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

/* 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

#else

#define M	ARG1

#define X	ARG2

#define INCX	ARG3

#endif

#define XX	%r10

#define I	%rax

#include "l1param.h"

	PROLOGUE

	PROFCODE

#ifdef WINDOWS_ABI

	movq	40(%rsp), X

	movq	48(%rsp), INCX

	movaps	%xmm3, %xmm0	

#endif

	SAVEREGISTERS

	testq	M, M

	jle	.L999

	leaq	(, INCX, SIZE), INCX

	xorps	%xmm1, %xmm1

	comisd	%xmm0, %xmm1

	jne	.L100		# Alpha != ZERO

	jp	.L100		# For Alpha = NaN

/* Alpha == ZERO */

	cmpq	$SIZE, INCX

	jne	.L50

/* INCX == 1 */

	testq	$15, X		# aligned for quad word?

	je	.L05

	movsd	%xmm1, 0 * SIZE(X)

	addq	$SIZE, X

	decq	M

	jle	.L999

	ALIGN_3

.L05:

/* Aligned Mode */

	movq	M,  I	# rcx = n

	sarq	$4, I

	jle	.L12

	ALIGN_4

.L11:

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	movaps	%xmm1,  0 * SIZE(X)

	movaps	%xmm1,  2 * SIZE(X)

	movaps	%xmm1,  4 * SIZE(X)

	movaps	%xmm1,  6 * SIZE(X)

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X)

#endif

	movaps	%xmm1,  8 * SIZE(X)

	movaps	%xmm1, 10 * SIZE(X)

	movaps	%xmm1, 12 * SIZE(X)

	movaps	%xmm1, 14 * SIZE(X)

	addq	$16 * SIZE, X

	decq	I

	jg	.L11

	ALIGN_4

.L12:

	testq	$15, M

	je	.L999

	testq	$8, M

	je	.L13

	movaps	%xmm1,  0 * SIZE(X)

	movaps	%xmm1,  2 * SIZE(X)

	movaps	%xmm1,  4 * SIZE(X)

	movaps	%xmm1,  6 * SIZE(X)

	addq	$8 * SIZE, X

	ALIGN_3

.L13:

	testq  $4, M

	je    .L14

	movaps	%xmm1,  0 * SIZE(X)

	movaps	%xmm1,  2 * SIZE(X)

	addq	$4 * SIZE, X

	ALIGN_3

.L14:

	testq  $2, M

	je    .L15

	movaps	%xmm1,  0 * SIZE(X)

	addq	$2 * SIZE, X

	ALIGN_3

.L15:

	testq  $1, M

	je    .L999

	movsd	%xmm1,  0 * SIZE(X)

	jmp	.L999

	ALIGN_4

.L50:

	movq	M,  I

	sarq	$3, I

	jle	.L52

	ALIGN_4

.L51:

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	decq	I

	jg	.L51

	ALIGN_4

.L52:

	testq	$7, M

	je	.L999

	testq	$4, M

	je	.L53

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	ALIGN_3

.L53:

	testq	$2, M

	je	.L54

	movsd	%xmm1, (X)

	addq	INCX, X

	movsd	%xmm1, (X)

	addq	INCX, X

	ALIGN_3

.L54:

	testq	$1, M

	je	.L999

	movsd	%xmm1, (X)

	jmp	.L999

	ALIGN_4

/* Alpha != ZERO */

.L100:

	unpcklpd %xmm0, %xmm0

	cmpq	$SIZE, INCX

	jne	.L150

	testq	$SIZE, X

	je	.L105

	movsd	0 * SIZE(X), %xmm1

	mulsd	%xmm0, %xmm1

	movsd	%xmm1, 0 * SIZE(X)

	addq	$SIZE, X

	decq	M

	jle	.L999

	ALIGN_3

.L105:

	subq	$-16 * SIZE, X

	movq	M,  I		# rcx = n

	sarq	$4, I

	jle	.L113

#if defined(BARCELONA) || defined(SHANGHAI)

	movaps  %xmm0, %xmm1

	mulpd	-16 * SIZE(X), %xmm1

	movaps  %xmm0, %xmm2

	mulpd	-14 * SIZE(X), %xmm2

	movaps  %xmm0, %xmm3

	mulpd	-12 * SIZE(X), %xmm3

	movaps  %xmm0, %xmm4

	mulpd	-10 * SIZE(X), %xmm4

	movaps  %xmm0, %xmm5

	mulpd	 -8 * SIZE(X), %xmm5

	movaps  %xmm0, %xmm6

	mulpd	 -6 * SIZE(X), %xmm6

	movaps  %xmm0, %xmm7

	mulpd	 -4 * SIZE(X), %xmm7

	movaps  %xmm0, %xmm8

	mulpd	 -2 * SIZE(X), %xmm8

	decq	I 

	jle	.L112

	ALIGN_4

.L111:

