Skip to content

Fix matmul_reduce_scatter wrapper passing unsupported bias argument #530

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
mawad-amd wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix matmul_reduce_scatter wrapper passing unsupported bias argument #530

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4da3077
Fix matmul_reduce_scatter wrapper passing unsupported bias argument
mawad-amd May 2, 2026
6798ff6
Remove unsupported bias parameter from matmul_all_reduce and matmul_r…
mawad-amd May 2, 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
12 changes: 5 additions & 7 deletions iris/ops/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,17 +65,16 @@ def __init__(self, shmem):
"""
self._shmem = shmem

def matmul_all_reduce(self, output_tensor, A, B, bias=None, async_op=False, config=None, workspace=None):
def matmul_all_reduce(self, output_tensor, A, B, async_op=False, config=None, workspace=None):
"""
Fused matrix multiplication and all-reduce.

Computes: output = all_reduce(A @ B + bias)
Computes: output = all_reduce(A @ B)

Args:
output_tensor: Output tensor (M, N)
A: Input matrix A (M, K)
B: Input matrix B (K, N)
bias: Optional bias vector (M,) or (N,)
async_op: If False, performs barrier at end
config: Optional FusedConfig for tuning
workspace: Optional pre-allocated workspace
Expand Down Expand Up @@ -141,17 +140,16 @@ def matmul_all_gather(self, output_tensor, A, B, bias=None, async_op=False, conf
"""
return matmul_all_gather(self._shmem, output_tensor, A, B, bias, async_op, config, workspace)

def matmul_reduce_scatter(self, output_tensor, A, B, bias=None, async_op=False, config=None, workspace=None):
def matmul_reduce_scatter(self, output_tensor, A, B, async_op=False, config=None, workspace=None):
"""
Fused matrix multiplication and reduce-scatter.

Computes: output = reduce_scatter(A @ B + bias) along N dimension
Computes: output = reduce_scatter(A @ B) along N dimension

Args:
output_tensor: Output tensor (M, N_local) where N_local = N / world_size
A: Input matrix A (M, K)
B: Input matrix B (K, N)
bias: Optional bias vector (M,) or (N,)
async_op: If False, performs barrier at end
config: Optional FusedConfig for tuning
workspace: Optional pre-allocated workspace
Expand All @@ -164,7 +162,7 @@ def matmul_reduce_scatter(self, output_tensor, A, B, bias=None, async_op=False,
>>> output = shmem.zeros((M, N_local), dtype=torch.float16)
>>> shmem.ops.matmul_reduce_scatter(output, A, B)
"""
return matmul_reduce_scatter(self._shmem, output_tensor, A, B, bias, async_op, config, workspace)
return matmul_reduce_scatter(self._shmem, output_tensor, A, B, async_op, config, workspace)

@mawad-amd mawad-amd May 2, 2026

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. Fixed in 6798ff6 — removed bias from both matmul_all_reduce and matmul_reduce_scatter wrapper signatures. The underlying functions don't support bias; only all_gather_matmul and matmul_all_gather do.



# Export public API
Expand Down
Loading