Add ZEN support (tested for auto-detected static backend)

This commit is contained in:
Denis Steckelmacher
2017-03-19 15:32:50 +01:00
parent 99880f7906
commit c9ff735da6
46 changed files with 339 additions and 58 deletions

98
kernel/x86_64/KERNEL.ZEN Normal file
View File

@@ -0,0 +1,98 @@
DSCALKERNEL = dscal.c
CSCALKERNEL = cscal.c
ZSCALKERNEL = zscal.c
SGEMVNKERNEL = sgemv_n_4.c
SGEMVTKERNEL = sgemv_t_4.c
DGEMVNKERNEL = dgemv_n_4.c
DGEMVTKERNEL = dgemv_t_4.c
ZGEMVNKERNEL = zgemv_n_4.c
ZGEMVTKERNEL = zgemv_t_4.c
CGEMVNKERNEL = cgemv_n_4.c
CGEMVTKERNEL = cgemv_t_4.c
SSYMV_L_KERNEL = ssymv_L.c
SSYMV_U_KERNEL = ssymv_U.c
DSYMV_L_KERNEL = dsymv_L.c
DSYMV_U_KERNEL = dsymv_U.c
SDOTKERNEL = sdot.c
DDOTKERNEL = ddot.c
CDOTKERNEL = cdot.c
ZDOTKERNEL = zdot.c
SAXPYKERNEL = saxpy.c
DAXPYKERNEL = daxpy.c
CAXPYKERNEL = caxpy.c
ZAXPYKERNEL = zaxpy.c
STRMMKERNEL = sgemm_kernel_16x4_haswell.S
SGEMMKERNEL = sgemm_kernel_16x4_haswell.S
SGEMMINCOPY = ../generic/gemm_ncopy_16.c
SGEMMITCOPY = ../generic/gemm_tcopy_16.c
SGEMMONCOPY = ../generic/gemm_ncopy_4.c
SGEMMOTCOPY = ../generic/gemm_tcopy_4.c
SGEMMINCOPYOBJ = sgemm_incopy$(TSUFFIX).$(SUFFIX)
SGEMMITCOPYOBJ = sgemm_itcopy$(TSUFFIX).$(SUFFIX)
SGEMMONCOPYOBJ = sgemm_oncopy$(TSUFFIX).$(SUFFIX)
SGEMMOTCOPYOBJ = sgemm_otcopy$(TSUFFIX).$(SUFFIX)
DTRMMKERNEL = dtrmm_kernel_4x8_haswell.c
DGEMMKERNEL = dgemm_kernel_4x8_haswell.S
DGEMMINCOPY = ../generic/gemm_ncopy_4.c
DGEMMITCOPY = ../generic/gemm_tcopy_4.c
DGEMMONCOPY = ../generic/gemm_ncopy_8.c
DGEMMOTCOPY = ../generic/gemm_tcopy_8.c
DGEMMINCOPYOBJ = dgemm_incopy$(TSUFFIX).$(SUFFIX)
DGEMMITCOPYOBJ = dgemm_itcopy$(TSUFFIX).$(SUFFIX)
DGEMMONCOPYOBJ = dgemm_oncopy$(TSUFFIX).$(SUFFIX)
DGEMMOTCOPYOBJ = dgemm_otcopy$(TSUFFIX).$(SUFFIX)
CTRMMKERNEL = cgemm_kernel_8x2_haswell.S
CGEMMKERNEL = cgemm_kernel_8x2_haswell.S
CGEMMINCOPY = ../generic/zgemm_ncopy_8.c
CGEMMITCOPY = ../generic/zgemm_tcopy_8.c
CGEMMONCOPY = ../generic/zgemm_ncopy_2.c
CGEMMOTCOPY = ../generic/zgemm_tcopy_2.c
CGEMMINCOPYOBJ = cgemm_incopy$(TSUFFIX).$(SUFFIX)
CGEMMITCOPYOBJ = cgemm_itcopy$(TSUFFIX).$(SUFFIX)
CGEMMONCOPYOBJ = cgemm_oncopy$(TSUFFIX).$(SUFFIX)
CGEMMOTCOPYOBJ = cgemm_otcopy$(TSUFFIX).$(SUFFIX)
ZTRMMKERNEL = zgemm_kernel_4x2_haswell.S
ZGEMMKERNEL = zgemm_kernel_4x2_haswell.S
ZGEMMINCOPY = ../generic/zgemm_ncopy_4.c
ZGEMMITCOPY = ../generic/zgemm_tcopy_4.c
ZGEMMONCOPY = ../generic/zgemm_ncopy_2.c
ZGEMMOTCOPY = ../generic/zgemm_tcopy_2.c
ZGEMMINCOPYOBJ = zgemm_incopy$(TSUFFIX).$(SUFFIX)
ZGEMMITCOPYOBJ = zgemm_itcopy$(TSUFFIX).$(SUFFIX)
ZGEMMONCOPYOBJ = zgemm_oncopy$(TSUFFIX).$(SUFFIX)
ZGEMMOTCOPYOBJ = zgemm_otcopy$(TSUFFIX).$(SUFFIX)
STRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
STRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
STRSMKERNEL_RN = ../generic/trsm_kernel_RN.c
STRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
DTRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
DTRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
DTRSMKERNEL_RN = dtrsm_kernel_RN_haswell.c
DTRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
CTRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
CTRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
CTRSMKERNEL_RN = ../generic/trsm_kernel_RN.c
CTRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
ZTRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
ZTRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
ZTRSMKERNEL_RN = ../generic/trsm_kernel_RN.c
ZTRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
CGEMM3MKERNEL = zgemm3m_kernel_4x8_nehalem.S
ZGEMM3MKERNEL = zgemm3m_kernel_2x8_nehalem.S

