Skip to content

fix kernel block size, port of #1439 #1453

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
kamil-kaczor merged 2 commits into from
May 19, 2026
Merged

fix kernel block size, port of #1439 #1453

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b00f29b
fix kernel block size, #1439
iboiko-habana May 18, 2026
38c7a9e
Merge branch 'main' into pr1439_port
iboiko-habana May 18, 2026
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
9 changes: 5 additions & 4 deletions vllm_gaudi/v1/attention/backends/hpu_attn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,10 +35,11 @@ def get_metadata_cls() -> type["AttentionMetadata"]:

@staticmethod
def get_supported_kernel_block_sizes() -> list[Union[int, MultipleOf]]:
# 128 is the standard HPU kernel block size; 528 is required for
# Granite 4.0-H (granitemoehybrid) without prefix caching (16-token
# FA alignment), 768 with prefix caching (chunk-aligned).
return [128, 528, 768]
# 16 is supported for testing/smaller models; 128 is the standard HPU
# kernel block size; 528 is required for Granite 4.0-H
# (granitemoehybrid) without prefix caching (16-token FA alignment),
# 768 with prefix caching (chunk-aligned).
return [16, 128, 528, 768]

@classmethod
def get_preferred_block_size(cls, default_block_size: int) -> int:
Expand Down
Loading