libflame
12600
|
Functions | |
FLA_Error | FLA_Gemv_external (FLA_Trans transa, FLA_Obj alpha, FLA_Obj A, FLA_Obj x, FLA_Obj beta, FLA_Obj y) |
FLA_Error FLA_Gemv_external | ( | FLA_Trans | transa, |
FLA_Obj | alpha, | ||
FLA_Obj | A, | ||
FLA_Obj | x, | ||
FLA_Obj | beta, | ||
FLA_Obj | y | ||
) |
References bl1_cgemv(), bl1_dgemv(), bl1_sgemv(), bl1_zgemv(), FLA_Check_error_level(), FLA_Gemv_check(), FLA_Obj_col_stride(), FLA_Obj_datatype(), FLA_Obj_has_zero_dim(), FLA_Obj_length(), FLA_Obj_row_stride(), FLA_Obj_vector_inc(), FLA_Obj_width(), FLA_Param_map_flame_to_blis_conj(), FLA_Param_map_flame_to_blis_trans(), and FLA_Scal_external().
Referenced by FLA_Accum_T_UT_fc_unb_var1(), FLA_Apply_H2_UT_r_unb_var1(), FLA_CAQR2_UT_unb_var1(), FLA_Gemm_cn_unb_var3(), FLA_Gemm_cn_unb_var4(), FLA_Gemm_ct_unb_var3(), FLA_Gemm_ct_unb_var4(), FLA_Gemm_hn_unb_var3(), FLA_Gemm_hn_unb_var4(), FLA_Gemm_ht_unb_var3(), FLA_Gemm_ht_unb_var4(), FLA_Gemm_nc_unb_var1(), FLA_Gemm_nc_unb_var2(), FLA_Gemm_nh_unb_var1(), FLA_Gemm_nh_unb_var2(), FLA_Gemm_nn_unb_var1(), FLA_Gemm_nn_unb_var2(), FLA_Gemm_nn_unb_var3(), FLA_Gemm_nn_unb_var4(), FLA_Gemm_nt_unb_var1(), FLA_Gemm_nt_unb_var2(), FLA_Gemm_nt_unb_var3(), FLA_Gemm_nt_unb_var4(), FLA_Gemm_tc_unb_var1(), FLA_Gemm_tc_unb_var2(), FLA_Gemm_th_unb_var1(), FLA_Gemm_th_unb_var2(), FLA_Gemm_tn_unb_var1(), FLA_Gemm_tn_unb_var2(), FLA_Gemm_tn_unb_var3(), FLA_Gemm_tn_unb_var4(), FLA_Gemm_tt_unb_var1(), FLA_Gemm_tt_unb_var2(), FLA_Gemm_tt_unb_var3(), FLA_Gemm_tt_unb_var4(), FLA_Gemv(), FLA_Gemv_h_task(), FLA_Gemv_n_task(), FLA_Gemv_t_task(), FLA_Gemv_task(), FLA_Hemm_ll_unb_var1(), FLA_Hemm_ll_unb_var2(), FLA_Hemm_ll_unb_var6(), FLA_Hemm_ll_unb_var7(), FLA_Hemm_lu_unb_var2(), FLA_Hemm_lu_unb_var3(), FLA_Hemm_lu_unb_var5(), FLA_Hemm_lu_unb_var6(), FLA_Hemm_rl_unb_var2(), FLA_Hemm_rl_unb_var3(), FLA_Hemm_rl_unb_var5(), FLA_Hemm_rl_unb_var6(), FLA_Hemm_ru_unb_var1(), FLA_Hemm_ru_unb_var2(), FLA_Hemm_ru_unb_var6(), FLA_Hemm_ru_unb_var7(), FLA_Her2k_lh_unb_var2(), FLA_Her2k_lh_unb_var3(), FLA_Her2k_lh_unb_var4(), FLA_Her2k_lh_unb_var5(), FLA_Her2k_lh_unb_var6(), FLA_Her2k_lh_unb_var7(), FLA_Her2k_ln_unb_var1(), FLA_Her2k_ln_unb_var2(), FLA_Her2k_ln_unb_var3(), FLA_Her2k_ln_unb_var6(), FLA_Her2k_ln_unb_var7(), FLA_Her2k_ln_unb_var8(), FLA_Her2k_uh_unb_var2(), FLA_Her2k_uh_unb_var3(), FLA_Her2k_uh_unb_var4(), FLA_Her2k_uh_unb_var5(), FLA_Her2k_uh_unb_var6(), FLA_Her2k_uh_unb_var7(), FLA_Her2k_un_unb_var1(), FLA_Her2k_un_unb_var2(), FLA_Her2k_un_unb_var3(), FLA_Her2k_un_unb_var6(), FLA_Her2k_un_unb_var7(), FLA_Her2k_un_unb_var