(
combined_num_tokens, [])
except BucketingFailedException:
return False
return target_bs <= self.max_prefill_batch_size and\
target_bs * target_seq <= self.max_num_tokens

Expand Down