Use AVX512 also for DGEMM

this required switching to the generic gemm_beta code (which is faster anyway on SKX)
for both DGEMM and SGEMM

Performance for the not-retuned version is in the 30% range

kernel/x86_64/KERNEL.SKYLAKEX

@@ -2,3 +2,18 @@ include $(KERNELDIR)/KERNEL.HASWELL
 
 SGEMMKERNEL    =  sgemm_kernel_16x4_skylakex.S
 
+
+DTRMMKERNEL    =  ../generic/trmmkernel_16x2.c
+DGEMMKERNEL    =  dgemm_kernel_16x2_skylakex.S
+DGEMMINCOPY    =  ../generic/gemm_ncopy_16.c
+DGEMMITCOPY    =  ../generic/gemm_tcopy_16.c
+DGEMMONCOPY    =  ../generic/gemm_ncopy_2.c
+DGEMMOTCOPY    =  ../generic/gemm_tcopy_2.c
+DGEMMINCOPYOBJ =  dgemm_incopy$(TSUFFIX).$(SUFFIX)
+DGEMMITCOPYOBJ =  dgemm_itcopy$(TSUFFIX).$(SUFFIX)
+DGEMMONCOPYOBJ =  dgemm_oncopy$(TSUFFIX).$(SUFFIX)
+DGEMMOTCOPYOBJ =  dgemm_otcopy$(TSUFFIX).$(SUFFIX)
+
+
+SGEMM_BETA = ../generic/gemm_beta.c
+DGEMM_BETA = ../generic/gemm_beta.c

kernel/x86_64/sgemm_kernel_16x4_skylakex.S

@@ -159,7 +159,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-16 * SIZE(AO), %zmm0
 	vbroadcastss	 -4 * SIZE(BO), %zmm2
 	vbroadcastss	 -3 * SIZE(BO), %zmm3
-	prefetcht0	A_PR1(AO)
+#	prefetcht0	A_PR1(AO)
 
 	VFMADD231PS_(  	%zmm4,%zmm2,%zmm0 )
 	VFMADD231PS_(  	%zmm6,%zmm3,%zmm0 )
@@ -183,7 +183,6 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 	vmovups 	-16 * SIZE(AO), %zmm0
 	vbroadcastss	 -4 * SIZE(BO), %zmm2
 	vbroadcastss	 -3 * SIZE(BO), %zmm3
-	prefetcht0	A_PR1(AO)
 
 	VFMADD231PS_(  	%zmm4,%zmm2,%zmm0 )
 	VFMADD231PS_(  	%zmm6,%zmm3,%zmm0 )