Skip to content

Optimize F32 <- QAI8DXP 1x8 (LHS) x QSI4C32P 8x8 (RHS) for 1x8 sdot

Michael Kozlov requested to merge 1x8_1x8_int4_sdot_gemv into main
  • Add new assembly ukernel optimized with FEAT_DOTPROD for matrix multiplication with 1x8 block size.
  • Update build script.
  • Add to unit test.

Signed-off-by: Michael Kozlov michael.kozlov@arm.com

Merge request reports