Skip to content

CUDA: Add fastdiv to k_bin_bcast*, giving 1-3% E2E performance#15872

Merged
JohannesGaessler merged 4 commits intoggml-org:masterfrom
ORippler:osimons/add_fastdiv_to_k_bin_bcast
Sep 10, 2025
Merged

CUDA: Add `fastdiv` to `k_bin_bcast*`, giving 1-3% E2E performance#15872
JohannesGaessler merged 4 commits intoggml-org:masterfrom
ORippler:osimons/add_fastdiv_to_k_bin_bcast

Commits

Commits on Sep 8, 2025

Commits on Sep 10, 2025