本文整理汇总了C++中FLA_Part_1x2函数的典型用法代码示例。如果您正苦于以下问题：C++ FLA_Part_1x2函数的具体用法？C++ FLA_Part_1x2怎么用？C++ FLA_Part_1x2使用的例子？那么恭喜您, 这里精选的函数代码示例或许可以为您提供帮助。
在下文中一共展示了FLA_Part_1x2函数的15个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的C++代码示例。
示例1: Symm_ru_blk_var6
int Symm_ru_blk_var6( FLA_Obj A, FLA_Obj B, FLA_Obj C, int nb_alg )
{
FLA_Obj ATL, ATR, A00, A01, A02,
ABL, ABR, A10, A11, A12,
A20, A21, A22;
FLA_Obj BL, BR, B0, B1, B2;
FLA_Obj CL, CR, C0, C1, C2;
int b;
FLA_Part_2x2( A, &ATL, &ATR,
&ABL, &ABR, 0, 0, FLA_BR );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
FLA_Part_1x2( C, &CL, &CR, 0, FLA_RIGHT );
while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
b = min( FLA_Obj_length( ATL ), nb_alg );
FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &A01, /**/ &A02,
&A10, &A11, /**/ &A12,
/* ************* */ /* ******************** */
ABL, /**/ ABR, &A20, &A21, /**/ &A22,
b, b, FLA_TL );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &B1, /**/ &B2,
b, FLA_LEFT );
FLA_Repart_1x2_to_1x3( CL, /**/ CR, &C0, &C1, /**/ &C2,
b, FLA_LEFT );
/*------------------------------------------------------------*/
/*C1 = B0 * A01 + B1 * A11 + B2 * A12' + C1;*/
FLA_Gemm(FLA_NO_TRANSPOSE, FLA_NO_TRANSPOSE, FLA_ONE, B0, A01, FLA_ONE, C1);
FLA_Gemm(FLA_NO_TRANSPOSE, FLA_NO_TRANSPOSE, FLA_ONE, B1, A11, FLA_ONE, C1);
FLA_Gemm(FLA_NO_TRANSPOSE, FLA_TRANSPOSE, FLA_ONE, B2, A12, FLA_ONE, C1);
/*------------------------------------------------------------*/
FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ A01, A02,
/* ************** */ /* ****************** */
A10, /**/ A11, A12,
&ABL, /**/ &ABR, A20, /**/ A21, A22,
FLA_BR );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ B1, B2,
FLA_RIGHT );
FLA_Cont_with_1x3_to_1x2( &CL, /**/ &CR, C0, /**/ C1, C2,
FLA_RIGHT );
}
return FLA_SUCCESS;
}示例2: FLA_Axpy_sync_circular
FLA_Error FLA_Axpy_sync_circular( FLA_Obj alpha, FLA_Obj X, FLA_Obj B )
{
FLA_Obj XL, XR, X0, X1, X2;
FLA_Obj BL, BR, B0, B1, B2;
int n_stages = FLA_omp_get_num_stages();
int stage_width = FLA_omp_compute_stage_width( X );
int thread_num = omp_get_thread_num();
int n_done = 0;
int b, i;
// Start thread i on the ith panel partition of B.
FLA_Part_1x2( X, &XL, &XR, stage_width*thread_num, FLA_LEFT );
FLA_Part_1x2( B, &BL, &BR, stage_width*thread_num, FLA_LEFT );
while ( n_done++ < n_stages ){
// The last lockable partition may be smaller than the others.
b = min( FLA_Obj_width( XR ), stage_width );
FLA_Repart_1x2_to_1x3( XL, /**/ XR, &X0, /**/ &X1, &X2,
b, FLA_RIGHT );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
b, FLA_RIGHT );
/*------------------------------------------------------------*/
// Get the index of the current partition.
i = FLA_Obj_width(XL)/stage_width;
// Acquire lock[i] (the lock for X1 and B1).
omp_set_lock( &fla_omp_lock[i] );
// B1 := alpha * X1 + B1
FLA_Axpy_external( alpha, X1, B1 );
// Release lock[i] (the lock for X1 and B1).
omp_unset_lock( &fla_omp_lock[i] );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &XL, /**/ &XR, X0, X1, /**/ X2,
FLA_LEFT );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
FLA_LEFT );
// If this thread reaches the last partition, wrap back around to
// the first partition for the next iteration.
