PR #24087 CUDA: enroll mul_mat_vec_q_moe into pdl

CUDA: enroll mul_mat_vec_q_moe into pdl #24087

JohannesGaessler merged 2 commits into ggml-org:master from ORippler:osimons/enroll_mul_mat_vec_q_moe_into_PDL

Enroll mul_mat_vec_q_moe into PDL, boosting MTP performance on BW

770a38f5

LC to overlap with following kernels

b508845c

ORippler requested a review 44 days ago

github-actions added Nvidia GPU

github-actions added ggml

JohannesGaessler approved these changes on 2026-06-04

am17an approved these changes on 2026-06-05

JohannesGaessler merged 2154a0fd into master 43 days ago

ORippler deleted the osimons/enroll_mul_mat_vec_q_moe_into_PDL branch 40 days ago

Reviewers

am17an

JohannesGaessler

Assignees

No one assigned

Labels

Nvidia GPU ggml

Milestone

No milestone