llama.cpp
Introduction of gemm4xN and gemmMx4 for Q4_0 and Q8_0 for better performance results
#8908
Merged
Merged