if( FLA_Obj_width( XL ) == FLA_Obj_width( X ) )
{
FLA_Part_1x2( X, &XL, &XR, 0, FLA_LEFT );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
}
}
return FLA_SUCCESS;
}示例3: Gemm_unb_var2
int Gemm_unb_var2( FLA_Obj A, FLA_Obj B, FLA_Obj C )
{
FLA_Obj BL, BR, B0, b1, B2;
FLA_Obj CL, CR, C0, c1, C2;
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
FLA_Part_1x2( C, &CL, &CR, 0, FLA_LEFT );
while ( FLA_Obj_width( BL ) < FLA_Obj_width( B ) ){
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
1, FLA_RIGHT );
FLA_Repart_1x2_to_1x3( CL, /**/ CR, &C0, /**/ &c1, &C2,
1, FLA_RIGHT );
/*------------------------------------------------------------*/
FLA_Gemv( FLA_NO_TRANSPOSE, FLA_ONE, A, b1, FLA_ONE, c1 );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
FLA_LEFT );
FLA_Cont_with_1x3_to_1x2( &CL, /**/ &CR, C0, c1, /**/ C2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例4: FLA_Symm_lu_unb_var10
FLA_Error FLA_Symm_lu_unb_var10( FLA_Obj alpha, FLA_Obj A, FLA_Obj B, FLA_Obj beta, FLA_Obj C )
{
FLA_Obj BL, BR, B0, b1t, B2;
FLA_Obj CL, CR, C0, c1t, C2;
FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
FLA_Part_1x2( C, &CL, &CR, 0, FLA_RIGHT );
while ( FLA_Obj_width( BR ) < FLA_Obj_width( B ) ){
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &b1t, /**/ &B2,
1, FLA_LEFT );
FLA_Repart_1x2_to_1x3( CL, /**/ CR, &C0, &c1t, /**/ &C2,
1, FLA_LEFT );
/*------------------------------------------------------------*/
/* c1t = c1t + A * b1t */
FLA_Symv_external( FLA_UPPER_TRIANGULAR, alpha, A, b1t, beta, c1t );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ b1t, B2,
FLA_RIGHT );
FLA_Cont_with_1x3_to_1x2( &CL, /**/ &CR, C0, /**/ c1t, C2,
FLA_RIGHT );
}
return FLA_SUCCESS;
}示例5: Symm_ru_unb_var4
int Symm_ru_unb_var4( FLA_Obj A, FLA_Obj B, FLA_Obj C )
{
FLA_Obj ATL, ATR, A00, a01, A02,
ABL, ABR, a10t, alpha11, a12t,
A20, a21, A22;
FLA_Obj BL, BR, B0, b1, B2;
FLA_Obj CL, CR, C0, c1, C2;
FLA_Part_2x2( A, &ATL, &ATR,
&ABL, &ABR, 0, 0, FLA_TL );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
FLA_Part_1x2( C, &CL, &CR, 0, FLA_LEFT );
while ( FLA_Obj_length( ATL ) < FLA_Obj_length( A ) ){
FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, /**/ &a01, &A02,
/* ************* */ /* ************************** */
&a10t, /**/ &alpha11, &a12t,
ABL, /**/ ABR, &A20, /**/ &a21, &A22,
1, 1, FLA_BR );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
1, FLA_RIGHT );
FLA_Repart_1x2_to_1x3( CL, /**/ CR, &C0, /**/ &c1, &C2,
1, FLA_RIGHT );
/*------------------------------------------------------------*/
//c1 = (b1 * alpha11) + c1;
FLA_Axpy(alpha11, b1, c1);
//c1 = (B2 * a12t') + c1;
FLA_Gemv(FLA_NO_TRANSPOSE, FLA_ONE, B2, a12t, FLA_ONE, c1);
//C2 = (b1 * a12t) + C2
FLA_Ger(FLA_ONE, a12t, b1, C0);
/*------------------------------------------------------------*/
FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, a01, /**/ A02,
a10t, alpha11, /**/ a12t,
/* ************** */ /* ************************ */
&ABL, /**/ &ABR, A20, a21, /**/ A22,
FLA_TL );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
FLA_LEFT );
FLA_Cont_with_1x3_to_1x2( &CL, /**/ &CR, C0, c1, /**/ C2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例6: FLA_Axpy_sync_pipeline
FLA_Error FLA_Axpy_sync_pipeline( FLA_Obj alpha, FLA_Obj X, FLA_Obj B )
{
FLA_Obj XL, XR, X0, X1, X2;
FLA_Obj BL, BR, B0, B1, B2;
int b, i, nb_alg;
FLA_Part_1x2( X, &XL, &XR, 0, FLA_LEFT );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
// Compute the width of one lockable partition.
