Skip to content

Commit e58f6dc

Browse files
authored
Add extensions ?GEMM_BATCH_STRIDED and CBLAS_?GEMM_BATCH_STRIDED (#5458)
* Add ?GEMM_BATCH_STRIDED and CBLAS_?GEMM_BATCH_STRIDED
1 parent c92f7f6 commit e58f6dc

File tree

4 files changed

+479
-10
lines changed

4 files changed

+479
-10
lines changed

cblas.h

Lines changed: 9 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -456,6 +456,14 @@ void cblas_cgemm_batch(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enu
456456
void cblas_zgemm_batch(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE * TransA_array, OPENBLAS_CONST enum CBLAS_TRANSPOSE * TransB_array, OPENBLAS_CONST blasint * M_array, OPENBLAS_CONST blasint * N_array, OPENBLAS_CONST blasint * K_array,
457457
OPENBLAS_CONST void * alpha_array, OPENBLAS_CONST void ** A_array, OPENBLAS_CONST blasint * lda_array, OPENBLAS_CONST void ** B_array, OPENBLAS_CONST blasint * ldb_array, OPENBLAS_CONST void * beta_array, void ** C_array, OPENBLAS_CONST blasint * ldc_array, OPENBLAS_CONST blasint group_count, OPENBLAS_CONST blasint * group_size);
458458

459+
void cblas_sgemm_batch_strided(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float alpha, OPENBLAS_CONST float * A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST blasint stridea, OPENBLAS_CONST float * B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST blasint strideb, OPENBLAS_CONST float beta, float * C, OPENBLAS_CONST blasint ldc, OPENBLAS_CONST blasint stridec, OPENBLAS_CONST blasint group_size);
460+
461+
void cblas_dgemm_batch_strided(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST double alpha, OPENBLAS_CONST double * A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST blasint stridea, OPENBLAS_CONST double * B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST blasint strideb, OPENBLAS_CONST double beta, double * C, OPENBLAS_CONST blasint ldc, OPENBLAS_CONST blasint stridec, OPENBLAS_CONST blasint group_size);
462+
463+
void cblas_cgemm_batch_strided(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void * alpha, OPENBLAS_CONST void * A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST blasint stridea, OPENBLAS_CONST void * B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST blasint strideb, OPENBLAS_CONST void * beta, void * C, OPENBLAS_CONST blasint ldc, OPENBLAS_CONST blasint stridec, OPENBLAS_CONST blasint group_size);
464+
465+
void cblas_zgemm_batch_strided(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST void * alpha, OPENBLAS_CONST void * A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST blasint stridea, OPENBLAS_CONST void * B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST blasint strideb, OPENBLAS_CONST void * beta, void * C, OPENBLAS_CONST blasint ldc, OPENBLAS_CONST blasint stridec, OPENBLAS_CONST blasint group_size);
466+
459467
/*** BFLOAT16 and INT8 extensions ***/
460468
/* convert float array to BFLOAT16 array by rounding */
461469
void cblas_sbstobf16(OPENBLAS_CONST blasint n, OPENBLAS_CONST float *in, OPENBLAS_CONST blasint incin, bfloat16 *out, OPENBLAS_CONST blasint incout);
@@ -477,6 +485,7 @@ void cblas_sbgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum C
477485
void cblas_sbgemm_batch(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE * TransA_array, OPENBLAS_CONST enum CBLAS_TRANSPOSE * TransB_array, OPENBLAS_CONST blasint * M_array, OPENBLAS_CONST blasint * N_array, OPENBLAS_CONST blasint * K_array,
478486
OPENBLAS_CONST float * alpha_array, OPENBLAS_CONST bfloat16 ** A_array, OPENBLAS_CONST blasint * lda_array, OPENBLAS_CONST bfloat16 ** B_array, OPENBLAS_CONST blasint * ldb_array, OPENBLAS_CONST float * beta_array, float ** C_array, OPENBLAS_CONST blasint * ldc_array, OPENBLAS_CONST blasint group_count, OPENBLAS_CONST blasint * group_size);
479487