8(), FLA_Herk_lh_unb_var2(), FLA_Herk_lh_unb_var3(), FLA_Herk_ln_unb_var4(), FLA_Herk_uh_unb_var2(), FLA_Herk_uh_unb_var4(), FLA_Herk_un_unb_var1(), FLA_Herk_un_unb_var3(), FLA_LU_nopiv_unb_var2(), FLA_LU_nopiv_unb_var3(), FLA_LU_nopiv_unb_var4(), FLA_LU_piv_unb_var3(), FLA_LU_piv_unb_var3b(), FLA_LU_piv_unb_var4(), FLA_QR2_UT_unb_var1(), FLA_QR_UT_piv_unb_var1(), FLA_QR_UT_piv_unb_var2(), FLA_QR_UT_unb_var2(), FLA_Symm_ll_unb_var1(), FLA_Symm_ll_unb_var2(), FLA_Symm_ll_unb_var3(), FLA_Symm_ll_unb_var5(), FLA_Symm_ll_unb_var6(), FLA_Symm_ll_unb_var7(), FLA_Symm_lu_unb_var1(), FLA_Symm_lu_unb_var2(), FLA_Symm_lu_unb_var3(), FLA_Symm_lu_unb_var5(), FLA_Symm_lu_unb_var6(), FLA_Symm_lu_unb_var7(), FLA_Symm_rl_unb_var1(), FLA_Symm_rl_unb_var2(), FLA_Symm_rl_unb_var3(), FLA_Symm_rl_unb_var5(), FLA_Symm_rl_unb_var6(), FLA_Symm_rl_unb_var7(), FLA_Symm_ru_unb_var1(), FLA_Symm_ru_unb_var2(), FLA_Symm_ru_unb_var3(), FLA_Symm_ru_unb_var5(), FLA_Symm_ru_unb_var6(), FLA_Symm_ru_unb_var7(), FLA_Syr2k_ln_unb_var1(), FLA_Syr2k_ln_unb_var2(), FLA_Syr2k_ln_unb_var3(), FLA_Syr2k_ln_unb_var4(), FLA_Syr2k_ln_unb_var5(), FLA_Syr2k_ln_unb_var6(), FLA_Syr2k_ln_unb_var7(), FLA_Syr2k_ln_unb_var8(), FLA_Syr2k_lt_unb_var1(), FLA_Syr2k_lt_unb_var2(), FLA_Syr2k_lt_unb_var3(), FLA_Syr2k_lt_unb_var4(), FLA_Syr2k_lt_unb_var5(), FLA_Syr2k_lt_unb_var6(), FLA_Syr2k_lt_unb_var7(), FLA_Syr2k_lt_unb_var8(), FLA_Syr2k_un_unb_var1(), FLA_Syr2k_un_unb_var2(), FLA_Syr2k_un_unb_var3(), FLA_Syr2k_un_unb_var4(), FLA_Syr2k_un_unb_var5(), FLA_Syr2k_un_unb_var6(), FLA_Syr2k_un_unb_var7(), FLA_Syr2k_un_unb_var8(), FLA_Syr2k_ut_unb_var1(), FLA_Syr2k_ut_unb_var2(), FLA_Syr2k_ut_unb_var3(), FLA_Syr2k_ut_unb_var4(), FLA_Syr2k_ut_unb_var5(), FLA_Syr2k_ut_unb_var6(), FLA_Syr2k_ut_unb_var7(), FLA_Syr2k_ut_unb_var8(), FLA_Syrk_ln_unb_var1(), FLA_Syrk_ln_unb_var2(), FLA_Syrk_ln_unb_var3(), FLA_Syrk_ln_unb_var4(), FLA_Syrk_lt_unb_var1(), FLA_Syrk_lt_unb_var2(), FLA_Syrk_lt_unb_var3(), FLA_Syrk_lt_unb_var4(), FLA_Syrk_un_unb_var1(), FLA_Syrk_un_unb_var2(), FLA_Syrk_un_unb_var3(), FLA_Syrk_un_unb_var4(), FLA_Syrk_ut_unb_var1(), FLA_Syrk_ut_unb_var2(), FLA_Syrk_ut_unb_var3(), FLA_Syrk_ut_unb_var4(), FLA_Trmm_lln_unb_var1(), FLA_Trmm_llt_unb_var1(), FLA_Trmm_lun_unb_var1(), FLA_Trmm_lut_unb_var1(), FLA_Trmm_rln_unb_var1(), FLA_Trmm_rlt_unb_var1(), FLA_Trmm_run_unb_var1(), FLA_Trmm_rut_unb_var1(), FLA_Trsm_lln_unb_var1(), FLA_Trsm_llt_unb_var1(), FLA_Trsm_lun_unb_var1(), FLA_Trsm_lut_unb_var1(), FLA_Trsm_rln_unb_var1(), FLA_Trsm_rlt_unb_var1(), FLA_Trsm_run_unb_var1(), and FLA_Trsm_rut_unb_var1().
