@@ -72,15 +72,16 @@ SBLAS3OBJS = \
7272 sgemm.$(SUFFIX ) ssymm.$(SUFFIX ) strmm.$(SUFFIX ) \
7373 strsm.$(SUFFIX ) ssyrk.$(SUFFIX ) ssyr2k.$(SUFFIX ) \
7474 somatcopy.$(SUFFIX ) simatcopy.$(SUFFIX ) \
75- sgeadd.$(SUFFIX ) sgemmt.$(SUFFIX ) sgemmtr.$(SUFFIX )
75+ sgeadd.$(SUFFIX ) sgemmt.$(SUFFIX ) sgemmtr.$(SUFFIX ) \
76+ sgemm_batch.$(SUFFIX ) sgemm_batch_strided.$(SUFFIX )
7677
7778ifeq ($(BUILD_BFLOAT16 ) ,1)
7879BBLAS3OBJS = bgemm.$(SUFFIX )
7980BBLAS2OBJS = bgemv.$(SUFFIX )
8081BBLAS1OBJS = bscal.$(SUFFIX )
8182SBBLAS1OBJS = sbdot.$(SUFFIX )
8283SBBLAS2OBJS = sbgemv.$(SUFFIX )
83- SBBLAS3OBJS = sbgemm.$(SUFFIX ) sbgemmt.$(SUFFIX ) sbgemmtr.$(SUFFIX )
84+ SBBLAS3OBJS = sbgemm.$(SUFFIX ) sbgemmt.$(SUFFIX ) sbgemmtr.$(SUFFIX ) sbgemm_batch. $( SUFFIX ) sbgemm_batch_strided. $( SUFFIX )
8485SBEXTOBJS = sbstobf16.$(SUFFIX ) sbdtobf16.$(SUFFIX ) sbf16tos.$(SUFFIX ) dbf16tod.$(SUFFIX )
8586endif
8687
@@ -111,7 +112,8 @@ DBLAS3OBJS = \
111112 dgemm.$(SUFFIX ) dsymm.$(SUFFIX ) dtrmm.$(SUFFIX ) \
112113 dtrsm.$(SUFFIX ) dsyrk.$(SUFFIX ) dsyr2k.$(SUFFIX ) \
113114 domatcopy.$(SUFFIX ) dimatcopy.$(SUFFIX ) \
114- dgeadd.$(SUFFIX ) dgemmt.$(SUFFIX ) dgemmtr.$(SUFFIX )
115+ dgeadd.$(SUFFIX ) dgemmt.$(SUFFIX ) dgemmtr.$(SUFFIX ) \
116+ dgemm_batch.$(SUFFIX ) dgemm_batch_strided.$(SUFFIX )
115117
116118CBLAS1OBJS = \
117119 caxpy.$(SUFFIX ) caxpyc.$(SUFFIX ) cswap.$(SUFFIX ) \
@@ -140,7 +142,8 @@ CBLAS3OBJS = \
140142 ctrsm.$(SUFFIX ) csyrk.$(SUFFIX ) csyr2k.$(SUFFIX ) \
141143 chemm.$(SUFFIX ) cherk.$(SUFFIX ) cher2k.$(SUFFIX ) \
142144 comatcopy.$(SUFFIX ) cimatcopy.$(SUFFIX ) \
143- cgeadd.$(SUFFIX ) cgemmt.$(SUFFIX ) cgemmtr.$(SUFFIX )
145+ cgeadd.$(SUFFIX ) cgemmt.$(SUFFIX ) cgemmtr.$(SUFFIX ) \
146+ cgemm_batch.$(SUFFIX ) cgemm_batch_strided.$(SUFFIX )
144147
145148ZBLAS1OBJS = \
146149 zaxpy.$(SUFFIX ) zaxpyc.$(SUFFIX ) zswap.$(SUFFIX ) \
@@ -169,7 +172,8 @@ ZBLAS3OBJS = \
169172 ztrsm.$(SUFFIX ) zsyrk.$(SUFFIX ) zsyr2k.$(SUFFIX ) \
170173 zhemm.$(SUFFIX ) zherk.$(SUFFIX ) zher2k.$(SUFFIX ) \
171174 zomatcopy.$(SUFFIX ) zimatcopy.$(SUFFIX ) \
172- zgeadd.$(SUFFIX ) zgemmt.$(SUFFIX ) zgemmtr.$(SUFFIX )
175+ zgeadd.$(SUFFIX ) zgemmt.$(SUFFIX ) zgemmtr.$(SUFFIX ) \
176+ zgemm_batch.$(SUFFIX ) zgemm_batch_strided.$(SUFFIX )
173177
174178ifeq ($(SUPPORT_GEMM3M ) , 1)
175179
@@ -317,15 +321,15 @@ CSBLAS2OBJS = \
317321CSBLAS3OBJS = \
318322 cblas_sgemm.$(SUFFIX ) cblas_ssymm.$(SUFFIX ) cblas_strmm.$(SUFFIX ) cblas_strsm.$(SUFFIX ) \
319323 cblas_ssyrk.$(SUFFIX ) cblas_ssyr2k.$(SUFFIX ) cblas_somatcopy.$(SUFFIX ) cblas_simatcopy.$(SUFFIX ) \
320- 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. $( SUFFIX )
321325
322326ifeq ($(BUILD_BFLOAT16 ) ,1)
323327CBBLAS3OBJS = cblas_bgemm.$(SUFFIX )
324328CBBLAS2OBJS = cblas_bgemv.$(SUFFIX )
325329CBBLAS1OBJS = cblas_bscal.$(SUFFIX )
