Skip to content

feat: mx9: add fused MX9 quantization kernel with pack/unpack #33

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
Jiarwang77 wants to merge 7 commits into
base: main
Choose a base branch
from
Open

feat: mx9: add fused MX9 quantization kernel with pack/unpack #33

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
2de0111
feat: mx9: add fused MX9 quantization kernel with pack/unpack
Jiarwang77 Jun 29, 2026
95b1114
feat: mx9: wire packed kernel into patcher and fix convert_from_mx9
Jiarwang77 Jun 30, 2026
d3ce92b
feat: mx6: add packed quantization kernel
Jiarwang77 Jul 14, 2026
db210c0
perf: mx9: consume strided quantization inputs
Jiarwang77 Jul 14, 2026
c5ce915
feat: mx6: add convert_to_mx6 / convert_from_mx6 host wrappers
Jiarwang77 Jul 17, 2026
abf9559
refactor: mx: pack mx9 into a single uint8 tensor and dispatch mx6/mx…
Jiarwang77 Jul 17, 2026
087a876
refactor: mx: centralize shared MX6/MX9 helpers in _mx_common
Jiarwang77 Jul 23, 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
9 changes: 9 additions & 0 deletions alto/kernels/mx/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
# Copyright (c) 2026 Advanced Micro Devices, Inc.
#
# SPDX-License-Identifier: MIT
"""MX real (fused Triton) kernels.

@hann-wang hann-wang Jun 30, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

This is misleading. We don't have real mx gemm now.


The fake-quant reference / emulation lives in
``alto.modifiers.quantization.mx``; this package holds the fused GPU kernel
implementation of the same MX QDQ.
"""
Loading