Skip to content

GEMV kernel <- f16 LHS x QSI4c32p RHS

Hugo OKeeffe requested to merge hugoke01/int4_gemv into main

Micro kernel for GEMV of packed LHS matrix with f16 and RHS matrix with symmetric 4-bit integer with per-block quantisation and accumulation into a single precision matrix. This MR also implements testing for the used LHS Advanced SIMD packing kernel. The GEMV kernel has been optimised for SME2

Signed-off-by: Hugo OKeeffe hugo.okeeffe@arm.com

Edited by Hugo OKeeffe

Merge request reports