326330CSBBLAS1OBJS = cblas_sbdot.$(SUFFIX )
327331CSBBLAS2OBJS = cblas_sbgemv.$(SUFFIX )
328- CSBBLAS3OBJS = cblas_sbgemm.$(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. $( SUFFIX )
329333CSBEXTOBJS = cblas_sbstobf16.$(SUFFIX ) cblas_sbdtobf16.$(SUFFIX ) cblas_sbf16tos.$(SUFFIX ) cblas_dbf16tod.$(SUFFIX )
330334ifeq ($(ONLY_CBLAS ) ,1)
331335CSBEXTOBJS += sbstobf16.$(SUFFIX ) sbdtobf16.$(SUFFIX ) sbf16tos.$(SUFFIX ) dbf16tod.$(SUFFIX )
@@ -353,7 +357,7 @@ CDBLAS2OBJS = \
353357CDBLAS3OBJS += \
354358 cblas_dgemm.$(SUFFIX ) cblas_dsymm.$(SUFFIX ) cblas_dtrmm.$(SUFFIX ) cblas_dtrsm.$(SUFFIX ) \
355359 cblas_dsyrk.$(SUFFIX ) cblas_dsyr2k.$(SUFFIX ) cblas_domatcopy.$(SUFFIX ) cblas_dimatcopy.$(SUFFIX ) \
356- 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. $( SUFFIX )
357361
358362CCBLAS1OBJS = \
359363 cblas_icamax.$(SUFFIX ) cblas_icamin.$(SUFFIX ) cblas_scasum.$(SUFFIX ) cblas_caxpy.$(SUFFIX ) \
@@ -378,7 +382,7 @@ CCBLAS3OBJS = \
378382 cblas_csyrk.$(SUFFIX ) cblas_csyr2k.$(SUFFIX ) \
379383 cblas_chemm.$(SUFFIX ) cblas_cherk.$(SUFFIX ) cblas_cher2k.$(SUFFIX ) \
380384 cblas_comatcopy.$(SUFFIX ) cblas_cimatcopy.$(SUFFIX ) \
381- 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. $( SUFFIX )
382386
383387CXERBLAOBJ = \
384388 cblas_xerbla.$(SUFFIX )
@@ -409,7 +413,7 @@ CZBLAS3OBJS = \
409413 cblas_zsyrk.$(SUFFIX ) cblas_zsyr2k.$(SUFFIX ) \
410414 cblas_zhemm.$(SUFFIX ) cblas_zherk.$(SUFFIX ) cblas_zher2k.$(SUFFIX ) \
411415 cblas_zomatcopy.$(SUFFIX ) cblas_zimatcopy.$(SUFFIX ) \
412- cblas_zgeadd.$(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. $( SUFFIX )
413417
414418
415419ifeq ($(SUPPORT_GEMM3M ) , 1)
@@ -2539,3 +2543,48 @@ cblas_cgemm_batch.$(SUFFIX) cblas_cgemm_batch.$(PSUFFIX) : gemm_batch.c ../param
25392543
25402544cblas_zgemm_batch.$(SUFFIX ) cblas_zgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
25412545 $(CC ) -c $(CFLAGS ) -DCBLAS $< -o $(@F )
2546+
2547+ cblas_sbgemm_batch_strided.$(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 $(@F )
2561+
2562+ sbgemm_batch.$(SUFFIX ) sbgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
2563+ $(CC ) -c $(CFLAGS ) -UCBLAS $< -o $(@F )
2564+
2565+ sgemm_batch.$(SUFFIX ) sgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
2566+ $(CC ) -c $(CFLAGS ) -UCBLAS $< -o $(@F )
2567+
2568+ dgemm_batch.$(SUFFIX ) dgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
2569+ $(CC ) -c $(CFLAGS ) -UCBLAS $< -o $(@F )
2570+
2571+ cgemm_batch.$(SUFFIX ) cgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
2572+ $(CC ) -c $(CFLAGS ) -UCBLAS $< -o $(@F )
2573+
2574+ zgemm_batch.$(SUFFIX ) zgemm_batch.$(PSUFFIX ) : gemm_batch.c ../param.h
2575+ $(CC ) -c $(CFLAGS ) -UCBLAS $< -o $(@F )
2576+
2577+ sbgemm_batch_strided.$(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