View File

@@ -33,7 +33,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "caxpy_microk_steamroller-2.c"
#elif defined(BULLDOZER)
#include "caxpy_microk_bulldozer-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "caxpy_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "caxpy_microk_sandy-2.c"

View File

@@ -34,7 +34,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "cdot_microk_bulldozer-2.c"
#elif defined(STEAMROLLER) || defined(PILEDRIVER) || defined(EXCAVATOR)
#include "cdot_microk_steamroller-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "cdot_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "cdot_microk_sandy-2.c"

View File

@@ -29,7 +29,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include <stdio.h>
#include "common.h"
#if defined(HASWELL)
#if defined(HASWELL) || defined(ZEN)
#include "cgemv_n_microk_haswell-4.c"
#elif defined(BULLDOZER) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "cgemv_n_microk_bulldozer-4.c"

View File

@@ -28,7 +28,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "common.h"
#if defined(HASWELL)
#if defined(HASWELL) || defined(ZEN)
#include "cgemv_t_microk_haswell-4.c"
#elif defined(BULLDOZER) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "cgemv_t_microk_bulldozer-4.c"

View File

@@ -28,7 +28,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "common.h"
#if defined(HASWELL)
#if defined(HASWELL) || defined(ZEN)
#include "cscal_microk_haswell-2.c"
#elif defined(BULLDOZER) || defined(PILEDRIVER)
#include "cscal_microk_bulldozer-2.c"

View File

@@ -37,7 +37,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "daxpy_microk_steamroller-2.c"
#elif defined(PILEDRIVER)
#include "daxpy_microk_piledriver-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "daxpy_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "daxpy_microk_sandy-2.c"

View File

@@ -37,7 +37,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "ddot_microk_piledriver-2.c"
#elif defined(NEHALEM)
#include "ddot_microk_nehalem-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "ddot_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "ddot_microk_sandy-2.c"

View File

@@ -31,7 +31,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#if defined(NEHALEM)
#include "dgemv_n_microk_nehalem-4.c"
#elif defined(HASWELL) || defined(STEAMROLLER) || defined(EXCAVATOR)
#elif defined(HASWELL) || defined(ZEN) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "dgemv_n_microk_haswell-4.c"
#endif

View File

@@ -28,7 +28,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "common.h"
#if defined(HASWELL) || defined(STEAMROLLER) || defined(EXCAVATOR)
#if defined(HASWELL) || defined(ZEN) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "dgemv_t_microk_haswell-4.c"
#endif

