Skip to content

Commit 42f0201

Browse files
authored
Merge pull request #20 from xianyi/develop
rebase
2 parents fa9a30b + 22db876 commit 42f0201

File tree

7 files changed

+30
-8
lines changed

7 files changed

+30
-8
lines changed

getarch_2nd.c

Lines changed: 9 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,15 @@
44
#else
55
#include "config_kernel.h"
66
#endif
7-
#include "common.h"
7+
#if (defined(__WIN32__) || defined(__WIN64__) || defined(__CYGWIN32__) || defined(__CYGWIN64__) || defined(_WIN32) || defined(_WIN64)) && defined(__64BIT__)
8+
typedef long long BLASLONG;
9+
typedef unsigned long long BLASULONG;
10+
#else
11+
typedef long BLASLONG;
12+
typedef unsigned long BLASULONG;
13+
#endif
14+
15+
#include "param.h"
816

917
int main(int argc, char **argv) {
1018

kernel/power/KERNEL.POWER10

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -169,8 +169,13 @@ ZROTKERNEL = zrot.c
169169
#
170170
SSCALKERNEL = sscal.c
171171
DSCALKERNEL = dscal.c
172+
ifeq ($(C_COMPILER), PGI)
173+
CSCALKERNEL = ../arm/zscal.c
174+
ZSCALKERNEL = ../arm/zscal.c
175+
else
172176
CSCALKERNEL = zscal.c
173177
ZSCALKERNEL = zscal.c
178+
endif
174179
#
175180
SSWAPKERNEL = sswap.c
176181
DSWAPKERNEL = dswap.c

kernel/power/KERNEL.POWER8

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -242,8 +242,13 @@ ZROTKERNEL = zrot.c
242242
#
243243
SSCALKERNEL = sscal.c
244244
DSCALKERNEL = dscal.c
245+
ifeq ($(C_COMPILER), PGI)
246+
CSCALKERNEL = ../arm/zscal.c
247+
ZSCALKERNEL = ../arm/zscal.c
248+
else
245249
CSCALKERNEL = zscal.c
246250
ZSCALKERNEL = zscal.c
251+
endif
247252
#
248253
SSWAPKERNEL = sswap.c
249254
DSWAPKERNEL = dswap.c

kernel/power/KERNEL.POWER9

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -166,8 +166,13 @@ ZROTKERNEL = zrot.c
166166
#
167167
SSCALKERNEL = sscal.c
168168
DSCALKERNEL = dscal.c
169+
ifeq ($(C_COMPILER), PGI)
170+
CSCALKERNEL = ../arm/zscal.c
171+
ZSCALKERNEL = ../arm/zscal.c
172+
else
169173
CSCALKERNEL = zscal.c
170174
ZSCALKERNEL = zscal.c
175+
endif
171176
#
172177
SSWAPKERNEL = sswap.c
173178
DSWAPKERNEL = dswap.c

kernel/x86_64/KERNEL

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -491,4 +491,3 @@ SSUMKERNEL = ../arm/sum.c
491491
DSUMKERNEL = ../arm/sum.c
492492

493493
SOMATCOPY_RT = omatcopy_rt.c
494-
DOMATCOPY_RT = omatcopy_rt.c

kernel/x86_64/sgemm_direct_skylakex.c

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,10 @@
11
/* the direct sgemm code written by Arjan van der Ven */
2-
2+
#include "common.h"
33

44
#if defined(SKYLAKEX) || defined (COOPERLAKE)
55

66
#include <immintrin.h>
7-
#include "common.h"
7+
88

99
/*
1010
* "Direct sgemm" code. This code operates directly on the inputs and outputs
@@ -472,7 +472,7 @@ void CNAME (BLASLONG M, BLASLONG N, BLASLONG K, float * __restrict A, BLASLONG s
472472
}
473473
}
474474
#else
475-
#include "common.h"
475+
476476
void CNAME (BLASLONG M, BLASLONG N, BLASLONG K, float * __restrict A, BLASLONG strideA, float * __restrict B, BLASLONG strideB , float * __restrict R, BLASLONG strideR)
477477
{}
478478
#endif

param.h

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2466,13 +2466,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24662466

24672467
#define SGEMM_DEFAULT_P 512
24682468
#define DGEMM_DEFAULT_P 384
2469-
#define CGEMM_DEFAULT_P 512
2469+
#define CGEMM_DEFAULT_P 512
24702470
#define ZGEMM_DEFAULT_P 256
24712471

24722472
#define SGEMM_DEFAULT_Q 512
24732473
#define DGEMM_DEFAULT_Q 512
2474-
#define CGEMM_DEFAULT_Q 1026
2475-
#define ZGEMM_DEFAULT_Q 1026
2474+
#define CGEMM_DEFAULT_Q 384
2475+
#define ZGEMM_DEFAULT_Q 384
24762476

24772477
#define SGEMM_DEFAULT_R 4096
24782478
#define DGEMM_DEFAULT_R 4096

0 commit comments

Comments
 (0)