488+
void cblas_sbgemm_batch_strided(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K, OPENBLAS_CONST float alpha, OPENBLAS_CONST bfloat16 * A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST blasint stridea, OPENBLAS_CONST bfloat16 * B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST blasint strideb, OPENBLAS_CONST float beta, float * C, OPENBLAS_CONST blasint ldc, OPENBLAS_CONST blasint stridec, OPENBLAS_CONST blasint group_size);
480489
/*** FLOAT16 extensions ***/
481490
void cblas_shgemm(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransA, OPENBLAS_CONST enum CBLAS_TRANSPOSE TransB, OPENBLAS_CONST blasint M, OPENBLAS_CONST blasint N, OPENBLAS_CONST blasint K,
482491
OPENBLAS_CONST float alpha, OPENBLAS_CONST hfloat16 *A, OPENBLAS_CONST blasint lda, OPENBLAS_CONST hfloat16 *B, OPENBLAS_CONST blasint ldb, OPENBLAS_CONST float beta, float *C, OPENBLAS_CONST blasint ldc);

interface/CMakeLists.txt

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -125,6 +125,7 @@ foreach (CBLAS_FLAG ${CBLAS_FLAGS})
125125
if (BUILD_SINGLE OR BUILD_DOUBLE)
126126
GenerateNamedObjects("sdsdot.c" "" "sdsdot" ${CBLAS_FLAG} "" "" true "SINGLE")
127127
GenerateNamedObjects("gemm_batch.c" "" "gemm_batch" ${CBLAS_FLAG} "" "" false)
128+
GenerateNamedObjects("gemm_batch_strided.c" "" "gemm_batch_strided" ${CBLAS_FLAG} "" "" false)
128129
endif ()
129130
if (BUILD_DOUBLE)
130131
GenerateNamedObjects("dsdot.c" "" "dsdot" ${CBLAS_FLAG} "" "" true "SINGLE")
@@ -161,6 +162,7 @@ if (BUILD_BFLOAT16)
161162
GenerateNamedObjects("bf16to.c" "SINGLE_PREC" "sbf16tos" ${CBLAS_FLAG} "" "" true "BFLOAT16")
162163
GenerateNamedObjects("bf16to.c" "DOUBLE_PREC" "dbf16tod" ${CBLAS_FLAG} "" "" true "BFLOAT16")
163164
GenerateNamedObjects("gemm_batch.c" "" "sbgemm_batch" ${CBLAS_FLAG} "" "" true "BFLOAT16")
165+
GenerateNamedObjects("gemm_batch_strided.c" "" "sbgemm_batch_strided" ${CBLAS_FLAG} "" "" true "BFLOAT16")
164166
endif ()
165167
if (BUILD_HFLOAT16)
166168
GenerateNamedObjects("gemm.c" "" "shgemm" ${CBLAS_FLAG} "" "" true "HFLOAT16")
@@ -194,6 +196,7 @@ foreach (float_type ${FLOAT_TYPES})
194196
GenerateNamedObjects("asum.c" "" "scasum" ${CBLAS_FLAG} "" "" true "COMPLEX")
195197
GenerateNamedObjects("sum.c" "" "scsum" ${CBLAS_FLAG} "" "" true "COMPLEX")
196198
GenerateNamedObjects("gemm_batch.c" "" "cgemm_batch" ${CBLAS_FLAG} "" "" true "COMPLEX")
199+
GenerateNamedObjects("gemm_batch_strided.c" "" "cgemm_batch_strided" ${CBLAS_FLAG} "" "" true "COMPLEX")
197200
endif ()
198201
if (${float_type} STREQUAL "ZCOMPLEX")
199202
GenerateNamedObjects("zscal.c" "SSCAL" "dscal" ${CBLAS_FLAG} "" "" false "ZCOMPLEX")
@@ -204,6 +207,7 @@ foreach (float_type ${FLOAT_TYPES})
204207
GenerateNamedObjects("asum.c" "" "dzasum" ${CBLAS_FLAG} "" "" true "ZCOMPLEX")
205208
GenerateNamedObjects("sum.c" "" "dzsum" ${CBLAS_FLAG} "" "" true "ZCOMPLEX")
206209
GenerateNamedObjects("gemm_batch.c" "" "zgemm_batch" ${CBLAS_FLAG} "" "" true "ZCOMPLEX")
210+
GenerateNamedObjects("gemm_batch_strided.c" "" "zgemm_batch_strided" ${CBLAS_FLAG} "" "" true "ZCOMPLEX")
207211
endif ()
208212
endforeach ()
209213

