@@ -456,6 +456,14 @@ void cblas_cgemm_batch(OPENBLAS_CONST enum CBLAS_ORDER Order, OPENBLAS_CONST enu
456456void 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 */
461469void 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
477485void 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 ***/
481490void 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 );
0 commit comments