MAGMA  2.3.0 Matrix Algebra for GPU and Multicore Architectures
laqps: Partial factorization; used by geqp3

## Functions

magma_int_t magma_claqps (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaFloatComplex *A, magma_int_t lda, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaFloatComplex *tau, float *vn1, float *vn2, magmaFloatComplex *auxv, magmaFloatComplex *F, magma_int_t ldf, magmaFloatComplex_ptr dF, magma_int_t lddf)
CLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. More...

magma_int_t magma_claqps_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaFloatComplex *tau, float *vn1, float *vn2, magmaFloatComplex_ptr dauxv, magmaFloatComplex_ptr dF, magma_int_t lddf)

magma_int_t magma_dlaqps (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, double *A, magma_int_t lda, magmaDouble_ptr dA, magma_int_t ldda, magma_int_t *jpvt, double *tau, double *vn1, double *vn2, double *auxv, double *F, magma_int_t ldf, magmaDouble_ptr dF, magma_int_t lddf)
DLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. More...

magma_int_t magma_dlaqps_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaDouble_ptr dA, magma_int_t ldda, magma_int_t *jpvt, double *tau, double *vn1, double *vn2, magmaDouble_ptr dauxv, magmaDouble_ptr dF, magma_int_t lddf)

magma_int_t magma_slaqps (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, float *A, magma_int_t lda, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t *jpvt, float *tau, float *vn1, float *vn2, float *auxv, float *F, magma_int_t ldf, magmaFloat_ptr dF, magma_int_t lddf)
SLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. More...

magma_int_t magma_slaqps_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t *jpvt, float *tau, float *vn1, float *vn2, magmaFloat_ptr dauxv, magmaFloat_ptr dF, magma_int_t lddf)

magma_int_t magma_zlaqps (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaDoubleComplex *A, magma_int_t lda, magmaDoubleComplex_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaDoubleComplex *tau, double *vn1, double *vn2, magmaDoubleComplex *auxv, magmaDoubleComplex *F, magma_int_t ldf, magmaDoubleComplex_ptr dF, magma_int_t lddf)
ZLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. More...

magma_int_t magma_zlaqps_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaDoubleComplex_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaDoubleComplex *tau, double *vn1, double *vn2, magmaDoubleComplex_ptr dauxv, magmaDoubleComplex_ptr dF, magma_int_t lddf)

magma_int_t magma_claqps2_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaFloatComplex_ptr dtau, magmaFloat_ptr dvn1, magmaFloat_ptr dvn2, magmaFloatComplex_ptr dauxv, magmaFloatComplex_ptr dF, magma_int_t lddf, magmaFloat_ptr dlsticcs, magma_queue_t queue)
CLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. More...

DLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. More...

magma_int_t magma_slaqps2_gpu (magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t *kb, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t *jpvt, magmaFloat_ptr dtau, magmaFloat_ptr dvn1, magmaFloat_ptr dvn2, magmaFloat_ptr dauxv, magmaFloat_ptr dF, magma_int_t lddf, magmaFloat_ptr dlsticcs, magma_queue_t queue)
SLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. More...

ZLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. More...

## Function Documentation

 magma_int_t magma_claqps ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaFloatComplex * A, magma_int_t lda, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaFloatComplex * tau, float * vn1, float * vn2, magmaFloatComplex * auxv, magmaFloatComplex * F, magma_int_t ldf, magmaFloatComplex_ptr dF, magma_int_t lddf )

CLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] A COMPLEX array, dimension (LDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] lda INTEGER The leading dimension of the array A. LDA >= max(1,M). [in,out] dA COMPLEX array, dimension (LDA,N) Copy of A on the GPU. Portions of A are updated on the CPU; portions of dA are updated on the GPU. See code for details. [in] ldda INTEGER The leading dimension of the array dA. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau COMPLEX array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] vn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] auxv COMPLEX array, dimension (NB) Auxiliar vector. [in,out] F COMPLEX array, dimension (LDF,NB) Matrix F' = L*Y'*A. [in] ldf INTEGER The leading dimension of the array F. LDF >= max(1,N). [in,out] dF COMPLEX array, dimension (LDDF,NB) Copy of F on the GPU. See code for details. [in] lddf INTEGER The leading dimension of the array dF. LDDF >= max(1,N).
 magma_int_t magma_claqps_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaFloatComplex * tau, float * vn1, float * vn2, magmaFloatComplex_ptr dauxv, magmaFloatComplex_ptr dF, magma_int_t lddf )
Deprecated:

CLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA COMPLEX array, dimension (LDDA,N), on the GPU. On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau COMPLEX array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] vn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] dauxv COMPLEX array, dimension (NB), on the GPU Auxiliary vector. [in,out] dF COMPLEX array, dimension (LDDF,NB), on the GPU Matrix F' = L*Y'*A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N).
 magma_int_t magma_dlaqps ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, double * A, magma_int_t lda, magmaDouble_ptr dA, magma_int_t ldda, magma_int_t * jpvt, double * tau, double * vn1, double * vn2, double * auxv, double * F, magma_int_t ldf, magmaDouble_ptr dF, magma_int_t lddf )

DLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] A DOUBLE PRECISION array, dimension (LDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] lda INTEGER The leading dimension of the array A. LDA >= max(1,M). [in,out] dA DOUBLE PRECISION array, dimension (LDA,N) Copy of A on the GPU. Portions of A are updated on the CPU; portions of dA are updated on the GPU. See code for details. [in] ldda INTEGER The leading dimension of the array dA. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau DOUBLE PRECISION array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] vn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] auxv DOUBLE PRECISION array, dimension (NB) Auxiliar vector. [in,out] F DOUBLE PRECISION array, dimension (LDF,NB) Matrix F' = L*Y'*A. [in] ldf INTEGER The leading dimension of the array F. LDF >= max(1,N). [in,out] dF DOUBLE PRECISION array, dimension (LDDF,NB) Copy of F on the GPU. See code for details. [in] lddf INTEGER The leading dimension of the array dF. LDDF >= max(1,N).
 magma_int_t magma_dlaqps_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaDouble_ptr dA, magma_int_t ldda, magma_int_t * jpvt, double * tau, double * vn1, double * vn2, magmaDouble_ptr dauxv, magmaDouble_ptr dF, magma_int_t lddf )
Deprecated:

DLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA DOUBLE PRECISION array, dimension (LDDA,N), on the GPU. On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau DOUBLE PRECISION array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] vn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] dauxv DOUBLE PRECISION array, dimension (NB), on the GPU Auxiliary vector. [in,out] dF DOUBLE PRECISION array, dimension (LDDF,NB), on the GPU Matrix F' = L*Y'*A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N).
 magma_int_t magma_slaqps ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, float * A, magma_int_t lda, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t * jpvt, float * tau, float * vn1, float * vn2, float * auxv, float * F, magma_int_t ldf, magmaFloat_ptr dF, magma_int_t lddf )

SLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] A REAL array, dimension (LDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] lda INTEGER The leading dimension of the array A. LDA >= max(1,M). [in,out] dA REAL array, dimension (LDA,N) Copy of A on the GPU. Portions of A are updated on the CPU; portions of dA are updated on the GPU. See code for details. [in] ldda INTEGER The leading dimension of the array dA. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau REAL array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] vn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] auxv REAL array, dimension (NB) Auxiliar vector. [in,out] F REAL array, dimension (LDF,NB) Matrix F' = L*Y'*A. [in] ldf INTEGER The leading dimension of the array F. LDF >= max(1,N). [in,out] dF REAL array, dimension (LDDF,NB) Copy of F on the GPU. See code for details. [in] lddf INTEGER The leading dimension of the array dF. LDDF >= max(1,N).
 magma_int_t magma_slaqps_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t * jpvt, float * tau, float * vn1, float * vn2, magmaFloat_ptr dauxv, magmaFloat_ptr dF, magma_int_t lddf )
Deprecated:

SLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3. It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA REAL array, dimension (LDDA,N), on the GPU. On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau REAL array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] vn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] dauxv REAL array, dimension (NB), on the GPU Auxiliary vector. [in,out] dF REAL array, dimension (LDDF,NB), on the GPU Matrix F' = L*Y'*A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N).
 magma_int_t magma_zlaqps ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaDoubleComplex * A, magma_int_t lda, magmaDoubleComplex_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaDoubleComplex * tau, double * vn1, double * vn2, magmaDoubleComplex * auxv, magmaDoubleComplex * F, magma_int_t ldf, magmaDoubleComplex_ptr dF, magma_int_t lddf )

ZLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] A COMPLEX_16 array, dimension (LDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] lda INTEGER The leading dimension of the array A. LDA >= max(1,M). [in,out] dA COMPLEX_16 array, dimension (LDA,N) Copy of A on the GPU. Portions of A are updated on the CPU; portions of dA are updated on the GPU. See code for details. [in] ldda INTEGER The leading dimension of the array dA. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau COMPLEX_16 array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] vn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] auxv COMPLEX_16 array, dimension (NB) Auxiliar vector. [in,out] F COMPLEX_16 array, dimension (LDF,NB) Matrix F' = L*Y'*A. [in] ldf INTEGER The leading dimension of the array F. LDF >= max(1,N). [in,out] dF COMPLEX_16 array, dimension (LDDF,NB) Copy of F on the GPU. See code for details. [in] lddf INTEGER The leading dimension of the array dF. LDDF >= max(1,N).
 magma_int_t magma_zlaqps_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaDoubleComplex_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaDoubleComplex * tau, double * vn1, double * vn2, magmaDoubleComplex_ptr dauxv, magmaDoubleComplex_ptr dF, magma_int_t lddf )
Deprecated:

ZLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3. It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA COMPLEX_16 array, dimension (LDDA,N), on the GPU. On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] tau COMPLEX_16 array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] vn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] vn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] dauxv COMPLEX_16 array, dimension (NB), on the GPU Auxiliary vector. [in,out] dF COMPLEX_16 array, dimension (LDDF,NB), on the GPU Matrix F' = L*Y'*A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N).
 magma_int_t magma_claqps2_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaFloatComplex_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaFloatComplex_ptr dtau, magmaFloat_ptr dvn1, magmaFloat_ptr dvn2, magmaFloatComplex_ptr dauxv, magmaFloatComplex_ptr dF, magma_int_t lddf, magmaFloat_ptr dlsticcs, magma_queue_t queue )

CLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA COMPLEX array, dimension (LDDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] dtau COMPLEX array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] dvn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] dvn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] dauxv COMPLEX array, dimension (NB) Auxiliar vector. [in,out] dF COMPLEX array, dimension (LDDF,NB) Matrix F**H = L * Y**H * A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N). dlsticcs TODO: undocumented [in] queue magma_queue_t Queue to execute in.

DLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA DOUBLE PRECISION array, dimension (LDDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] dtau DOUBLE PRECISION array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] dvn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] dvn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] dauxv DOUBLE PRECISION array, dimension (NB) Auxiliar vector. [in,out] dF DOUBLE PRECISION array, dimension (LDDF,NB) Matrix F**H = L * Y**H * A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N). dlsticcs TODO: undocumented [in] queue magma_queue_t Queue to execute in.
 magma_int_t magma_slaqps2_gpu ( magma_int_t m, magma_int_t n, magma_int_t offset, magma_int_t nb, magma_int_t * kb, magmaFloat_ptr dA, magma_int_t ldda, magma_int_t * jpvt, magmaFloat_ptr dtau, magmaFloat_ptr dvn1, magmaFloat_ptr dvn2, magmaFloat_ptr dauxv, magmaFloat_ptr dF, magma_int_t lddf, magmaFloat_ptr dlsticcs, magma_queue_t queue )

SLAQPS computes a step of QR factorization with column pivoting of a real M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA REAL array, dimension (LDDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] dtau REAL array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] dvn1 REAL array, dimension (N) The vector with the partial column norms. [in,out] dvn2 REAL array, dimension (N) The vector with the exact column norms. [in,out] dauxv REAL array, dimension (NB) Auxiliar vector. [in,out] dF REAL array, dimension (LDDF,NB) Matrix F**H = L * Y**H * A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N). dlsticcs TODO: undocumented [in] queue magma_queue_t Queue to execute in.

ZLAQPS computes a step of QR factorization with column pivoting of a complex M-by-N matrix A by using Blas-3.

It tries to factorize NB columns from A starting from the row OFFSET+1, and updates all of the matrix with Blas-3 xGEMM.

In some cases, due to catastrophic cancellations, it cannot factorize NB columns. Hence, the actual number of factorized columns is returned in KB.

Block A(1:OFFSET,1:N) is accordingly pivoted, but not factorized.

Parameters
 [in] m INTEGER The number of rows of the matrix A. M >= 0. [in] n INTEGER The number of columns of the matrix A. N >= 0 [in] offset INTEGER The number of rows of A that have been factorized in previous steps. [in] nb INTEGER The number of columns to factorize. [out] kb INTEGER The number of columns actually factorized. [in,out] dA COMPLEX*16 array, dimension (LDDA,N) On entry, the M-by-N matrix A. On exit, block A(OFFSET+1:M,1:KB) is the triangular factor obtained and block A(1:OFFSET,1:N) has been accordingly pivoted, but no factorized. The rest of the matrix, block A(OFFSET+1:M,KB+1:N) has been updated. [in] ldda INTEGER The leading dimension of the array A. LDDA >= max(1,M). [in,out] jpvt INTEGER array, dimension (N) JPVT(I) = K <==> Column K of the full matrix A has been permuted into position I in AP. [out] dtau COMPLEX*16 array, dimension (KB) The scalar factors of the elementary reflectors. [in,out] dvn1 DOUBLE PRECISION array, dimension (N) The vector with the partial column norms. [in,out] dvn2 DOUBLE PRECISION array, dimension (N) The vector with the exact column norms. [in,out] dauxv COMPLEX*16 array, dimension (NB) Auxiliar vector. [in,out] dF COMPLEX*16 array, dimension (LDDF,NB) Matrix F**H = L * Y**H * A. [in] lddf INTEGER The leading dimension of the array F. LDDF >= max(1,N). dlsticcs TODO: undocumented [in] queue magma_queue_t Queue to execute in.