Added 30xk support for packm ukernels.

Details:
- Updated bli_kernel_*_macro_defs.h headers to include default
  definitions for 30xk packm kernels.
- Extended function pointer arrays in bli_packm_cxk_*() out to 31 and
  included 30xk kernels.
- Addex 30xk kernels to frame/1m/packm/ukernels/bli_packm_ref_cxk_*.c.
This commit is contained in:
Field G. Van Zee
2014-09-22 16:06:15 -05:00
parent b6d4bd792e
commit 4a7df04e8a
16 changed files with 1356 additions and 67 deletions

View File

@@ -186,6 +186,16 @@
#define BLIS_ZPACKM_16XK_3M_KERNEL BLIS_ZPACKM_16XK_3M_KERNEL_REF
#endif
// packm_30xk_3m kernels
#ifndef BLIS_CPACKM_30XK_3M_KERNEL
#define BLIS_CPACKM_30XK_3M_KERNEL BLIS_CPACKM_30XK_3M_KERNEL_REF
#endif
#ifndef BLIS_ZPACKM_30XK_3M_KERNEL
#define BLIS_ZPACKM_30XK_3M_KERNEL BLIS_ZPACKM_30XK_3M_KERNEL_REF
#endif
#endif

View File

@@ -186,6 +186,16 @@
#define BLIS_ZPACKM_16XK_4M_KERNEL BLIS_ZPACKM_16XK_4M_KERNEL_REF
#endif
// packm_30xk_4m kernels
#ifndef BLIS_CPACKM_30XK_4M_KERNEL
#define BLIS_CPACKM_30XK_4M_KERNEL BLIS_CPACKM_30XK_4M_KERNEL_REF
#endif
#ifndef BLIS_ZPACKM_30XK_4M_KERNEL
#define BLIS_ZPACKM_30XK_4M_KERNEL BLIS_ZPACKM_30XK_4M_KERNEL_REF
#endif
#endif

View File

@@ -333,6 +333,24 @@
#define BLIS_ZPACKM_16XK_KERNEL BLIS_ZPACKM_16XK_KERNEL_REF
#endif
// packm_30xk kernels
#ifndef BLIS_SPACKM_30XK_KERNEL
#define BLIS_SPACKM_30XK_KERNEL BLIS_SPACKM_30XK_KERNEL_REF
#endif
#ifndef BLIS_DPACKM_30XK_KERNEL
#define BLIS_DPACKM_30XK_KERNEL BLIS_DPACKM_30XK_KERNEL_REF
#endif
#ifndef BLIS_CPACKM_30XK_KERNEL
#define BLIS_CPACKM_30XK_KERNEL BLIS_CPACKM_30XK_KERNEL_REF
#endif
#ifndef BLIS_ZPACKM_30XK_KERNEL
#define BLIS_ZPACKM_30XK_KERNEL BLIS_ZPACKM_30XK_KERNEL_REF
#endif
// unpackm_2xk kernels
#ifndef BLIS_SUNPACKM_2XK_KERNEL

View File

@@ -212,6 +212,13 @@
#define BLIS_CPACKM_16XK_KERNEL_REF bli_cpackm_ref_16xk
#define BLIS_ZPACKM_16XK_KERNEL_REF bli_zpackm_ref_16xk
// packm_30xk kernels
#define BLIS_SPACKM_30XK_KERNEL_REF bli_spackm_ref_30xk
#define BLIS_DPACKM_30XK_KERNEL_REF bli_dpackm_ref_30xk
#define BLIS_CPACKM_30XK_KERNEL_REF bli_cpackm_ref_30xk
#define BLIS_ZPACKM_30XK_KERNEL_REF bli_zpackm_ref_30xk
// packm_2xk_4m kernels
#define BLIS_CPACKM_2XK_4M_KERNEL_REF bli_cpackm_ref_2xk_4m
@@ -252,6 +259,11 @@
#define BLIS_CPACKM_16XK_4M_KERNEL_REF bli_cpackm_ref_16xk_4m
#define BLIS_ZPACKM_16XK_4M_KERNEL_REF bli_zpackm_ref_16xk_4m
// packm_30xk_4m kernels
#define BLIS_CPACKM_30XK_4M_KERNEL_REF bli_cpackm_ref_30xk_4m
#define BLIS_ZPACKM_30XK_4M_KERNEL_REF bli_zpackm_ref_30xk_4m
// packm_2xk_3m kernels
#define BLIS_CPACKM_2XK_3M_KERNEL_REF bli_cpackm_ref_2xk_3m
@@ -292,6 +304,11 @@
#define BLIS_CPACKM_16XK_3M_KERNEL_REF bli_cpackm_ref_16xk_3m
#define BLIS_ZPACKM_16XK_3M_KERNEL_REF bli_zpackm_ref_16xk_3m
// packm_30xk_3m kernels
#define BLIS_CPACKM_30XK_3M_KERNEL_REF bli_cpackm_ref_30xk_3m
#define BLIS_ZPACKM_30XK_3M_KERNEL_REF bli_zpackm_ref_30xk_3m
// packm_2xk_rih kernels
#define BLIS_CPACKM_2XK_RIH_KERNEL_REF bli_cpackm_ref_2xk_rih
@@ -332,6 +349,11 @@
#define BLIS_CPACKM_16XK_RIH_KERNEL_REF bli_cpackm_ref_16xk_rih
#define BLIS_ZPACKM_16XK_RIH_KERNEL_REF bli_zpackm_ref_16xk_rih
// packm_30xk_rih kernels
#define BLIS_CPACKM_30XK_RIH_KERNEL_REF bli_cpackm_ref_30xk_rih
#define BLIS_ZPACKM_30XK_RIH_KERNEL_REF bli_zpackm_ref_30xk_rih
// unpack_2xk kernels
#define BLIS_SUNPACKM_2XK_KERNEL_REF bli_sunpackm_ref_2xk

View File

@@ -163,6 +163,16 @@
#define BLIS_ZPACKM_16XK_RIH_KERNEL BLIS_ZPACKM_16XK_RIH_KERNEL_REF
#endif
// packm_30xk_rih kernels
#ifndef BLIS_CPACKM_30XK_RIH_KERNEL
#define BLIS_CPACKM_30XK_RIH_KERNEL BLIS_CPACKM_30XK_RIH_KERNEL_REF
#endif
#ifndef BLIS_ZPACKM_30XK_RIH_KERNEL
#define BLIS_ZPACKM_30XK_RIH_KERNEL BLIS_ZPACKM_30XK_RIH_KERNEL_REF
#endif
#endif