nb_alg = FLA_omp_compute_stage_width( X );
while ( FLA_Obj_width( XL ) < FLA_Obj_width( X ) ){
b = min( FLA_Obj_width( XR ), nb_alg );
FLA_Repart_1x2_to_1x3( XL, /**/ XR, &X0, /**/ &X1, &X2,
b, FLA_RIGHT );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
b, FLA_RIGHT );
/*------------------------------------------------------------*/
// Get the index of the current partition.
i = FLA_Obj_width(XL)/nb_alg;
// Acquire lock[i] (the lock for X1 and B1).
omp_set_lock( &fla_omp_lock[i] );
// B1 := alpha * X1 + B1
FLA_Axpy_external( alpha, X1, B1 );
// Release lock[i] (the lock for X1 and B1).
omp_unset_lock( &fla_omp_lock[i] );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &XL, /**/ &XR, X0, X1, /**/ X2,
FLA_LEFT );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例7: FLA_Trsm_llt_unb_var3
FLA_Error FLA_Trsm_llt_unb_var3( FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B )
{
FLA_Obj BL, BR, B0, b1, B2;
FLA_Scal_external( alpha, B );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
while ( FLA_Obj_width( BL ) < FLA_Obj_width( B ) ){
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
1, FLA_RIGHT );
/*------------------------------------------------------------*/
/* b1 = tril( A' ) \ b1; */
FLA_Trsv_external( FLA_LOWER_TRIANGULAR, FLA_TRANSPOSE, diagA, A, b1 );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例8: FLA_Herk_un_unb_var5
FLA_Error FLA_Herk_un_unb_var5( FLA_Obj alpha, FLA_Obj A, FLA_Obj beta, FLA_Obj C )
{
FLA_Obj AL, AR, A0, a1, A2;
FLA_Scalr_external( FLA_UPPER_TRIANGULAR, beta, C );
FLA_Part_1x2( A, &AL, &AR, 0, FLA_LEFT );
while ( FLA_Obj_width( AL ) < FLA_Obj_width( A ) ){
FLA_Repart_1x2_to_1x3( AL, /**/ AR, &A0, /**/ &a1, &A2,
1, FLA_RIGHT );
/*------------------------------------------------------------*/
/* C := C + a1 * a1' */
FLA_Her_external( FLA_UPPER_TRIANGULAR, alpha, a1, C );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &AL, /**/ &AR, A0, a1, /**/ A2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例9: FLASH_LQ_UT_solve
FLA_Error FLASH_LQ_UT_solve( FLA_Obj A, FLA_Obj T, FLA_Obj B, FLA_Obj X )
{
FLA_Obj W;
FLA_Obj AL, AR;
FLA_Obj XT, XB;
// Check parameters.
