Merge pull request #4474 from ChipKerchner/sgemmIncopy_PR

Vectorize in-copy packing/copying for SGEMM - up to 4X faster.

Author: martin-frbg

kernel/power/KERNEL.POWER10

@@ -25,7 +25,7 @@ ZTRMMKERNEL	= zgemm_kernel_power10.S
 endif
 
 SGEMMKERNEL    =  sgemm_kernel_power10.c
-SGEMMINCOPY    = ../generic/gemm_ncopy_16.c
+SGEMMINCOPY    = sgemm_ncopy_16_power.c
 SGEMMITCOPY    = sgemm_tcopy_16_power8.S
 SGEMMONCOPY    =  ../generic/gemm_ncopy_8.c
 SGEMMOTCOPY    = sgemm_tcopy_8_power8.S

kernel/power/KERNEL.POWER8

@@ -50,7 +50,7 @@ CTRMMKERNEL	= ctrmm_kernel_8x4_power8.S
 ZTRMMKERNEL	= ztrmm_kernel_8x2_power8.S
 
 SGEMMKERNEL    =  sgemm_kernel_16x8_power8.S
-SGEMMINCOPY    = ../generic/gemm_ncopy_16.c
+SGEMMINCOPY    = sgemm_ncopy_16_power.c
 SGEMMITCOPY    = sgemm_tcopy_16_power8.S
 SGEMMONCOPY    =  ../generic/gemm_ncopy_8.c
 SGEMMOTCOPY    = sgemm_tcopy_8_power8.S

kernel/power/KERNEL.POWER9

@@ -13,7 +13,7 @@ CTRMMKERNEL	= cgemm_kernel_power9.S
 ZTRMMKERNEL	= zgemm_kernel_power9.S
 
 SGEMMKERNEL    =  sgemm_kernel_power9.S
-SGEMMINCOPY    = ../generic/gemm_ncopy_16.c
+SGEMMINCOPY    = sgemm_ncopy_16_power.c
 SGEMMITCOPY    = sgemm_tcopy_16_power8.S
 SGEMMONCOPY    =  ../generic/gemm_ncopy_8.c
 SGEMMOTCOPY    = sgemm_tcopy_8_power8.S