@@ -255,6 +259,7 @@ if ( BUILD_COMPLEX AND NOT BUILD_SINGLE)
255259
GenerateNamedObjects("gemv.c" "" "gemv" 0 "" "" false "SINGLE")
256260
GenerateNamedObjects("gemm.c" "" "gemm" 0 "" "" false "SINGLE")
257261
GenerateNamedObjects("gemm_batch.c" "" "gemm_batch" 0 "" "" false "SINGLE")
262+
GenerateNamedObjects("gemm_batch_strided.c" "" "gemm_batch_strided" 0 "" "" false "SINGLE")
258263
GenerateNamedObjects("asum.c" "" "asum" 0 "" "" false "SINGLE")
259264
GenerateNamedObjects("swap.c" "" "swap" 0 "" "" false "SINGLE")
260265
GenerateNamedObjects("axpy.c" "" "axpy" 0 "" "" false "SINGLE")
@@ -269,6 +274,7 @@ if ( BUILD_COMPLEX16 AND NOT BUILD_DOUBLE)
269274
GenerateNamedObjects("gemv.c" "" "gemv" 0 "" "" false "DOUBLE")
270275
GenerateNamedObjects("gemm.c" "" "gemm" 0 "" "" false "DOUBLE")
271276
GenerateNamedObjects("gemm_batch.c" "" "gemm_batch" 0 "" "" false "DOUBLE")
277+
GenerateNamedObjects("gemm_batch_strided.c" "" "gemm_batch_strided" 0 "" "" false "DOUBLE")
272278
GenerateNamedObjects("asum.c" "" "asum" 0 "" "" false "DOUBLE")
273279
GenerateNamedObjects("swap.c" "" "swap" 0 "" "" false "DOUBLE")
274280
GenerateNamedObjects("axpy.c" "" "axpy" 0 "" "" false "DOUBLE")

interface/Makefile

Lines changed: 39 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -73,15 +73,15 @@ SBLAS3OBJS = \
7373
strsm.$(SUFFIX) ssyrk.$(SUFFIX) ssyr2k.$(SUFFIX) \
7474
somatcopy.$(SUFFIX) simatcopy.$(SUFFIX)\
7575
sgeadd.$(SUFFIX) sgemmt.$(SUFFIX) sgemmtr.$(SUFFIX) \
76-
sgemm_batch.$(SUFFIX)
76+
sgemm_batch.$(SUFFIX) sgemm_batch_strided.$(SUFFIX)
7777

7878
ifeq ($(BUILD_BFLOAT16),1)
7979
BBLAS3OBJS = bgemm.$(SUFFIX)
8080
BBLAS2OBJS = bgemv.$(SUFFIX)
8181
BBLAS1OBJS = bscal.$(SUFFIX)
8282
SBBLAS1OBJS = sbdot.$(SUFFIX)
8383
SBBLAS2OBJS = sbgemv.$(SUFFIX)
84-
SBBLAS3OBJS = sbgemm.$(SUFFIX) sbgemmt.$(SUFFIX) sbgemmtr.$(SUFFIX) sbgemm_batch.$(SUFFIX)
84+
SBBLAS3OBJS = sbgemm.$(SUFFIX) sbgemmt.$(SUFFIX) sbgemmtr.$(SUFFIX) sbgemm_batch.$(SUFFIX) sbgemm_batch_strided.$(SUFFIX)
8585
SBEXTOBJS = sbstobf16.$(SUFFIX) sbdtobf16.$(SUFFIX) sbf16tos.$(SUFFIX) dbf16tod.$(SUFFIX)
8686
endif
8787

@@ -113,7 +113,7 @@ DBLAS3OBJS = \
113113
dtrsm.$(SUFFIX) dsyrk.$(SUFFIX) dsyr2k.$(SUFFIX) \
114114
domatcopy.$(SUFFIX) dimatcopy.$(SUFFIX)\
115115
dgeadd.$(SUFFIX) dgemmt.$(SUFFIX) dgemmtr.$(SUFFIX) \
116-
dgemm_batch.$(SUFFIX)
116+
dgemm_batch.$(SUFFIX) dgemm_batch_strided.$(SUFFIX)
117117

118118
CBLAS1OBJS = \
119119
caxpy.$(SUFFIX) caxpyc.$(SUFFIX) cswap.$(SUFFIX) \
@@ -143,7 +143,7 @@ CBLAS3OBJS = \
143143
chemm.$(SUFFIX) cherk.$(SUFFIX) cher2k.$(SUFFIX) \
144144
comatcopy.$(SUFFIX) cimatcopy.$(SUFFIX)\
145145
cgeadd.$(SUFFIX) cgemmt.$(SUFFIX) cgemmtr.$(SUFFIX) \
146-
cgemm_batch.$(SUFFIX)
146+
cgemm_batch.$(SUFFIX) cgemm_batch_strided.$(SUFFIX)
147147

