Skip to content

Fix a potential buffer size misaligning issue in TMA description of partition attention #3

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

Open
yuxiaoguo wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix a potential buffer size misaligning issue in TMA description of partition attention #3

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: 5 additions & 1 deletion megakernels/demos/latency/scheduler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,10 @@ def make_buffer(shape, buffer_dtype=dtype):

stacked_params = model.stacked_params

max_attn_partitions = get_sm_count(device)
def align_sm_to_partition_cacheline(x: int) -> int:
return int(math.ceil(x / 16) * 16)

max_attn_partitions = align_sm_to_partition_cacheline(get_sm_count(device))

barriers = torch.zeros(
[
Expand All @@ -60,6 +63,7 @@ def make_buffer(shape, buffer_dtype=dtype):
device=device,
)


return Globals(
# model params
qkv_proj_weights=stacked_params.qkv_proj,
Expand Down