View File

@@ -31,7 +31,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "dscal_microk_bulldozer-2.c"
#elif defined(SANDYBRIDGE)
#include "dscal_microk_sandy-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "dscal_microk_haswell-2.c"
#endif

View File

@@ -30,7 +30,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#if defined(BULLDOZER) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "dsymv_L_microk_bulldozer-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "dsymv_L_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "dsymv_L_microk_sandy-2.c"

View File

@@ -31,7 +31,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#if defined(BULLDOZER) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "dsymv_U_microk_bulldozer-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "dsymv_U_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "dsymv_U_microk_sandy-2.c"

View File

@@ -31,7 +31,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#if defined(NEHALEM)
#include "saxpy_microk_nehalem-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "saxpy_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "saxpy_microk_sandy-2.c"

View File

@@ -34,7 +34,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "sdot_microk_steamroller-2.c"
#elif defined(NEHALEM)
#include "sdot_microk_nehalem-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "sdot_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "sdot_microk_sandy-2.c"

View File

@@ -35,7 +35,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "sgemv_n_microk_nehalem-4.c"
#elif defined(SANDYBRIDGE)
#include "sgemv_n_microk_sandy-4.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "sgemv_n_microk_haswell-4.c"
#endif

View File

@@ -34,7 +34,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "sgemv_t_microk_bulldozer-4.c"
#elif defined(SANDYBRIDGE)
#include "sgemv_t_microk_sandy-4.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "sgemv_t_microk_haswell-4.c"
#endif

View File

@@ -32,7 +32,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "ssymv_L_microk_bulldozer-2.c"
#elif defined(NEHALEM)
#include "ssymv_L_microk_nehalem-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "ssymv_L_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "ssymv_L_microk_sandy-2.c"

View File

@@ -33,7 +33,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "ssymv_U_microk_bulldozer-2.c"
#elif defined(NEHALEM)
#include "ssymv_U_microk_nehalem-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "ssymv_U_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "ssymv_U_microk_sandy-2.c"

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 12)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 12)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 12)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 12)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 12)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 12)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 12)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 24)

View File

@@ -33,7 +33,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "zaxpy_microk_bulldozer-2.c"
#elif defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "zaxpy_microk_steamroller-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "zaxpy_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "zaxpy_microk_sandy-2.c"

View File

@@ -33,7 +33,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "zdot_microk_bulldozer-2.c"
#elif defined(STEAMROLLER) || defined(PILEDRIVER) || defined(EXCAVATOR)
#include "zdot_microk_steamroller-2.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "zdot_microk_haswell-2.c"
#elif defined(SANDYBRIDGE)
#include "zdot_microk_sandy-2.c"

View File

@@ -30,7 +30,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "common.h"
#if defined(HASWELL)
#if defined(HASWELL) || defined(ZEN)
#include "zgemv_n_microk_haswell-4.c"
#elif defined(SANDYBRIDGE)
#include "zgemv_n_microk_sandy-4.c"

View File

@@ -31,7 +31,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#if defined(BULLDOZER) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "zgemv_t_microk_bulldozer-4.c"
#elif defined(HASWELL)
#elif defined(HASWELL) || defined(ZEN)
#include "zgemv_t_microk_haswell-4.c"
#endif

View File

@@ -28,7 +28,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "common.h"
#if defined(HASWELL)
#if defined(HASWELL) || defined(ZEN)
#include "zscal_microk_haswell-2.c"
#elif defined(BULLDOZER) || defined(PILEDRIVER)
#include "zscal_microk_bulldozer-2.c"

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 24)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 24)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 24)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 24)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 24)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 24)

View File

@@ -57,7 +57,7 @@
#define PREFETCHSIZE (16 * 24)
#endif
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL)
#if defined(NEHALEM) || defined(SANDYBRIDGE) || defined(HASWELL) || defined(ZEN)
#define PREFETCH prefetcht0
#define PREFETCHW prefetcht0
#define PREFETCHSIZE (16 * 24)