llama.cpp
CUDA: MMVQ nwarps=8 for bs=1 for dense models on DGX Spark
#26843

Merged

CUDA: MMVQ nwarps=8 for bs=1 for dense models on DGX Spark #26843

am17an merged 3 commits into ggml-org:master from ynankani:ynankani/mmvq_nwarp8

CUDA: MMVQ nwarps=8 for bs=1 for dense models on DGX Spark

713ae20b

github-actions added ggml

github-actions added CUDA

skip moe experts and allow others based on k geometry (allow only sma…

3eb76495

ynankani marked this pull request as ready for review 11 days ago

ynankani requested a review 11 days ago

am17an approved these changes on 2026-08-13

rename MMVQ DGX Spark params to GB10 and fix MSVC constexpr lambda ca…

2c1c2e58

am17an added merge ready

gaugarg-nv approved these changes on 2026-08-18

am17an merged 25ae3a9b into master 6 days ago

Reviewers

gaugarg-nv

am17an

Assignees

No one assigned

Labels

ggml merge ready CUDA

Milestone

No milestone