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	movaps	%xmm1, -16 * SIZE(X)

	movaps	%xmm2, -14 * SIZE(X)

	movaps	%xmm3, -12 * SIZE(X)

	movaps	%xmm4, -10 * SIZE(X)

	movaps  %xmm0, %xmm1

	mulpd	  0 * SIZE(X), %xmm1

	movaps  %xmm0, %xmm2

	mulpd	  2 * SIZE(X), %xmm2

	movaps  %xmm0, %xmm3

	mulpd	  4 * SIZE(X), %xmm3

	movaps  %xmm0, %xmm4

	mulpd	  6 * SIZE(X), %xmm4

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X)

#endif

	movaps	%xmm5,  -8 * SIZE(X)

	movaps	%xmm6,  -6 * SIZE(X)

	movaps	%xmm7,  -4 * SIZE(X)

	movaps	%xmm8,  -2 * SIZE(X)

	movaps  %xmm0, %xmm5

	mulpd	  8 * SIZE(X), %xmm5

	movaps  %xmm0, %xmm6

	mulpd	 10 * SIZE(X), %xmm6

	movaps  %xmm0, %xmm7

	mulpd	 12 * SIZE(X), %xmm7

	movaps  %xmm0, %xmm8

	mulpd	 14 * SIZE(X), %xmm8

	subq	$-16 * SIZE, X

	decq	I

	jg	.L111

	ALIGN_4

.L112:

	movaps	%xmm1, -16 * SIZE(X)

	movaps	%xmm2, -14 * SIZE(X)

	movaps	%xmm3, -12 * SIZE(X)

	movaps	%xmm4, -10 * SIZE(X)

	movaps	%xmm5,  -8 * SIZE(X)

	movaps	%xmm6,  -6 * SIZE(X)

	movaps	%xmm7,  -4 * SIZE(X)

	movaps	%xmm8,  -2 * SIZE(X)

#else

	movaps	-16 * SIZE(X), %xmm1

	movaps	-14 * SIZE(X), %xmm2

	movaps	-12 * SIZE(X), %xmm3

	movaps	-10 * SIZE(X), %xmm4

	movaps	 -8 * SIZE(X), %xmm5

	movaps	 -6 * SIZE(X), %xmm6

	movaps	 -4 * SIZE(X), %xmm7

	movaps	 -2 * SIZE(X), %xmm8

	decq	I 

	jle	.L112

	ALIGN_4

.L111:

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	mulpd	%xmm0, %xmm1

	movaps	%xmm1, -16 * SIZE(X)

	movaps	 0 * SIZE(X), %xmm1

	mulpd	%xmm0, %xmm2

	movaps	%xmm2, -14 * SIZE(X)

	movaps	 2 * SIZE(X), %xmm2

	mulpd	%xmm0, %xmm3

	movaps	%xmm3, -12 * SIZE(X)

	movaps	 4  * SIZE(X), %xmm3

	mulpd	%xmm0, %xmm4

	movaps	%xmm4, -10 * SIZE(X)

	movaps	 6 * SIZE(X), %xmm4

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X)

#endif

	mulpd	%xmm0, %xmm5

	movaps	%xmm5, -8 * SIZE(X)

	movaps	 8 * SIZE(X), %xmm5

	mulpd	%xmm0, %xmm6

	movaps	%xmm6, -6 * SIZE(X)

	movaps	10 * SIZE(X), %xmm6

	mulpd	%xmm0, %xmm7

	movaps	%xmm7, -4 * SIZE(X)

	movaps	12 * SIZE(X), %xmm7

	mulpd	%xmm0, %xmm8

	movaps	%xmm8, -2 * SIZE(X)

	movaps	14 * SIZE(X), %xmm8

	subq	$-16 * SIZE, X

	decq	I

	jg	.L111

	ALIGN_4

.L112:

	mulpd	%xmm0, %xmm1

	movaps	%xmm1, -16 * SIZE(X)

	mulpd	%xmm0, %xmm2

	movaps	%xmm2, -14 * SIZE(X)

	mulpd	%xmm0, %xmm3

	movaps	%xmm3, -12 * SIZE(X)

	mulpd	%xmm0, %xmm4

	movaps	%xmm4, -10 * SIZE(X)

	mulpd	%xmm0, %xmm5

	movaps	%xmm5,  -8 * SIZE(X)

	mulpd	%xmm0, %xmm6

	movaps	%xmm6,  -6 * SIZE(X)

	mulpd	%xmm0, %xmm7

	movaps	%xmm7,  -4 * SIZE(X)

	mulpd	%xmm0, %xmm8

	movaps	%xmm8,  -2 * SIZE(X)

