#include <lapacke.h>
#include "common.h"

Include dependency graph for core_ztslqt.c:

Macros
#define	COMPLEX

Functions
int	CORE_ztslqt (int M, int N, int IB, PLASMA_Complex64_t A1, int LDA1, PLASMA_Complex64_t A2, int LDA2, PLASMA_Complex64_t T, int LDT, PLASMA_Complex64_t TAU, PLASMA_Complex64_t *WORK)
void	QUARK_CORE_ztslqt (Quark quark, Quark_Task_Flags task_flags, int m, int n, int ib, int nb, PLASMA_Complex64_t A1, int lda1, PLASMA_Complex64_t A2, int lda2, PLASMA_Complex64_t *T, int ldt)
void	CORE_ztslqt_quark (Quark *quark)

Detailed Description

PLASMA core_blas kernel PLASMA is a software package provided by Univ. of Tennessee, Univ. of California Berkeley and Univ. of Colorado Denver

Version:: 2.4.5

Author:: Hatem Ltaief; Mathieu Faverge; Jakub Kurzak

Date:: 2010-11-15 normal z -> c d s

Definition in file core_ztslqt.c.

Macro Definition Documentation

#define COMPLEX

Definition at line 20 of file core_ztslqt.c.

Function Documentation

int CORE_ztslqt	(	int	M,
		int	N,
		int	IB,
		PLASMA_Complex64_t *	A1,
		int	LDA1,
		PLASMA_Complex64_t *	A2,
		int	LDA2,
		PLASMA_Complex64_t *	T,
		int	LDT,
		PLASMA_Complex64_t *	TAU,
		PLASMA_Complex64_t *	WORK
	)

CORE_ztslqt computes a LQ factorization of a rectangular matrix formed by coupling side-by-side a complex M-by-M lower triangular tile A1 and a complex M-by-N tile A2:

| A1 A2 | = L * Q

The tile Q is represented as a product of elementary reflectors

Q = H(k)' . . . H(2)' H(1)', where k = min(M,N).

Each H(i) has the form

H(i) = I - tau * v * v'

where tau is a complex scalar, and v is a complex vector with v(1:i-1) = 0 and v(i) = 1; conjg(v(i+1:n)) is stored on exit in A2(i,1:n), and tau in TAU(i).

Parameters:

[in]	M	The number of rows of the tile A1 and A2. M >= 0. The number of columns of the tile A1.
[in]	N	The number of columns of the tile A2. N >= 0.
[in]	IB	The inner-blocking size. IB >= 0.
[in,out]	A1	On entry, the M-by-M tile A1. On exit, the elements on and below the diagonal of the array contain the M-by-M lower trapezoidal tile L; the elements above the diagonal are not referenced.
[in]	LDA1	The leading dimension of the array A1. LDA1 >= max(1,M).
[in,out]	A2	On entry, the M-by-N tile A2. On exit, all the elements with the array TAU, represent the unitary tile Q as a product of elementary reflectors (see Further Details).
[in]	LDA2	The leading dimension of the tile A2. LDA2 >= max(1,M).
[out]	T	The IB-by-N triangular factor T of the block reflector. T is upper triangular by block (economic storage); The rest of the array is not referenced.
[in]	LDT	The leading dimension of the array T. LDT >= IB.
[out]	TAU	The scalar factors of the elementary reflectors (see Further Details).
[out]	WORK

Returns:

Return values:

PLASMA_SUCCESS	successful exit
<0	if -i, the i-th argument had an illegal value

Definition at line 107 of file core_ztslqt.c.

References CBLAS_SADDR, cblas_zaxpy(), cblas_zcopy(), cblas_zgemv(), cblas_zgerc(), cblas_ztrmv(), CblasColMajor, CORE_ztsmlq(), coreblas_error, max, min, PLASMA_SUCCESS, PlasmaConjTrans, PlasmaNonUnit, PlasmaNoTrans, PlasmaRight, and PlasmaUpper.