if ( FLA_Check_error_level() >= FLA_MIN_ERROR_CHECKING )
FLA_LQ_UT_solve_check( A, T, B, X );
FLASH_Apply_Q_UT_create_workspace( T, X, &W );
FLA_Part_1x2( A, &AL, &AR, FLA_Obj_length( A ), FLA_LEFT );
FLA_Part_2x1( X, &XT,
&XB, FLA_Obj_length( B ), FLA_TOP );
FLASH_Copy( B, XT );
FLASH_Trsm( FLA_LEFT, FLA_LOWER_TRIANGULAR, FLA_NO_TRANSPOSE,
FLA_NONUNIT_DIAG, FLA_ONE, AL, XT );
FLASH_Set( FLA_ZERO, XB );
FLASH_Apply_Q_UT( FLA_LEFT, FLA_NO_TRANSPOSE, FLA_FORWARD, FLA_ROWWISE,
A, T, W, X );
FLASH_Obj_free( &W );
return FLA_SUCCESS;
}示例10: FLA_Trmm_lun_unb_var4
FLA_Error FLA_Trmm_lun_unb_var4( FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B )
{
FLA_Obj BL, BR, B0, b1, B2;
FLA_Scal_external( alpha, B );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
while ( FLA_Obj_width( BR ) < FLA_Obj_width( B ) ){
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &b1, /**/ &B2,
1, FLA_LEFT );
/*------------------------------------------------------------*/
/* b1 = triu( A ) * b1 */
FLA_Trmv_external( FLA_UPPER_TRIANGULAR, FLA_NO_TRANSPOSE, diagA, A, b1 );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ b1, B2,
FLA_RIGHT );
}
return FLA_SUCCESS;
}示例11: FLA_Trsm_lun_blk_var4
FLA_Error FLA_Trsm_lun_blk_var4( FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trsm_t* cntl )
{
FLA_Obj BL, BR, B0, B1, B2;
dim_t b;
FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
while ( FLA_Obj_width( BR ) < FLA_Obj_width( B ) ){
b = FLA_Determine_blocksize( BL, FLA_LEFT, FLA_Cntl_blocksize( cntl ) );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &B1, /**/ &B2,
b, FLA_LEFT );
/*------------------------------------------------------------*/
/* B1 = triu( A ) \ B1; */
FLA_Trsm_internal( FLA_LEFT, FLA_UPPER_TRIANGULAR, FLA_NO_TRANSPOSE, diagA,
alpha, A, B1,
FLA_Cntl_sub_trsm( cntl ) );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ B1, B2,
FLA_RIGHT );
}
return FLA_SUCCESS;
}示例12: FLA_Herk_ln_blk_var5
FLA_Error FLA_Herk_ln_blk_var5( FLA_Obj alpha, FLA_Obj A, FLA_Obj beta, FLA_Obj C, fla_herk_t* cntl )
{
FLA_Obj AL, AR, A0, A1, A2;
dim_t b;
FLA_Scalr_internal( FLA_LOWER_TRIANGULAR, beta, C,
FLA_Cntl_sub_scalr( cntl ) );
FLA_Part_1x2( A, &AL, &AR, 0, FLA_LEFT );
while ( FLA_Obj_width( AL ) < FLA_Obj_width( A ) ) {
b = FLA_Determine_blocksize( AR, FLA_RIGHT, FLA_Cntl_blocksize( cntl ) );
FLA_Repart_1x2_to_1x3( AL, /**/ AR, &A0, /**/ &A1, &A2,
b, FLA_RIGHT );
/*------------------------------------------------------------*/
/* C = C + A1 * A1' */
FLA_Herk_internal( FLA_LOWER_TRIANGULAR, FLA_NO_TRANSPOSE,
alpha, A1, FLA_ONE, C,
FLA_Cntl_sub_herk( cntl ) );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &AL, /**/ &AR, A0, A1, /**/ A2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例13: FLA_Trmm_rlt_blk_var2
FLA_Error FLA_Trmm_rlt_blk_var2( FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t* cntl )
{
FLA_Obj ATL, ATR, A00, A01, A02,
ABL, ABR, A10, A11, A12,
A20, A21, A22;
FLA_Obj BL, BR, B0, B1, B2;
dim_t b;
FLA_Scal_internal( alpha, B,
FLA_Cntl_sub_scal( cntl ) );
FLA_Part_2x2( A, &ATL, &ATR,
&ABL, &ABR, 0, 0, FLA_BR );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
b = FLA_Determine_blocksize( ATL, FLA_TL, FLA_Cntl_blocksize( cntl ) );
FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &A01, /**/ &A02,
&A10, &A11, /**/ &A12,
/* ************* */ /* ******************** */