#endif

	subq	$-16 * SIZE, X

	ALIGN_3

.L113:

	testq	$15, M

	je	.L999

	testq	$8, M

	je	.L114

	movaps	-16 * SIZE(X), %xmm1

	movaps	-14 * SIZE(X), %xmm2

	movaps	-12 * SIZE(X), %xmm3

	movaps	-10 * SIZE(X), %xmm4

	mulpd	%xmm0, %xmm1

	movaps	%xmm1, -16 * SIZE(X)

	mulpd	%xmm0, %xmm2

	movaps	%xmm2, -14 * SIZE(X)

	mulpd	%xmm0, %xmm3

	movaps	%xmm3, -12 * SIZE(X)

	mulpd	%xmm0, %xmm4

	movaps	%xmm4, -10 * SIZE(X)

	addq	$8 * SIZE, X

	ALIGN_3

.L114:

	testq	$4, M

	je	.L115

	movaps	-16 * SIZE(X), %xmm1

	movaps	-14 * SIZE(X), %xmm2

	mulpd	%xmm0, %xmm1

	movaps	%xmm1, -16 * SIZE(X)

	mulpd	%xmm0, %xmm2

	movaps	%xmm2, -14 * SIZE(X)

	addq	$4 * SIZE, X

	ALIGN_3

.L115:

	testq	$2, M

	je	.L116

	movaps	-16 * SIZE(X), %xmm1

	mulpd	%xmm0, %xmm1

	movaps	%xmm1, -16 * SIZE(X)

	addq	$2 * SIZE, X

	ALIGN_3

.L116:

	testq	$1, M

	je	.L999

	movsd	-16 * SIZE(X), %xmm1

	mulsd	%xmm0, %xmm1

	movsd	%xmm1, -16 * SIZE(X)

	jmp	.L999

	ALIGN_3

/* incx != 1 */

.L150:

	movq	X, XX

	movq	M,  I		# rcx = n

	sarq	$3, I		# (n >> 3)

	jle	.L152

	ALIGN_4

.L151:

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X)

#endif

	movsd	(X), %xmm1

	addq	INCX, X

	movsd	(X), %xmm2

	addq	INCX, X

	movsd	(X), %xmm3

	addq	INCX, X

	movsd	(X), %xmm4

	addq	INCX, X

	movsd	(X), %xmm5

	addq	INCX, X

	movsd	(X), %xmm6

	addq	INCX, X

	movsd	(X), %xmm7

	addq	INCX, X

	movsd	(X), %xmm8

	addq	INCX, X

	mulsd	%xmm0, %xmm1

	mulsd	%xmm0, %xmm2

	mulsd	%xmm0, %xmm3

	mulsd	%xmm0, %xmm4

	mulsd	%xmm0, %xmm5

	mulsd	%xmm0, %xmm6

	mulsd	%xmm0, %xmm7

	mulsd	%xmm0, %xmm8

	movsd	%xmm1, (XX)

	addq	INCX, XX

	movsd	%xmm2, (XX)

	addq	INCX, XX

	movsd	%xmm3, (XX)

	addq	INCX, XX

	movsd	%xmm4, (XX)

	addq	INCX, XX

	movsd	%xmm5, (XX)

	addq	INCX, XX

	movsd	%xmm6, (XX)

	addq	INCX, XX

	movsd	%xmm7, (XX)

	addq	INCX, XX

	movsd	%xmm8, (XX)

	addq	INCX, XX

	decq	I

	jg	.L151

	ALIGN_4

.L152:

	testq	$7, M

	je	.L999

	testq	$4, M

	je	.L153

	movsd	 (X), %xmm1

	addq	INCX, X

	movsd	 (X), %xmm2

	addq	INCX, X

	movsd	 (X), %xmm3

	addq	INCX, X

	movsd	 (X), %xmm4

	addq	INCX, X

	mulsd	%xmm0, %xmm1

	mulsd	%xmm0, %xmm2

	mulsd	%xmm0, %xmm3

	mulsd	%xmm0, %xmm4

	movsd	%xmm1,  (XX)

	addq	INCX, XX

	movsd	%xmm2,  (XX)

	addq	INCX, XX

	movsd	%xmm3,  (XX)

	addq	INCX, XX

	movsd	%xmm4,  (XX)

	addq	INCX, XX

	ALIGN_3

.L153:

	testq	$2, M

	je	.L154

	movsd	 (X), %xmm1

	addq	INCX, X

	movsd	 (X), %xmm2

	addq	INCX, X

	mulsd	%xmm0, %xmm1

	mulsd	%xmm0, %xmm2

	movsd	%xmm1,  (XX)

	addq	INCX, XX

	movsd	%xmm2,  (XX)

	addq	INCX, XX

	ALIGN_3

.L154:

	testq	$1, M

	je	.L999

	movsd	 (X), %xmm1

	mulsd	%xmm0, %xmm1

	movsd	%xmm1,  (X)

	ALIGN_4

.L999:

	xorq	%rax, %rax

	RESTOREREGISTERS

	ret

	EPILOGUE