Skip to content

cuda: opt-in split-K for the PQ2_0 one-column ffn_down mat-vec on Hopper (GGML_CUDA_PQ2_SPLITK) - #424

Merged
bri-prism merged 1 commit into
prismfrom
cuda/pq2-splitk-ffn-down
Oct 31, 2026
Merged

bri-prism merged 1 commit into
prismfrom
cuda/pq2-splitk-ffn-down

Commits

  1. Commits on Oct 31, 2026