{
    static PLASMA_Complex64_t zone  = 1.0;
    static PLASMA_Complex64_t zzero = 0.0;
    PLASMA_Complex64_t alpha;
    int i, ii, sb;
    /* Check input arguments */
    if (M < 0) {
        coreblas_error(1, "Illegal value of M");
        return -1;
    }
    if (N < 0) {
        coreblas_error(2, "Illegal value of N");
        return -2;
    }
    if (IB < 0) {
        coreblas_error(3, "Illegal value of IB");
        return -3;
    }
    if ((LDA2 < max(1,M)) && (M > 0)) {
        coreblas_error(8, "Illegal value of LDA2");
        return -8;
    }
    /* Quick return */
    if ((M == 0) || (N == 0) || (IB == 0))
        return PLASMA_SUCCESS;
    for(ii = 0; ii < M; ii += IB) {
        sb = min(M-ii, IB);
        for(i = 0; i < sb; i++) {
            /*
             * Generate elementary reflector H( II*IB+I ) to annihilate A( II*IB+I, II*IB+I:N ).
             */
#ifdef COMPLEX
            LAPACKE_zlacgv_work(N, &A2[ii+i], LDA2);
            LAPACKE_zlacgv_work(1, &A1[LDA1*(ii+i)+ii+i], LDA1);
#endif
            LAPACKE_zlarfg_work(N+1, &A1[LDA1*(ii+i)+ii+i], &A2[ii+i], LDA2, &TAU[ii+i]);
            alpha = -(TAU[ii+i]);
            if (ii+i+1 < M) {
                /*
                 * Apply H( II+I-1 ) to A( II+I:II+IB-1, II+I-1:N  ) from the right.
                 */
                cblas_zcopy(
                    sb-i-1,
                    &A1[LDA1*(ii+i)+(ii+i+1)], 1,
                    WORK, 1);
                cblas_zgemv(
                    CblasColMajor, (CBLAS_TRANSPOSE)PlasmaNoTrans,
                    sb-i-1, N,
                    CBLAS_SADDR(zone), &A2[ii+i+1], LDA2,
                    &A2[ii+i], LDA2,
                    CBLAS_SADDR(zone), WORK, 1);
                cblas_zaxpy(
                    sb-i-1, CBLAS_SADDR(alpha),
                    WORK, 1,
                    &A1[LDA1*(ii+i)+ii+i+1], 1);
                cblas_zgerc(
                    CblasColMajor, sb-i-1, N,
                    CBLAS_SADDR(alpha), WORK, 1,
                    &A2[ii+i], LDA2,
                    &A2[ii+i+1], LDA2);
            }
            /*
             * Calculate T.
             */
            cblas_zgemv(
                CblasColMajor, (CBLAS_TRANSPOSE)PlasmaNoTrans, i, N,
                CBLAS_SADDR(alpha), &A2[ii], LDA2,
                &A2[ii+i], LDA2,
                CBLAS_SADDR(zzero), &T[LDT*(ii+i)], 1);
#ifdef COMPLEX
            LAPACKE_zlacgv_work(N, &A2[ii+i], LDA2 );
            LAPACKE_zlacgv_work(1, &A1[LDA1*(ii+i)+ii+i], LDA1 );
#endif
            cblas_ztrmv(
                CblasColMajor, (CBLAS_UPLO)PlasmaUpper,
                (CBLAS_TRANSPOSE)PlasmaNoTrans, (CBLAS_DIAG)PlasmaNonUnit, i,
                &T[LDT*ii], LDT,
                &T[LDT*(ii+i)], 1);
            T[LDT*(ii+i)+i] = TAU[ii+i];
        }
        if (M > ii+sb) {
            CORE_ztsmlq(
                PlasmaRight, PlasmaConjTrans,
                M-(ii+sb), sb, M-(ii+sb), N, IB, IB,
                &A1[LDA1*ii+ii+sb], LDA1,
                &A2[ii+sb], LDA2,
                &A2[ii], LDA2,
                &T[LDT*ii], LDT,
                WORK, LDA1);
        }
    }
    return PLASMA_SUCCESS;
}

Here is the call graph for this function:

Here is the caller graph for this function:

void CORE_ztslqt_quark ( Quark * quark )

Definition at line 247 of file core_ztslqt.c.

References CORE_ztslqt(), quark_unpack_args_11, T, and TAU.

{
    int m;
    int n;
    int ib;
    PLASMA_Complex64_t *A1;
    int lda1;
    PLASMA_Complex64_t *A2;
    int lda2;
    PLASMA_Complex64_t *T;
    int ldt;
    PLASMA_Complex64_t *TAU;
    PLASMA_Complex64_t *WORK;
    quark_unpack_args_11(quark, m, n, ib, A1, lda1, A2, lda2, T, ldt, TAU, WORK);
    CORE_ztslqt(m, n, ib, A1, lda1, A2, lda2, T, ldt, TAU, WORK);
}

Here is the call graph for this function:

Here is the caller graph for this function:

void QUARK_CORE_ztslqt	(	Quark *	quark,
		Quark_Task_Flags *	task_flags,
		int	m,
		int	n,
		int	ib,
		int	nb,
		PLASMA_Complex64_t *	A1,
		int	lda1,
		PLASMA_Complex64_t *	A2,
		int	lda2,
		PLASMA_Complex64_t *	T,
		int	ldt
	)

Definition at line 218 of file core_ztslqt.c.

References CORE_ztslqt_quark(), DAG_CORE_TSLQT, INOUT, LOCALITY, OUTPUT, QUARK_Insert_Task(), QUARK_REGION_D, QUARK_REGION_L, SCRATCH, and VALUE.

{
    DAG_CORE_TSLQT;
    QUARK_Insert_Task(quark, CORE_ztslqt_quark, task_flags,
        sizeof(int),                        &m,     VALUE,
        sizeof(int),                        &n,     VALUE,
        sizeof(int),                        &ib,    VALUE,
        sizeof(PLASMA_Complex64_t)*nb*nb,    A1,            INOUT | QUARK_REGION_D | QUARK_REGION_L,
        sizeof(int),                        &lda1,  VALUE,
        sizeof(PLASMA_Complex64_t)*nb*nb,    A2,            INOUT | LOCALITY,
        sizeof(int),                        &lda2,  VALUE,
        sizeof(PLASMA_Complex64_t)*ib*nb,    T,             OUTPUT,
        sizeof(int),                        &ldt,   VALUE,
        sizeof(PLASMA_Complex64_t)*nb,       NULL,          SCRATCH,
        sizeof(PLASMA_Complex64_t)*ib*nb,    NULL,          SCRATCH,
        0);
}

Here is the call graph for this function:

Here is the caller graph for this function:

Macros

Functions

Detailed Description

Macro Definition Documentation

Function Documentation