/*********************************************************************/

/* Copyright 2009, 2010 The University of Texas at Austin.           */

/* All rights reserved.                                              */

/*                                                                   */

/* Redistribution and use in source and binary forms, with or        */

/* without modification, are permitted provided that the following   */

/* conditions are met:                                               */

/*                                                                   */

/*   1. Redistributions of source code must retain the above         */

/*      copyright notice, this list of conditions and the following  */

/*      disclaimer.                                                  */

/*                                                                   */

/*   2. Redistributions in binary form must reproduce the above      */

/*      copyright notice, this list of conditions and the following  */

/*      disclaimer in the documentation and/or other materials       */

/*      provided with the distribution.                              */

/*                                                                   */

/*    THIS  SOFTWARE IS PROVIDED  BY THE  UNIVERSITY OF  TEXAS AT    */

/*    AUSTIN  ``AS IS''  AND ANY  EXPRESS OR  IMPLIED WARRANTIES,    */

/*    INCLUDING, BUT  NOT LIMITED  TO, THE IMPLIED  WARRANTIES OF    */

/*    MERCHANTABILITY  AND FITNESS FOR  A PARTICULAR  PURPOSE ARE    */

/*    DISCLAIMED.  IN  NO EVENT SHALL THE UNIVERSITY  OF TEXAS AT    */

/*    AUSTIN OR CONTRIBUTORS BE  LIABLE FOR ANY DIRECT, INDIRECT,    */

/*    INCIDENTAL,  SPECIAL, EXEMPLARY,  OR  CONSEQUENTIAL DAMAGES    */

/*    (INCLUDING, BUT  NOT LIMITED TO,  PROCUREMENT OF SUBSTITUTE    */

/*    GOODS  OR  SERVICES; LOSS  OF  USE,  DATA,  OR PROFITS;  OR    */

/*    BUSINESS INTERRUPTION) HOWEVER CAUSED  AND ON ANY THEORY OF    */

/*    LIABILITY, WHETHER  IN CONTRACT, STRICT  LIABILITY, OR TORT    */

/*    (INCLUDING NEGLIGENCE OR OTHERWISE)  ARISING IN ANY WAY OUT    */

/*    OF  THE  USE OF  THIS  SOFTWARE,  EVEN  IF ADVISED  OF  THE    */

/*    POSSIBILITY OF SUCH DAMAGE.                                    */

/*                                                                   */

/* The views and conclusions contained in the software and           */

/* documentation are those of the authors and should not be          */

/* interpreted as representing official policies, either expressed   */

/* or implied, of The University of Texas at Austin.                 */

/*********************************************************************/

#define ASSEMBLER

#include "common.h"

#ifndef WINDOWS_ABI

#define M	ARG1

#define X	ARG4

#define INCX	ARG5

#define Y	ARG6

#define INCY	ARG2

#else

#define M	ARG1

#define X	ARG2

#define INCX	ARG3

#define Y	ARG4

#define INCY	%r10

#endif

#define	YY	%r11

#define ALPHA	%xmm15

#include "l1param.h"

	PROLOGUE

	PROFCODE

#ifndef WINDOWS_ABI

#ifndef XDOUBLE

	movq	 8(%rsp), INCY

#else

	movq	24(%rsp), INCY

#endif

	movaps	%xmm0,  ALPHA

#else

	movaps	%xmm3,  ALPHA

	movq	40(%rsp), X

	movq	48(%rsp), INCX

	movq	56(%rsp), Y

	movq	64(%rsp), INCY

#endif

	SAVEREGISTERS

	leaq	(, INCX, SIZE), INCX

	leaq	(, INCY, SIZE), INCY

	testq	M, M

	jle	.L29

	cmpq	$SIZE, INCX

	jne	.L20

	cmpq	$SIZE, INCY

	jne	.L20

	movq	M,  %rax

	sarq	$3, %rax

	jle	.L13

	movsd	0 * SIZE(X), %xmm0

	movsd	1 * SIZE(X), %xmm1

	movsd	2 * SIZE(X), %xmm2

	movsd	3 * SIZE(X), %xmm3

	movsd	0 * SIZE(Y), %xmm4

	movsd	1 * SIZE(Y), %xmm5

	movsd	2 * SIZE(Y), %xmm6

	movsd	3 * SIZE(Y), %xmm7

	movsd	4 * SIZE(X), %xmm8

	mulsd	ALPHA, %xmm0

	movsd	5 * SIZE(X), %xmm9

	mulsd	ALPHA, %xmm1

	movsd	6 * SIZE(X), %xmm10

	mulsd	ALPHA, %xmm2

	movsd	7 * SIZE(X), %xmm11

	mulsd	ALPHA, %xmm3

	decq	%rax

	jle .L12

	ALIGN_3

.L11:

#ifdef PREFETCH

	PREFETCH (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	addsd	%xmm4, %xmm0

	movsd	4 * SIZE(Y), %xmm4

	addsd	%xmm5, %xmm1

	movsd	5 * SIZE(Y), %xmm5

	addsd	%xmm6, %xmm2

	movsd	6 * SIZE(Y), %xmm6

	addsd	%xmm7, %xmm3

	movsd	7 * SIZE(Y), %xmm7

	movsd	%xmm0, 0 * SIZE(Y)

	mulsd	ALPHA, %xmm8

	movsd	 8 * SIZE(X), %xmm0

	movsd	%xmm1, 1 * SIZE(Y)

	mulsd	ALPHA, %xmm9

	movsd	 9 * SIZE(X), %xmm1

	movsd	%xmm2, 2 * SIZE(Y)

	mulsd	ALPHA, %xmm10

	movsd	10 * SIZE(X), %xmm2

	movsd	%xmm3, 3 * SIZE(Y)

	mulsd	ALPHA, %xmm11

	movsd	11 * SIZE(X), %xmm3

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(Y)

#endif

	addsd	%xmm4, %xmm8

	movsd	 8 * SIZE(Y), %xmm4

	addsd	%xmm5, %xmm9

	movsd	 9 * SIZE(Y), %xmm5

	addsd	%xmm6, %xmm10

	movsd	10 * SIZE(Y), %xmm6

	addsd	%xmm7, %xmm11

	movsd	11 * SIZE(Y), %xmm7

	movsd	%xmm8,  4 * SIZE(Y)

	mulsd	ALPHA, %xmm0

	movsd	12 * SIZE(X), %xmm8

	movsd	%xmm9,  5 * SIZE(Y)

	mulsd	ALPHA, %xmm1

	movsd	13 * SIZE(X), %xmm9

	movsd	%xmm10, 6 * SIZE(Y)

	mulsd	ALPHA, %xmm2

 	movsd	14 * SIZE(X), %xmm10

	movsd	%xmm11, 7 * SIZE(Y)

	mulsd	ALPHA, %xmm3

	movsd	15 * SIZE(X), %xmm11

	addq	$8 * SIZE, Y

	addq	$8 * SIZE, X

	decq	%rax

	jg	.L11

	ALIGN_3

.L12:

	addsd	%xmm4, %xmm0

	movsd	4 * SIZE(Y), %xmm4

	addsd	%xmm5, %xmm1

	movsd	5 * SIZE(Y), %xmm5

	addsd	%xmm6, %xmm2

	movsd	6 * SIZE(Y), %xmm6

	addsd	%xmm7, %xmm3

	movsd	7 * SIZE(Y), %xmm7

	movsd	%xmm0, 0 * SIZE(Y)

	mulsd	ALPHA, %xmm8

	movsd	%xmm1, 1 * SIZE(Y)

	mulsd	ALPHA, %xmm9

	movsd	%xmm2, 2 * SIZE(Y)

	mulsd	ALPHA, %xmm10

	movsd	%xmm3, 3 * SIZE(Y)

	mulsd	ALPHA, %xmm11

	addsd	%xmm4, %xmm8

	addsd	%xmm5, %xmm9

	addsd	%xmm6, %xmm10

	addsd	%xmm7, %xmm11

	movsd	%xmm8,  4 * SIZE(Y)

	movsd	%xmm9,  5 * SIZE(Y)

	movsd	%xmm10, 6 * SIZE(Y)

	movsd	%xmm11, 7 * SIZE(Y)

	addq	$8 * SIZE, X

	addq	$8 * SIZE, Y

	ALIGN_3

.L13:

	movq	M,  %rax

	andq	$4, %rax

	jle	.L15

	ALIGN_3

	movsd	0 * SIZE(X), %xmm0

	movsd	1 * SIZE(X), %xmm1

	movsd	2 * SIZE(X), %xmm2

	movsd	3 * SIZE(X), %xmm3

	movsd	0 * SIZE(Y), %xmm4

	mulsd	ALPHA, %xmm0

	movsd	1 * SIZE(Y), %xmm5

	mulsd	ALPHA, %xmm1

	movsd	2 * SIZE(Y), %xmm6

	mulsd	ALPHA, %xmm2

	movsd	3 * SIZE(Y), %xmm7

	mulsd	ALPHA, %xmm3

	addsd	%xmm4, %xmm0

	addsd	%xmm5, %xmm1

	addsd	%xmm6, %xmm2

	addsd	%xmm7, %xmm3

	movsd	%xmm0, 0 * SIZE(Y)

