Functional Assembly Kernels for CortexA57

Adding functional (non-optimized) kernels for Cortex-A57
with the following layouts.
SGEMM - 16x4, 8x8
CGEMM - 8x4
DGEMM - 8x4, 4x8

kernel/arm64/KERNEL.CORTEXA57

@@ -60,32 +60,55 @@ DGEMVTKERNEL = gemv_t.S
 CGEMVTKERNEL = zgemv_t.S
 ZGEMVTKERNEL = zgemv_t.S
 
-STRMMKERNEL = strmm_kernel_4x4.S
-DTRMMKERNEL = dtrmm_kernel_4x4.S
-CTRMMKERNEL = ctrmm_kernel_4x4.S
-ZTRMMKERNEL = ztrmm_kernel_4x4.S
-
-SGEMMKERNEL    =  sgemm_kernel_4x4.S
-SGEMMONCOPY    =  ../generic/gemm_ncopy_4.c
-SGEMMOTCOPY    =  ../generic/gemm_tcopy_4.c
+SGEMMKERNEL    =  sgemm_kernel_$(SGEMM_UNROLL_M)x$(SGEMM_UNROLL_N).S
+STRMMKERNEL    =  strmm_kernel_$(SGEMM_UNROLL_M)x$(SGEMM_UNROLL_N).S
+ifneq ($(SGEMM_UNROLL_M), $(SGEMM_UNROLL_N))
+SGEMMINCOPY    =  ../generic/gemm_ncopy_$(SGEMM_UNROLL_M).c
+SGEMMITCOPY    =  ../generic/gemm_tcopy_$(SGEMM_UNROLL_M).c
+SGEMMINCOPYOBJ =  sgemm_incopy.o
+SGEMMITCOPYOBJ =  sgemm_itcopy.o
+endif
+SGEMMONCOPY    =  ../generic/gemm_ncopy_$(SGEMM_UNROLL_N).c
+SGEMMOTCOPY    =  ../generic/gemm_tcopy_$(SGEMM_UNROLL_N).c
 SGEMMONCOPYOBJ =  sgemm_oncopy.o
 SGEMMOTCOPYOBJ =  sgemm_otcopy.o
 
-DGEMMKERNEL    =  dgemm_kernel_4x4.S
-DGEMMONCOPY    =  ../generic/gemm_ncopy_4.c
-DGEMMOTCOPY    =  ../generic/gemm_tcopy_4.c
+DGEMMKERNEL    =  dgemm_kernel_$(DGEMM_UNROLL_M)x$(DGEMM_UNROLL_N).S
+DTRMMKERNEL    =  dtrmm_kernel_$(DGEMM_UNROLL_M)x$(DGEMM_UNROLL_N).S
+ifneq ($(DGEMM_UNROLL_M), $(DGEMM_UNROLL_N))
+DGEMMINCOPY    =  ../generic/gemm_ncopy_$(DGEMM_UNROLL_M).c
+DGEMMITCOPY    =  ../generic/gemm_tcopy_$(DGEMM_UNROLL_M).c
+DGEMMINCOPYOBJ =  dgemm_incopy.o
+DGEMMITCOPYOBJ =  dgemm_itcopy.o
+endif
+DGEMMONCOPY    =  ../generic/gemm_ncopy_$(DGEMM_UNROLL_N).c
+DGEMMOTCOPY    =  ../generic/gemm_tcopy_$(DGEMM_UNROLL_N).c
 DGEMMONCOPYOBJ =  dgemm_oncopy.o
 DGEMMOTCOPYOBJ =  dgemm_otcopy.o
 
-CGEMMKERNEL    =  cgemm_kernel_4x4.S
-CGEMMONCOPY    =  ../generic/zgemm_ncopy_4.c
-CGEMMOTCOPY    =  ../generic/zgemm_tcopy_4.c
+CGEMMKERNEL    =  cgemm_kernel_$(CGEMM_UNROLL_M)x$(CGEMM_UNROLL_N).S
+CTRMMKERNEL    =  ctrmm_kernel_$(CGEMM_UNROLL_M)x$(CGEMM_UNROLL_N).S
+ifneq ($(CGEMM_UNROLL_M), $(CGEMM_UNROLL_N))
+CGEMMINCOPY    =  ../generic/zgemm_ncopy_$(CGEMM_UNROLL_M).c
+CGEMMITCOPY    =  ../generic/zgemm_tcopy_$(CGEMM_UNROLL_M).c
+CGEMMINCOPYOBJ =  cgemm_incopy.o
+CGEMMITCOPYOBJ =  cgemm_itcopy.o
+endif
+CGEMMONCOPY    =  ../generic/zgemm_ncopy_$(CGEMM_UNROLL_N).c
+CGEMMOTCOPY    =  ../generic/zgemm_tcopy_$(CGEMM_UNROLL_N).c
 CGEMMONCOPYOBJ =  cgemm_oncopy.o
 CGEMMOTCOPYOBJ =  cgemm_otcopy.o
 
-ZGEMMKERNEL    =  zgemm_kernel_4x4.S
-ZGEMMONCOPY    =  ../generic/zgemm_ncopy_4.c
-ZGEMMOTCOPY    =  ../generic/zgemm_tcopy_4.c
+ZGEMMKERNEL    =  zgemm_kernel_$(ZGEMM_UNROLL_M)x$(ZGEMM_UNROLL_N).S
+ZTRMMKERNEL    =  ztrmm_kernel_$(ZGEMM_UNROLL_M)x$(ZGEMM_UNROLL_N).S
+ifneq ($(ZGEMM_UNROLL_M), $(ZGEMM_UNROLL_N))
+ZGEMMINCOPY    =  ../generic/zgemm_ncopy_$(ZGEMM_UNROLL_M).c
+ZGEMMITCOPY    =  ../generic/zgemm_tcopy_$(ZGEMM_UNROLL_M).c
+ZGEMMINCOPYOBJ =  zgemm_incopy.o
+ZGEMMITCOPYOBJ =  zgemm_itcopy.o
+endif
+ZGEMMONCOPY    =  ../generic/zgemm_ncopy_$(ZGEMM_UNROLL_N).c
+ZGEMMOTCOPY    =  ../generic/zgemm_tcopy_$(ZGEMM_UNROLL_N).c
 ZGEMMONCOPYOBJ =  zgemm_oncopy.o
 ZGEMMOTCOPYOBJ =  zgemm_otcopy.o