ABL, /**/ ABR, &A20, &A21, /**/ &A22,
b, b, FLA_TL );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &B1, /**/ &B2,
b, FLA_LEFT );
/*------------------------------------------------------------*/
/* B2 = B2 + B1 * A21'; */
FLA_Gemm_internal( FLA_NO_TRANSPOSE, FLA_TRANSPOSE,
FLA_ONE, B1, A21, FLA_ONE, B2,
FLA_Cntl_sub_gemm( cntl ) );
/* B1 = B1 * tril( A11 )'; */
FLA_Trmm_internal( FLA_RIGHT, FLA_LOWER_TRIANGULAR, FLA_TRANSPOSE, diagA,
FLA_ONE, A11, B1,
FLA_Cntl_sub_trmm( cntl ) );
/*------------------------------------------------------------*/
FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ A01, A02,
/* ************** */ /* ****************** */
A10, /**/ A11, A12,
&ABL, /**/ &ABR, A20, /**/ A21, A22,
FLA_BR );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ B1, B2,
FLA_RIGHT );
}
return FLA_SUCCESS;
}示例14: FLA_Sylv_nn_blk_var17
FLA_Error FLA_Sylv_nn_blk_var17( FLA_Obj isgn, FLA_Obj A, FLA_Obj B, FLA_Obj C, FLA_Obj scale, fla_sylv_t* cntl )
{
FLA_Obj BTL, BTR, B00, B01, B02,
BBL, BBR, B10, B11, B12,
B20, B21, B22;
FLA_Obj CL, CR, C0, C1, C2;
dim_t b;
FLA_Part_2x2( B, &BTL, &BTR,
&BBL, &BBR, 0, 0, FLA_TL );
FLA_Part_1x2( C, &CL, &CR, 0, FLA_LEFT );
while ( FLA_Obj_length( BTL ) < FLA_Obj_length( B ) ){
b = FLA_Determine_blocksize( CR, FLA_RIGHT, FLA_Cntl_blocksize( cntl ) );
FLA_Repart_2x2_to_3x3( BTL, /**/ BTR, &B00, /**/ &B01, &B02,
/* ************* */ /* ******************** */
&B10, /**/ &B11, &B12,
BBL, /**/ BBR, &B20, /**/ &B21, &B22,
b, b, FLA_BR );
FLA_Repart_1x2_to_1x3( CL, /**/ CR, &C0, /**/ &C1, &C2,
b, FLA_RIGHT );
// Loop Invariant:
// CL =
// CR =
/*------------------------------------------------------------*/
// C1 = sylv( A, B11, C1 -/+ C0 * B01 );
FLA_Gemm_internal( FLA_NO_TRANSPOSE, FLA_NO_TRANSPOSE,
FLA_NEGATE( isgn ), C0, B01, FLA_ONE, C1,
FLA_Cntl_sub_gemm1( cntl ) );
FLA_Sylv_internal( FLA_NO_TRANSPOSE, FLA_NO_TRANSPOSE,
isgn, A, B11, C1, scale,
FLA_Cntl_sub_sylv1( cntl ) );
/*------------------------------------------------------------*/
FLA_Cont_with_3x3_to_2x2( &BTL, /**/ &BTR, B00, B01, /**/ B02,
B10, B11, /**/ B12,
/* ************** */ /* ****************** */
&BBL, /**/ &BBR, B20, B21, /**/ B22,
FLA_TL );
FLA_Cont_with_1x3_to_1x2( &CL, /**/ &CR, C0, C1, /**/ C2,
FLA_LEFT );
}
return FLA_SUCCESS;
}示例15: FLA_Gemm_nt_blk_var5
FLA_Error FLA_Gemm_nt_blk_var5( FLA_Obj alpha, FLA_Obj A, FLA_Obj B, FLA_Obj beta, FLA_Obj C, fla_gemm_t* cntl )
{
FLA_Obj AL, AR, A0, A1, A2;
FLA_Obj BL, BR, B0, B1, B2;
dim_t b;
FLA_Scal_internal( beta, C,
FLA_Cntl_sub_scal( cntl ) );
FLA_Part_1x2( A, &AL, &AR, 0, FLA_LEFT );
FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
while ( FLA_Obj_width( AL ) < FLA_Obj_width( A ) ){
b = FLA_Determine_blocksize( AR, FLA_RIGHT, FLA_Cntl_blocksize( cntl ) );
FLA_Repart_1x2_to_1x3( AL, /**/ AR, &A0, /**/ &A1, &A2,
b, FLA_RIGHT );
FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
b, FLA_RIGHT );
/*------------------------------------------------------------*/
/* C = alpha * A1 * B1' + C; */
FLA_Gemm_internal( FLA_NO_TRANSPOSE, FLA_TRANSPOSE,
alpha, A1, B1, FLA_ONE, C,
FLA_Cntl_sub_gemm( cntl ) );
/*------------------------------------------------------------*/
FLA_Cont_with_1x3_to_1x2( &AL, /**/ &AR, A0, A1, /**/ A2,
FLA_LEFT );
FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
FLA_LEFT );
}
return FLA_SUCCESS;
}