Skip to content
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

Jump to bottom

Disable donated buffer when benchmarking layer_norm with backwards #88

Closed
wants to merge 3 commits into from
Closed

Disable donated buffer when benchmarking layer_norm with backwards #88

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
afa8e31
Disable donated buffer when benchmarking
xuzhao9 Dec 2, 2024
3f75a34
Lint
xuzhao9 Dec 2, 2024
f0e2a04
Fix unit test
xuzhao9 Dec 2, 2024
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
8 changes: 8 additions & 0 deletions tritonbench/operators/layer_norm/operator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,14 @@ def torch_layer_norm(self, *args):

@register_benchmark()
def torch_compile_layer_norm(self, *args):
# We need to run backward multiple times for proper benchmarking
# so donated buffer have to be disabled
if self.mode == Mode.BWD or self.mode == Mode.FWD_BWD:
import torch._functorch.config

torch._functorch.config.donated_buffer = False
import torch

@torch.compile
def inner(*args):
return F.layer_norm(*args)
Expand Down
Loading