Skip to content

fix mxfp8 matmul benchmark #3221

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
merged 1 commit into from
Oct 21, 2025
Merged

fix mxfp8 matmul benchmark #3221

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
7 changes: 7 additions & 0 deletions benchmarks/float8/bench_matmul.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

from torchao.ops import mx_fp4_bf16
from torchao.prototype.mx_formats.mx_tensor import to_mx
from torchao.prototype.mx_formats.utils import to_blocked
from torchao.testing.training.roofline_utils import get_specs
from torchao.utils import is_MI300

Expand Down Expand Up @@ -125,10 +126,16 @@ def run(
elif recipe in ("mxfp8_cublas", "mxfp4_cutlass"):
scale_a = torch.ones(M, K // 32, device=device, dtype=torch.float8_e8m0fnu)
scale_b = torch.ones(N, K // 32, device=device, dtype=torch.float8_e8m0fnu)
# pad if needed
scale_a = to_blocked(scale_a)
scale_b = to_blocked(scale_b)
elif recipe == "nvfp4":
# Use the blockwise scales from nvfp4_quantize
scale_a = A_scales.view(torch.float8_e4m3fn)
scale_b = B_scales.view(torch.float8_e4m3fn)
# pad if needed
scale_a = to_blocked(scale_a)
scale_b = to_blocked(scale_b)
else:
assert False, f"unknown recipe {recipe}"

Expand Down
Loading