148148
ZBLAS1OBJS = \
149149
zaxpy.$(SUFFIX) zaxpyc.$(SUFFIX) zswap.$(SUFFIX) \
@@ -173,7 +173,7 @@ ZBLAS3OBJS = \
173173
zhemm.$(SUFFIX) zherk.$(SUFFIX) zher2k.$(SUFFIX) \
174174
zomatcopy.$(SUFFIX) zimatcopy.$(SUFFIX)\
175175
zgeadd.$(SUFFIX) zgemmt.$(SUFFIX) zgemmtr.$(SUFFIX) \
176-
zgemm_batch.$(SUFFIX)
176+
zgemm_batch.$(SUFFIX) zgemm_batch_strided.$(SUFFIX)
177177

178178
ifeq ($(SUPPORT_GEMM3M), 1)
179179

@@ -321,15 +321,15 @@ CSBLAS2OBJS = \
321321
CSBLAS3OBJS = \
322322
cblas_sgemm.$(SUFFIX) cblas_ssymm.$(SUFFIX) cblas_strmm.$(SUFFIX) cblas_strsm.$(SUFFIX) \
323323
cblas_ssyrk.$(SUFFIX) cblas_ssyr2k.$(SUFFIX) cblas_somatcopy.$(SUFFIX) cblas_simatcopy.$(SUFFIX)\
324-
cblas_sgeadd.$(SUFFIX) cblas_sgemmt.$(SUFFIX) cblas_sgemmtr.$(SUFFIX) cblas_sgemm_batch.$(SUFFIX)
324+
cblas_sgeadd.$(SUFFIX) cblas_sgemmt.$(SUFFIX) cblas_sgemmtr.$(SUFFIX) cblas_sgemm_batch.$(SUFFIX) cblas_sgemm_batch_strided.$(SUFFIX)
325325

326326
ifeq ($(BUILD_BFLOAT16),1)
327327
CBBLAS3OBJS = cblas_bgemm.$(SUFFIX)
328328
CBBLAS2OBJS = cblas_bgemv.$(SUFFIX)
329329
CBBLAS1OBJS = cblas_bscal.$(SUFFIX)
330330
CSBBLAS1OBJS = cblas_sbdot.$(SUFFIX)
331331
CSBBLAS2OBJS = cblas_sbgemv.$(SUFFIX)
332-
CSBBLAS3OBJS = cblas_sbgemm.$(SUFFIX) cblas_sbgemmt.$(SUFFIX) cblas_sbgemmtr.$(SUFFIX) cblas_sbgemm_batch.$(SUFFIX)
332+
CSBBLAS3OBJS = cblas_sbgemm.$(SUFFIX) cblas_sbgemmt.$(SUFFIX) cblas_sbgemmtr.$(SUFFIX) cblas_sbgemm_batch.$(SUFFIX) cblas_sbgemm_batch_strided.$(SUFFIX)
333333
CSBEXTOBJS = cblas_sbstobf16.$(SUFFIX) cblas_sbdtobf16.$(SUFFIX) cblas_sbf16tos.$(SUFFIX) cblas_dbf16tod.$(SUFFIX)
334334
ifeq ($(ONLY_CBLAS),1)
335335
CSBEXTOBJS += sbstobf16.$(SUFFIX) sbdtobf16.$(SUFFIX) sbf16tos.$(SUFFIX) dbf16tod.$(SUFFIX)
@@ -357,7 +357,7 @@ CDBLAS2OBJS = \
357357
CDBLAS3OBJS += \
358358
cblas_dgemm.$(SUFFIX) cblas_dsymm.$(SUFFIX) cblas_dtrmm.$(SUFFIX) cblas_dtrsm.$(SUFFIX) \
359359
cblas_dsyrk.$(SUFFIX) cblas_dsyr2k.$(SUFFIX) cblas_domatcopy.$(SUFFIX) cblas_dimatcopy.$(SUFFIX) \
360-
cblas_dgeadd.$(SUFFIX) cblas_dgemmt.$(SUFFIX) cblas_dgemmtr.$(SUFFIX) cblas_dgemm_batch.$(SUFFIX)
360+
cblas_dgeadd.$(SUFFIX) cblas_dgemmt.$(SUFFIX) cblas_dgemmtr.$(SUFFIX) cblas_dgemm_batch.$(SUFFIX) cblas_dgemm_batch_strided.$(SUFFIX)
361361