	movsd	%xmm1, 1 * SIZE(Y)

	movsd	%xmm2, 2 * SIZE(Y)

	movsd	%xmm3, 3 * SIZE(Y)

	addq	$4 * SIZE, X

	addq	$4 * SIZE, Y

	ALIGN_3

.L15:

	movq	M,  %rax

	andq	$2, %rax

	jle	.L16

	ALIGN_3

	movsd	0 * SIZE(X), %xmm0

	movsd	0 * SIZE(Y), %xmm4

	movsd	1 * SIZE(X), %xmm1

	movsd	1 * SIZE(Y), %xmm5

	mulsd	ALPHA, %xmm0

	mulsd	ALPHA, %xmm1

	addsd	%xmm4, %xmm0

	addsd	%xmm5, %xmm1

	movsd	%xmm0, 0 * SIZE(Y)

	movsd	%xmm1, 1 * SIZE(Y)

	addq	$2 * SIZE, X

	addq	$2 * SIZE, Y

	ALIGN_3

.L16:

	movq	M,  %rax

	andq	$1, %rax

	jle	.L19

	ALIGN_3

	movsd	0 * SIZE(X), %xmm0

	mulsd	ALPHA, %xmm0

	addsd	0 * SIZE(Y), %xmm0

	movsd	%xmm0, 	0 * SIZE(Y)

	addq	$SIZE, Y

	ALIGN_3

.L19:

	xorq	%rax,%rax

	RESTOREREGISTERS

	ret

	ALIGN_3

.L20:

	movq	Y, YY

	movq	M,  %rax

	sarq	$3, %rax

	jle	.L23

	movsd	(X), %xmm0

	addq	INCX, X

	movsd	(X), %xmm1

	addq	INCX, X

	movsd	(X), %xmm2

	addq	INCX, X

	movsd	(X), %xmm3

	addq	INCX, X

	movsd	(Y), %xmm4

	addq	INCY, Y

	movsd	(Y), %xmm5

	addq	INCY, Y

	movsd	(Y), %xmm6

	addq	INCY, Y

	movsd	(Y), %xmm7

	addq	INCY, Y

	movsd	(X), %xmm8

	addq	INCX, X

	mulsd	ALPHA, %xmm0

	movsd	(X), %xmm9

	addq	INCX, X

	mulsd	ALPHA, %xmm1

	movsd	(X), %xmm10

	addq	INCX, X

	mulsd	ALPHA, %xmm2

	movsd	(X), %xmm11

	addq	INCX, X

	mulsd	ALPHA, %xmm3

	decq	%rax

	jle .L22

	ALIGN_3

.L21:

	addsd	%xmm4, %xmm0

	movsd	(Y), %xmm4

	addq	INCY, Y

	addsd	%xmm5, %xmm1

	movsd	(Y), %xmm5

	addq	INCY, Y

	addsd	%xmm6, %xmm2

	movsd	(Y), %xmm6

	addq	INCY, Y

	addsd	%xmm7, %xmm3

	movsd	(Y), %xmm7

	addq	INCY, Y

	movsd	%xmm0, (YY)

	addq	INCY, YY

	movsd	(X), %xmm0

	addq	INCX, X

	mulsd	ALPHA, %xmm8

	movsd	%xmm1, (YY)

	addq	INCY, YY

	movsd	(X), %xmm1

	addq	INCX, X

	mulsd	ALPHA, %xmm9

	movsd	%xmm2, (YY)

	addq	INCY, YY

	movsd	(X), %xmm2

	addq	INCX, X

	mulsd	ALPHA, %xmm10

	movsd	%xmm3, (YY)

	addq	INCY, YY

	movsd	(X), %xmm3

	addq	INCX, X

	mulsd	ALPHA, %xmm11

	addsd	%xmm4, %xmm8

	movsd	(Y), %xmm4

	addq	INCY, Y

	addsd	%xmm5, %xmm9

	movsd	(Y), %xmm5

	addq	INCY, Y

	addsd	%xmm6, %xmm10

	movsd	(Y), %xmm6

	addq	INCY, Y

	addsd	%xmm7, %xmm11

	movsd	(Y), %xmm7

	addq	INCY, Y

	movsd	%xmm8,  (YY)