{ FLA_Datatype datatype; int m_A, n_A; int rs_A, cs_A; int inc_x; int inc_y; trans1_t blis_transa; conj1_t blis_conjx; if ( FLA_Check_error_level() == FLA_FULL_ERROR_CHECKING ) FLA_Gemv_check( transa, alpha, A, x, beta, y ); if ( FLA_Obj_has_zero_dim( A ) ) { FLA_Scal_external( beta, y ); return FLA_SUCCESS; } datatype = FLA_Obj_datatype( A ); m_A = FLA_Obj_length( A ); n_A = FLA_Obj_width( A ); rs_A = FLA_Obj_row_stride( A ); cs_A = FLA_Obj_col_stride( A ); inc_x = FLA_Obj_vector_inc( x ); inc_y = FLA_Obj_vector_inc( y ); FLA_Param_map_flame_to_blis_trans( transa, &blis_transa ); FLA_Param_map_flame_to_blis_conj( FLA_NO_CONJUGATE, &blis_conjx ); switch( datatype ){ case FLA_FLOAT: { float *buff_A = ( float * ) FLA_FLOAT_PTR( A ); float *buff_x = ( float * ) FLA_FLOAT_PTR( x ); float *buff_y = ( float * ) FLA_FLOAT_PTR( y ); float *buff_alpha = ( float * ) FLA_FLOAT_PTR( alpha ); float *buff_beta = ( float * ) FLA_FLOAT_PTR( beta ); bl1_sgemv( blis_transa, blis_conjx, m_A, n_A, buff_alpha, buff_A, rs_A, cs_A, buff_x, inc_x, buff_beta, buff_y, inc_y ); break; } case FLA_DOUBLE: { double *buff_A = ( double * ) FLA_DOUBLE_PTR( A ); double *buff_x = ( double * ) FLA_DOUBLE_PTR( x ); double *buff_y = ( double * ) FLA_DOUBLE_PTR( y ); double *buff_alpha = ( double * ) FLA_DOUBLE_PTR( alpha ); double *buff_beta = ( double * ) FLA_DOUBLE_PTR( beta ); bl1_dgemv( blis_transa, blis_conjx, m_A, n_A, buff_alpha, buff_A, rs_A, cs_A, buff_x, inc_x, buff_beta, buff_y, inc_y ); break; } case FLA_COMPLEX: { scomplex *buff_A = ( scomplex * ) FLA_COMPLEX_PTR( A ); scomplex *buff_x = ( scomplex * ) FLA_COMPLEX_PTR( x ); scomplex *buff_y = ( scomplex * ) FLA_COMPLEX_PTR( y ); scomplex *buff_alpha = ( scomplex * ) FLA_COMPLEX_PTR( alpha ); scomplex *buff_beta = ( scomplex * ) FLA_COMPLEX_PTR( beta ); bl1_cgemv( blis_transa, blis_conjx, m_A, n_A, buff_alpha, buff_A, rs_A, cs_A, buff_x, inc_x, buff_beta, buff_y, inc_y ); break; } case FLA_DOUBLE_COMPLEX: { dcomplex *buff_A = ( dcomplex * ) FLA_DOUBLE_COMPLEX_PTR( A ); dcomplex *buff_x = ( dcomplex * ) FLA_DOUBLE_COMPLEX_PTR( x ); dcomplex *buff_y = ( dcomplex * ) FLA_DOUBLE_COMPLEX_PTR( y ); dcomplex *buff_alpha = ( dcomplex * ) FLA_DOUBLE_COMPLEX_PTR( alpha ); dcomplex *buff_beta = ( dcomplex * ) FLA_DOUBLE_COMPLEX_PTR( beta ); bl1_zgemv( blis_transa, blis_conjx, m_A, n_A, buff_alpha, buff_A, rs_A, cs_A, buff_x, inc_x, buff_beta, buff_y, inc_y ); break; } } return FLA_SUCCESS; }