362362
CCBLAS1OBJS = \
363363
cblas_icamax.$(SUFFIX) cblas_icamin.$(SUFFIX) cblas_scasum.$(SUFFIX) cblas_caxpy.$(SUFFIX) \
@@ -382,7 +382,7 @@ CCBLAS3OBJS = \
382382
cblas_csyrk.$(SUFFIX) cblas_csyr2k.$(SUFFIX) \
383383
cblas_chemm.$(SUFFIX) cblas_cherk.$(SUFFIX) cblas_cher2k.$(SUFFIX) \
384384
cblas_comatcopy.$(SUFFIX) cblas_cimatcopy.$(SUFFIX)\
385-
cblas_cgeadd.$(SUFFIX) cblas_cgemmt.$(SUFFIX) cblas_cgemmtr.$(SUFFIX) cblas_cgemm_batch.$(SUFFIX)
385+
cblas_cgeadd.$(SUFFIX) cblas_cgemmt.$(SUFFIX) cblas_cgemmtr.$(SUFFIX) cblas_cgemm_batch.$(SUFFIX) cblas_cgemm_batch_strided.$(SUFFIX)
386386

387387
CXERBLAOBJ = \
388388
cblas_xerbla.$(SUFFIX)
@@ -413,7 +413,7 @@ CZBLAS3OBJS = \
413413
cblas_zsyrk.$(SUFFIX) cblas_zsyr2k.$(SUFFIX) \
414414
cblas_zhemm.$(SUFFIX) cblas_zherk.$(SUFFIX) cblas_zher2k.$(SUFFIX)\
415415
cblas_zomatcopy.$(SUFFIX) cblas_zimatcopy.$(SUFFIX) \
416-
cblas_zgeadd.$(SUFFIX) cblas_zgemmt.$(SUFFIX) cblas_zgemmtr.$(SUFFIX) cblas_zgemm_batch.$(SUFFIX)
416+
cblas_zgeadd.$(SUFFIX) cblas_zgemmt.$(SUFFIX) cblas_zgemmtr.$(SUFFIX) cblas_zgemm_batch.$(SUFFIX) cblas_zgemm_batch_strided.$(SUFFIX)
417417

418418

419419
ifeq ($(SUPPORT_GEMM3M), 1)
@@ -2544,6 +2544,21 @@ cblas_cgemm_batch.$(SUFFIX) cblas_cgemm_batch.$(PSUFFIX) : gemm_batch.c ../param
25442544
cblas_zgemm_batch.$(SUFFIX) cblas_zgemm_batch.$(PSUFFIX) : gemm_batch.c ../param.h
25452545
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
25462546

2547+
cblas_sbgemm_batch_strided.$(SUFFIX) cblas_sbgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2548+
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
2549+
2550+
cblas_sgemm_batch_strided.$(SUFFIX) cblas_sgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2551+
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
2552+
2553+
cblas_dgemm_batch_strided.$(SUFFIX) cblas_dgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2554+
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
2555+
2556+
cblas_cgemm_batch_strided.$(SUFFIX) cblas_cgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2557+
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
2558+
2559+
cblas_zgemm_batch_strided.$(SUFFIX) cblas_zgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2560+
$(CC) -c $(CFLAGS) -DCBLAS $< -o $(@F)
2561+
25472562
sbgemm_batch.$(SUFFIX) sbgemm_batch.$(PSUFFIX) : gemm_batch.c ../param.h
25482563
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
25492564

@@ -2559,3 +2574,17 @@ cgemm_batch.$(SUFFIX) cgemm_batch.$(PSUFFIX) : gemm_batch.c ../param.h
25592574
zgemm_batch.$(SUFFIX) zgemm_batch.$(PSUFFIX) : gemm_batch.c ../param.h
25602575
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
25612576

2577+
sbgemm_batch_strided.$(SUFFIX) sbgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2578+
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
2579+
2580+
sgemm_batch_strided.$(SUFFIX) sgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2581+
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
2582+
2583+
dgemm_batch_strided.$(SUFFIX) dgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2584+
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
2585+
2586+
cgemm_batch_strided.$(SUFFIX) cgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2587+
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)
2588+
2589+
zgemm_batch_strided.$(SUFFIX) zgemm_batch_strided.$(PSUFFIX) : gemm_batch_strided.c ../param.h
2590+
$(CC) -c $(CFLAGS) -UCBLAS $< -o $(@F)

0 commit comments

Comments
 (0)