int CORE_sormqr(int side, int trans,
int M, int N, int K, int IB,
float *A, int LDA,
float *T, int LDT,
float *C, int LDC,
float *WORK, int LDWORK)
{
int i, kb;
int i1, i3;
int nq, nw;
int ic = 0;
int jc = 0;
int ni = N;
int mi = M;
/* Check input arguments */
if ((side != PlasmaLeft) && (side != PlasmaRight)) {
coreblas_error(1, "Illegal value of side");
return -1;
}
/*
* NQ is the order of Q and NW is the minimum dimension of WORK
*/
if (side == PlasmaLeft) {
nq = M;
nw = N;
}
else {
nq = N;
nw = M;
}
if ((trans != PlasmaNoTrans) && (trans != PlasmaTrans)) {
coreblas_error(2, "Illegal value of trans");
return -2;
}
if (M < 0) {
coreblas_error(3, "Illegal value of M");
return -3;
}
if (N < 0) {
coreblas_error(4, "Illegal value of N");
return -4;
}
if ((K < 0) || (K > nq)) {
coreblas_error(5, "Illegal value of K");
return -5;
}
if ((IB < 0) || ( (IB == 0) && ((M > 0) && (N > 0)) )) {
coreblas_error(6, "Illegal value of IB");
return -6;
}
if ((LDA < max(1,nq)) && (nq > 0)) {
coreblas_error(8, "Illegal value of LDA");
return -8;
}
if ((LDC < max(1,M)) && (M > 0)) {
coreblas_error(12, "Illegal value of LDC");
return -12;
}
if ((LDWORK < max(1,nw)) && (nw > 0)) {
coreblas_error(14, "Illegal value of LDWORK");
return -14;
}
/* Quick return */
if ((M == 0) || (N == 0) || (K == 0))
return PLASMA_SUCCESS;
if (((side == PlasmaLeft) && (trans != PlasmaNoTrans))
|| ((side == PlasmaRight) && (trans == PlasmaNoTrans))) {
i1 = 0;
i3 = IB;
}
else {
i1 = ( ( K-1 ) / IB )*IB;
i3 = -IB;
}
for(i = i1; (i >- 1) && (i < K); i+=i3 ) {
kb = min(IB, K-i);
if (side == PlasmaLeft) {
/*
* H or H' is applied to C(i:m,1:n)
*/
mi = M - i;
ic = i;
}
else {
/*
* H or H' is applied to C(1:m,i:n)
*/
ni = N - i;
jc = i;
}
/*
* Apply H or H'
*/
LAPACKE_slarfb_work(LAPACK_COL_MAJOR,
lapack_const(side),
lapack_const(trans),
lapack_const(PlasmaForward),
lapack_const(PlasmaColumnwise),
mi, ni, kb,
&A[LDA*i+i], LDA,
&T[LDT*i], LDT,
&C[LDC*jc+ic], LDC,
WORK, LDWORK);
}
return PLASMA_SUCCESS;
}
lapack_int LAPACKE_slarfb( int matrix_order, char side, char trans, char direct,
char storev, lapack_int m, lapack_int n,
lapack_int k, const float* v, lapack_int ldv,
const float* t, lapack_int ldt, float* c,
lapack_int ldc )
{
lapack_int info = 0;
lapack_int ldwork = ( side=='l')?n:(( side=='r')?m:1);
float* work = NULL;
if( matrix_order != LAPACK_COL_MAJOR && matrix_order != LAPACK_ROW_MAJOR ) {
LAPACKE_xerbla( "LAPACKE_slarfb", -1 );
return -1;
}
#ifndef LAPACK_DISABLE_NAN_CHECK
/* Optionally check input matrices for NaNs */
lapack_int ncols_v = LAPACKE_lsame( storev, 'c' ) ? k :
( ( LAPACKE_lsame( storev, 'r' ) &&
LAPACKE_lsame( side, 'l' ) ) ? m :
( ( LAPACKE_lsame( storev, 'r' ) &&
LAPACKE_lsame( side, 'r' ) ) ? n : 1) );
lapack_int nrows_v = ( LAPACKE_lsame( storev, 'c' ) &&
LAPACKE_lsame( side, 'l' ) ) ? m :
( ( LAPACKE_lsame( storev, 'c' ) &&
LAPACKE_lsame( side, 'r' ) ) ? n :
( LAPACKE_lsame( storev, 'r' ) ? k : 1) );
if( LAPACKE_sge_nancheck( matrix_order, m, n, c, ldc ) ) {
return -13;
}
if( LAPACKE_sge_nancheck( matrix_order, k, k, t, ldt ) ) {
return -11;
}
if( LAPACKE_lsame( storev, 'c' ) && LAPACKE_lsame( direct, 'f' ) ) {
if( LAPACKE_str_nancheck( matrix_order, 'l', 'u', k, v, ldv ) )
return -9;
if( LAPACKE_sge_nancheck( matrix_order, nrows_v-k, ncols_v, &v[k*ldv],
ldv ) )
return -9;
} else if( LAPACKE_lsame( storev, 'c' ) && LAPACKE_lsame( direct, 'b' ) ) {
if( k > nrows_v ) {
LAPACKE_xerbla( "LAPACKE_slarfb", -8 );
return -8;
}
if( LAPACKE_str_nancheck( matrix_order, 'u', 'u', k,
&v[(nrows_v-k)*ldv], ldv ) )
return -9;
if( LAPACKE_sge_nancheck( matrix_order, nrows_v-k, ncols_v, v, ldv ) )
return -9;
} else if( LAPACKE_lsame( storev, 'r' ) && LAPACKE_lsame( direct, 'f' ) ) {
if( LAPACKE_str_nancheck( matrix_order, 'u', 'u', k, v, ldv ) )
return -9;
if( LAPACKE_sge_nancheck( matrix_order, nrows_v, ncols_v-k, &v[k],
ldv ) )
return -9;
} else if( LAPACKE_lsame( storev, 'r' ) && LAPACKE_lsame( direct, 'f' ) ) {
if( k > ncols_v ) {
LAPACKE_xerbla( "LAPACKE_slarfb", -8 );
return -8;
}
if( LAPACKE_str_nancheck( matrix_order, 'l', 'u', k, &v[ncols_v-k],
ldv ) )
return -9;
if( LAPACKE_sge_nancheck( matrix_order, nrows_v, ncols_v-k, v, ldv ) )
return -9;
}
#endif
/* Allocate memory for working array(s) */
work = (float*)LAPACKE_malloc( sizeof(float) * ldwork * MAX(1,k) );
if( work == NULL ) {
info = LAPACK_WORK_MEMORY_ERROR;
goto exit_level_0;
}
/* Call middle-level interface */
info = LAPACKE_slarfb_work( matrix_order, side, trans, direct, storev, m, n,
k, v, ldv, t, ldt, c, ldc, work, ldwork );
/* Release memory and exit */
LAPACKE_free( work );
exit_level_0:
if( info == LAPACK_WORK_MEMORY_ERROR ) {
LAPACKE_xerbla( "LAPACKE_slarfb", info );
}
return info;
}