Fuse mat mul with activation following it #566
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Please run |
There was a problem hiding this comment.
@boris-drazic please look how fusion is handled here
https://github.com/tenstorrent/pytorch2.0_ttnn/blob/main/torch_ttnn/patterns/linear.py
boris-drazic
force-pushed
the
bdrazic/matmul_fuse
branch
from
7f09044 to
b1fc4ac
Compare
boris-drazic
force-pushed
the
bdrazic/matmul_fuse
branch
from
4292a08 to
ff08fd7
Compare
boris-drazic
force-pushed
the
bdrazic/matmul_fuse
branch
from
9572220 to
a763f32
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Problem description
torch.ops.aten.mm.defaultandtorch.ops.aten.bmm.defaultare lowered tottnn.mm, which has option to execute an activation function together with the multiplication. The three supported functions aretorch.nn.functional.relu,torch.nn.functional.gelu, andtorch.nn.functional.silu. Executing activation together with multiplication is faster, so mat mul and activation function should be fused.What's changed
MMFusionPassadded to perform the fusion of mat mul and activation function that follows it.Added
tests/optimization/test_mm_fusion.pyto test fusion.