onnxruntime
[JS/WebGPU] Optimize MatMulNBits
#19852
Merged

[JS/WebGPU] Optimize MatMulNBits #19852

satyajandhyala
satyajandhyala Temporarily remove uniforms to debug.
d114d9a2
satyajandhyala Vectorize MatMulNBits.
f1860048
satyajandhyala Use uppercase letters for M, N and K.
a825fbb6
satyajandhyala Restore uniforms.
daa37308
satyajandhyala Added a testcase with 8x8 output.
171362b0
satyajandhyala satyajandhyala added ep:WebGPU
satyajandhyala Added output values vectorization.
2361183d
satyajandhyala Revert "Use uppercase letters for M, N and K."
f8131e77
satyajandhyala Formating changes.
602f07f6
satyajandhyala Formatting
e65f0502
satyajandhyala typo
143faa79
satyajandhyala Use mat4x2 and mat2x4 instead of array when aComponent is 2 or 4.
e99fd21d
satyajandhyala satyajandhyala marked this pull request as ready for review 2 years ago
satyajandhyala Fix lint errors.
357a3a76
satyajandhyala satyajandhyala changed the title [WIP][JS/WebGPU] Optimize MatMulNBits [JS/WebGPU] Optimize MatMulNBits 2 years ago
guschmue
guschmue approved these changes on 2024-03-13
satyajandhyala satyajandhyala merged ed250b88 into main 2 years ago
satyajandhyala satyajandhyala deleted the sajandhy/webgpu_optimize_matmulnbits branch 2 years ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone