Skip to content

Stream-k mxfp4 kernel #74

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
xiaohuguo2023 wants to merge 19 commits into
base: main
Choose a base branch
from
Open

Stream-k mxfp4 kernel #74

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
884e881
the first working version of async_copy support, adding proper estima…
xiaohuguo2023 Feb 27, 2026
6f9e1e5
pass in num_stages for accurate LDS calculation, and implement fast m…
xiaohuguo2023 Feb 27, 2026
23c3875
add LDS tests
xiaohuguo2023 Mar 1, 2026
11ffa5c
add lds tests into ci workflow
xiaohuguo2023 Mar 1, 2026
cf7fcc6
update origamini commit
xiaohuguo2023 Mar 1, 2026
35b141e
add blank line
xiaohuguo2023 Mar 1, 2026
7e9a590
small tile size shouldn#t have LDS issues at all, just raise error
xiaohuguo2023 Mar 1, 2026
1a8c124
fix origami dtype assignment bug
xiaohuguo2023 Mar 12, 2026
8d3483c
remove the redundant lds test
xiaohuguo2023 Mar 12, 2026
bb6e44b
Added three replacement tests that can actually fail
xiaohuguo2023 Mar 12, 2026
2c2615f
make python 3.8 compatible
xiaohuguo2023 Mar 13, 2026
77974be
let's revert back to old api to fix issue in addmm tests
xiaohuguo2023 Mar 13, 2026
2e05e15
adds a torch.compiler.is_compiling() guard so that: During torch.comp…
xiaohuguo2023 Mar 13, 2026
1b0882a
the first version of working streamk mxfp4
xiaohuguo2023 Mar 13, 2026
f87eea3
registers a custom pytest marker called performance
xiaohuguo2023 Mar 13, 2026
0e9cb39
the same fix for non compatible api
xiaohuguo2023 Mar 13, 2026
9661e29
seperate persistent and stream-k fp4 matmul wrapper
xiaohuguo2023 Mar 13, 2026
93bf83b
forgot commit streamk fp4 kernel
xiaohuguo2023 Mar 13, 2026
78bf25c
Merge branch 'main' into streamk_mxfp4
xiaohuguo2023 Mar 15, 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
3 changes: 2 additions & 1 deletion include/tritonblas/kernels/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,9 @@

# FP4 kernel
from .fp4_matmul import fp4_matmul
from .fp4_streamk_gemm import fp4_streamk_matmul

# Export stages submodule
from . import stages

__all__ = ['persistent_matmul', 'streamk_matmul', 'fp4_matmul', 'stages']
__all__ = ['persistent_matmul', 'streamk_matmul', 'fp4_matmul', 'fp4_streamk_matmul', 'stages']
Loading
Loading