llama.cpp
hexagon: further optimization and tuning of matmul and dot kernels
#19407
Merged
Overview
Commits
7
Changes
View On GitHub
Commits
  • ggml-hexagon: implement 2x2 matmul kernel
    max-krasnyansky committed 121 days ago
  • hexmm: implement vec_dot_rx2x2 for Q8_0 and MXFP4
    max-krasnyansky committed 121 days ago
  • hexagon: fix editor config failures
    max-krasnyansky committed 121 days ago
  • hexagon: refactor matmul ops to use context struct and remove wrappers
    max-krasnyansky committed 121 days ago
  • hexagon: refactor dyn quantizers to use mmctx
    max-krasnyansky committed 121 days ago
  • hexagon: remove mm fastdiv from op_ctx
    max-krasnyansky committed 121 days ago
  • hexagon: refactor matmul entry point to reduce code duplication
    max-krasnyansky committed 121 days ago
FAQ Terms Privacy Refunds Impressum
Loading