	addq	INCY, YY

	movsd	(X), %xmm8

	addq	INCX, X

	mulsd	ALPHA, %xmm0

	movsd	%xmm9,  (YY)

	addq	INCY, YY

	movsd	(X), %xmm9

	addq	INCX, X

	mulsd	ALPHA, %xmm1

	movsd	%xmm10, (YY)

	addq	INCY, YY

	movsd	(X), %xmm10

	addq	INCX, X

	mulsd	ALPHA, %xmm2

	movsd	%xmm11, (YY)

	addq	INCY, YY

	movsd	(X), %xmm11

	addq	INCX, X

	mulsd	ALPHA, %xmm3

	decq	%rax

	jg	.L21

	ALIGN_3

.L22:

	addsd	%xmm4, %xmm0

	movsd	(Y), %xmm4

	addq	INCY, Y

	addsd	%xmm5, %xmm1

	movsd	(Y), %xmm5

	addq	INCY, Y

	addsd	%xmm6, %xmm2

	movsd	(Y), %xmm6

	addq	INCY, Y

	addsd	%xmm7, %xmm3

	movsd	(Y), %xmm7

	addq	INCY, Y

	movsd	%xmm0, (YY)

	addq	INCY, YY

	mulsd	ALPHA, %xmm8

	movsd	%xmm1, (YY)

	addq	INCY, YY

	mulsd	ALPHA, %xmm9

	movsd	%xmm2, (YY)

	addq	INCY, YY

	mulsd	ALPHA, %xmm10

	movsd	%xmm3, (YY)

	addq	INCY, YY

	mulsd	ALPHA, %xmm11

	addsd	%xmm4, %xmm8

	addsd	%xmm5, %xmm9

	addsd	%xmm6, %xmm10

	addsd	%xmm7, %xmm11

	movsd	%xmm8,  (YY)

	addq	INCY, YY

	movsd	%xmm9,  (YY)

	addq	INCY, YY

	movsd	%xmm10, (YY)

	addq	INCY, YY

	movsd	%xmm11, (YY)

	addq	INCY, YY

	ALIGN_3

.L23:

	movq	M,  %rax

	andq	$4, %rax

	jle	.L25

	ALIGN_3

	movsd	(X), %xmm0

	addq	INCX, X

	movsd	(Y), %xmm4

	addq	INCY, Y

	movsd	(X), %xmm1

	addq	INCX, X

	movsd	(Y), %xmm5

	addq	INCY, Y

	movsd	(X), %xmm2

	addq	INCX, X

	mulsd	ALPHA, %xmm0

	movsd	(Y), %xmm6

	addq	INCY, Y

	mulsd	ALPHA, %xmm1

	movsd	(X), %xmm3

	addq	INCX, X

	mulsd	ALPHA, %xmm2

	movsd	(Y), %xmm7

	addq	INCY, Y

	mulsd	ALPHA, %xmm3

	addsd	%xmm4, %xmm0

	addsd	%xmm5, %xmm1

	addsd	%xmm6, %xmm2

	addsd	%xmm7, %xmm3

	movsd	%xmm0, (YY)

	addq	INCY, YY

	movsd	%xmm1, (YY)

	addq	INCY, YY

	movsd	%xmm2, (YY)

	addq	INCY, YY

	movsd	%xmm3, (YY)

	addq	INCY, YY

	ALIGN_3

.L25:

	movq	M,  %rax

	andq	$2, %rax

	jle	.L26

	ALIGN_3

	movsd	(X), %xmm0

	addq	INCX, X

	movsd	(Y), %xmm4

	addq	INCY, Y

	movsd	(X), %xmm1

	addq	INCX, X

	movsd	(Y), %xmm5

	addq	INCY, Y

	mulsd	ALPHA, %xmm0

	mulsd	ALPHA, %xmm1

	addsd	%xmm4, %xmm0

	addsd	%xmm5, %xmm1

	movsd	%xmm0, (YY)

	addq	INCY, YY

	movsd	%xmm1, (YY)

	addq	INCY, YY

	ALIGN_3

.L26:

	movq	M,  %rax

	andq	$1, %rax

	jle	.L29

	ALIGN_3

	movsd	(X), %xmm0

	mulsd	ALPHA, %xmm0

	addsd	(Y), %xmm0

	movsd	%xmm0, 	(YY)

	addq	$SIZE, Y

	ALIGN_3

.L29:

	xorq	%rax, %rax

	RESTOREREGISTERS

	ret

	EPILOGUE