vllm
[ROCm][Perf] Fuse the MXFP4 activation quant into the GDN gated RMSNorm
#59553
Open
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
4
Changes
View On
GitHub
[ROCm][Perf] Fuse the MXFP4 activation quant into the GDN gated RMSNorm
#59553
mjkvaak-amd
wants to merge 4 commits into
vllm-project:main
from
mjkvaak-amd:rocm/gdn-rms-gated-mxfp4-gemm
[ROCm][Bugfix] Probe the preshuffled FP4 GEMM config by logical K
c92b9b70
[ROCm][Perf] Fuse the MXFP4 activation quant into the GDN gated RMSNorm
4e3168a7
[ROCm][Compile] Match gated-norm MXFP4 fusion through flatten(-2)
d601c302
mergify
added
quantization
mergify
added
rocm
mergify
added
torch.compile
[ROCm][Compile] Import the gated-norm MXFP4 quant directly
758e1594
Login to write a write a comment.
Login via GitHub
Reviewers
No reviews
Assignees
No one assigned
Labels
rocm
torch.compile
quantization
Milestone
No milestone
Login to write a write a comment.
Login via GitHub