Title: CUDA: few-row MMQ for MUL_MAT_ID #32065
CUDA: few-row MMQ for MUL_MAT_ID
e81135de
CUDA: few-row MMQ for MUL_MAT_ID from 2 tokens on DGX Spark
501d145f
CUDA: few-row MUL_MAT_ID: one launcher, tighter pair count, more tests
15dc909c
CUDA: few-row MUL_MAT_ID: dispatch order MMVQ, few-row, MMQ and revieā¦
501669ea
anujj
force pushed
from
a8cbbe52
to
501669ea
5 days ago
anujj
marked this pull request as ready for review 5 days ago
anujj
requested a review
5 days ago
am17an
approved these changes
on 2026-11-14
anujj
merged
8777c0c8
into master 1 day ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub