Introduction of gemm4xN and gemmMx4 for Q4_0 and Q8_0 for better performance results#8908

Merged

ggerganov merged 1 commit intoggerganov:masterfrom Srihari-mcw:q8_0_q4_0_fp16_delta_multiply_parallel

Aug 31, 2024

+149

Commits on Aug 7, 2024

Add loop unrolled 4xN and MX4 dimension GEMM functions with parallel delta multiplication
Srihari-mcw
committed