llama.cpp
vulkan: add IQ3_S MMQ matmul kernels
#28822
Merged
Overview
Commits
3
Changes
View On GitHub
Commits
  • vulkan: add IQ3_S MMQ matmul kernels
    vargad committed 12 days ago
  • Make block_a_to_shmem do 2-byte loads (110 bytes is divisible by 2)
    vargad committed 8 days ago
  • Align the check, IQ3_S is also using K tile size
    vargad committed 8 days ago
FAQ Terms Privacy Refunds Impressum
Loading