core_zswpab.c File Reference

#include <stdlib.h>
#include "common.h"
#include "quark.h"

Include dependency graph for core_zswpab.c:

Go to the source code of this file.

Functions
void	CORE_zswpab (int i, int n1, int n2, PLASMA_Complex64_t *A, PLASMA_Complex64_t *work)
void	QUARK_CORE_zswpab (Quark *quark, Quark_Task_Flags *task_flags, int i, int n1, int n2, PLASMA_Complex64_t *A, int szeA)
void	CORE_zswpab_quark (Quark *quark)

---

Detailed Description

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

This work is the implementation of an inplace transformation based on the GKK algorithm by Gustavson, Karlsson, Kagstrom and its fortran implementation.

Version:

2.4.5

Author:

Mathieu Faverge

Date:

2010-11-15

normal z -> c d s

Definition in file core_zswpab.c.

---

Function Documentation

void CORE_zswpab	(	int	i,
		int	n1,
		int	n2,
		PLASMA_Complex64_t *	A,
		PLASMA_Complex64_t *	work
	)

---

CORE_zswpab swaps two adjacent contiguous blocks of data.

n1                     n2

+————-+——————————-+

become : n2 n1 +——————————-+————-+

Parameters:

[in,out]	A	Array of size i+n1+n2. On entry, a block of size n1 followed by a block of size n2. On exit, the block of size n1 follows the block of size n2.
[in]	i	First block starts at A[i].
[in]	n1	Size of the first block to swap.
[in]	n2	Size of the second block to swap.
[out]	work	Workspace array of size min(n1, n2).

Definition at line 63 of file core_zswpab.c.

                                                                                         {
    PLASMA_Complex64_t *A0 = &(A[i]);
    PLASMA_Complex64_t *A1 = &(A[i+n1]);
    PLASMA_Complex64_t *A2 = &(A[i+n2]);
    int j;
    
    if( n1 < n2 ) {
        memcpy(work,  A0, n1*sizeof(PLASMA_Complex64_t));
        for (j=0; j<n2; j++)
            A0[j] = A1[j];
        memcpy(A2, work,  n1*sizeof(PLASMA_Complex64_t));
    } else {
        memcpy(work,  A1, n2*sizeof(PLASMA_Complex64_t));
        for (j=n1-1; j>-1; j--)
            A2[j] = A0[j];
        memcpy(A0, work,  n2*sizeof(PLASMA_Complex64_t));
    }
}

Here is the caller graph for this function:

void CORE_zswpab_quark

(

Quark *

quark

)

Definition at line 107 of file core_zswpab.c.

References A, CORE_zswpab(), and quark_unpack_args_5.

{
    int i;
    int n1;
    int n2;
    PLASMA_Complex64_t *A;
    PLASMA_Complex64_t *work;
    
    quark_unpack_args_5(quark, i, n1, n2, A, work);
    CORE_zswpab( i, n1, n2, A, work);
}

Here is the call graph for this function:

Here is the caller graph for this function:

void QUARK_CORE_zswpab	(	Quark *	quark,
		Quark_Task_Flags *	task_flags,
		int	i,
		int	n1,
		int	n2,
		PLASMA_Complex64_t *	A,
		int	szeA
	)

Definition at line 85 of file core_zswpab.c.

References CORE_zswpab_quark(), DAG_CORE_SWPAB, INOUT, min, QUARK_Insert_Task(), SCRATCH, and VALUE.

{
    DAG_CORE_SWPAB;
    QUARK_Insert_Task(
        quark, CORE_zswpab_quark, task_flags,
        sizeof(int),                           &i,   VALUE,
        sizeof(int),                           &n1,  VALUE,
        sizeof(int),                           &n2,  VALUE,
        sizeof(PLASMA_Complex64_t)*szeA,       A,            INOUT,
        sizeof(PLASMA_Complex64_t)*min(n1,n2), NULL,         SCRATCH,
        0);
}

Here is the call graph for this function: