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

/* 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_R	%xmm14

#define ALPHA_I	%xmm15

#include "l1param.h"

	PROLOGUE

	PROFCODE

#ifndef WINDOWS_ABI

#ifndef XDOUBLE

	movq	 8(%rsp), INCY

#else

	movq	40(%rsp), INCY

#endif

#else

	movaps	%xmm3, %xmm0

	movsd	40(%rsp), %xmm1	

	movq	48(%rsp), X

	movq	56(%rsp), INCX

	movq	64(%rsp), Y

	movq	72(%rsp), INCY

#endif

	SAVEREGISTERS

#ifndef CONJ

#define ADD1	subsd

#define ADD2	addsd

#else

#define ADD1	addsd

#define ADD2	subsd

#endif

	salq	$ZBASE_SHIFT, INCX

	movaps	%xmm0, ALPHA_R

	salq	$ZBASE_SHIFT, INCY

	movaps	%xmm1, ALPHA_I

	testq	M, M

	jle	.L999

	cmpq	$2 * SIZE, INCX

	jne	.L20

	cmpq	$2 * SIZE, INCY

	jne	.L20

	movq	M,  %rax

	sarq	$2, %rax

	jle	.L15

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	movsd	 0 * SIZE(Y), %xmm8

	movsd	 1 * SIZE(Y), %xmm9

	movsd	 2 * SIZE(X), %xmm4

	movsd	 3 * SIZE(X), %xmm5

	movsd	 2 * SIZE(Y), %xmm10

	movsd	 3 * SIZE(Y), %xmm11

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	mulsd	ALPHA_I, %xmm3

	mulsd	ALPHA_I, %xmm2

	movaps	 %xmm4,  %xmm6

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movsd	 4 * SIZE(X), %xmm0

	movaps	 %xmm5,  %xmm7

	mulsd	ALPHA_R, %xmm5

	ADD2	%xmm1, %xmm9

	movsd	 5 * SIZE(X), %xmm1

	decq	%rax

	jle	.L12

	ALIGN_3

.L11:

#ifdef PREFETCH

	PREFETCH (PREFETCHSIZE +  0) - PREOFFSET(X)

#endif

	mulsd	ALPHA_I, %xmm7

	movsd	 4 * SIZE(Y), %xmm12

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	movsd	 5 * SIZE(Y), %xmm13

	addsd	%xmm2, %xmm9

	addsd	%xmm4, %xmm10

	movsd	 6 * SIZE(X), %xmm4

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	ADD2	%xmm5, %xmm11

	movsd	 7 * SIZE(X), %xmm5

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD1	%xmm7, %xmm10

	movsd	%xmm8,   0 * SIZE(Y)

	mulsd	ALPHA_I, %xmm3

	addsd	%xmm6, %xmm11

	movsd	%xmm9,   1 * SIZE(Y)

	mulsd	ALPHA_I, %xmm2

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  2 * SIZE(Y)

	mulsd	ALPHA_R, %xmm4

	movsd	 6 * SIZE(Y), %xmm10

	addsd	%xmm0, %xmm12

	movsd	 8 * SIZE(X), %xmm0

#ifdef PREFETCHW

	PREFETCHW (PREFETCHSIZE +  0) - PREOFFSET(Y)

#endif

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  3 * SIZE(Y)

	mulsd	ALPHA_R, %xmm5

	movsd	 7 * SIZE(Y), %xmm11

	ADD2	%xmm1, %xmm13

	movsd	 9 * SIZE(X), %xmm1

	mulsd	ALPHA_I, %xmm7

	movsd	 8 * SIZE(Y), %xmm8

	ADD1	%xmm3, %xmm12

	mulsd	ALPHA_I, %xmm6

	movsd	 9 * SIZE(Y), %xmm9

	addsd	%xmm2, %xmm13

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	addsd	%xmm4, %xmm10

	movsd	10 * SIZE(X), %xmm4

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD2	%xmm5, %xmm11

	movsd	11 * SIZE(X), %xmm5

	mulsd	ALPHA_I, %xmm3

	movsd	%xmm12,   4 * SIZE(Y)

	ADD1	%xmm7, %xmm10

	mulsd	ALPHA_I, %xmm2

	movsd	%xmm13,   5 * SIZE(Y)

	addsd	%xmm6, %xmm11

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  6 * SIZE(Y)

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movsd	10 * SIZE(Y), %xmm10

	movsd	12 * SIZE(X), %xmm0

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  7 * SIZE(Y)

	mulsd	ALPHA_R, %xmm5

	movsd	11 * SIZE(Y), %xmm11

	ADD2	%xmm1, %xmm9

	movsd	13 * SIZE(X), %xmm1

	addq	$8 * SIZE, X

	addq	$8 * SIZE, Y

	decq	%rax

	jg	.L11

	ALIGN_3

.L12:

	mulsd	ALPHA_I, %xmm7

	movsd	 4 * SIZE(Y), %xmm12

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	movsd	 5 * SIZE(Y), %xmm13

	addsd	%xmm2, %xmm9

	addsd	%xmm4, %xmm10

	movsd	 6 * SIZE(X), %xmm4

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	ADD2	%xmm5, %xmm11

	movsd	 7 * SIZE(X), %xmm5

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD1	%xmm7, %xmm10

	movsd	%xmm8,   0 * SIZE(Y)

	mulsd	ALPHA_I, %xmm3

	addsd	%xmm6, %xmm11

	movsd	%xmm9,   1 * SIZE(Y)

	mulsd	ALPHA_I, %xmm2

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  2 * SIZE(Y)

	mulsd	ALPHA_R, %xmm4

	movsd	 6 * SIZE(Y), %xmm10

	addsd	%xmm0, %xmm12

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  3 * SIZE(Y)

	mulsd	ALPHA_R, %xmm5

	ADD2	%xmm1, %xmm13

	movsd	 7 * SIZE(Y), %xmm11

	mulsd	ALPHA_I, %xmm7

	ADD1	%xmm3, %xmm12

	mulsd	ALPHA_I, %xmm6

	addsd	%xmm2, %xmm13

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	addsd	%xmm4, %xmm10

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD2	%xmm5, %xmm11

	mulsd	ALPHA_I, %xmm3

	ADD1	%xmm7, %xmm10

	addsd	%xmm6, %xmm11

	mulsd	ALPHA_I, %xmm2

	movsd	%xmm12,  4 * SIZE(Y)

	movsd	%xmm13,  5 * SIZE(Y)

	movsd	%xmm10,  6 * SIZE(Y)

	movsd	%xmm11,  7 * SIZE(Y)

	addq	$8 * SIZE, X

	addq	$8 * SIZE, Y

	ALIGN_3

.L15:

	movq	M,  %rax

	andq	$2, %rax

	jle	.L17

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	movsd	 2 * SIZE(X), %xmm4

	movsd	 3 * SIZE(X), %xmm5

	movaps	 %xmm0,  %xmm2

	movsd	 0 * SIZE(Y), %xmm8

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	movsd	 1 * SIZE(Y), %xmm9

	mulsd	ALPHA_R, %xmm1

	movsd	 2 * SIZE(Y), %xmm10

	mulsd	ALPHA_I, %xmm3

	movsd	 3 * SIZE(Y), %xmm11

	mulsd	ALPHA_I, %xmm2

	movaps	 %xmm4,  %xmm6

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movaps	 %xmm5,  %xmm7

	mulsd	ALPHA_R, %xmm5

	ADD2	%xmm1, %xmm9

	mulsd	ALPHA_I, %xmm7

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	addsd	%xmm2, %xmm9

	addsd	%xmm4, %xmm10

	movsd	%xmm8,   0 * SIZE(Y)

	ADD2	%xmm5, %xmm11

	movsd	%xmm9,   1 * SIZE(Y)

	ADD1	%xmm7, %xmm10

	addsd	%xmm6, %xmm11

	movsd	%xmm10,   2 * SIZE(Y)

	movsd	%xmm11,   3 * SIZE(Y)

	addq	$4 * SIZE, X

	addq	$4 * SIZE, Y

	ALIGN_3

.L17:

	movq	M,  %rax

	andq	$1, %rax

	jle	.L999

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	movsd	 0 * SIZE(Y), %xmm8

	movsd	 1 * SIZE(Y), %xmm9

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	mulsd	ALPHA_I, %xmm3

	mulsd	ALPHA_I, %xmm2

	addsd	%xmm0, %xmm8

	ADD2	%xmm1, %xmm9

	ADD1	%xmm3, %xmm8

	addsd	%xmm2, %xmm9

	movsd	%xmm8,   0 * SIZE(Y)

	movsd	%xmm9,   1 * SIZE(Y)

	jmp	.L999

	ALIGN_3

.L20:

	movq	Y, YY

	movq	M,  %rax

	sarq	$2, %rax

	jle	.L25

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	addq	 INCX, X

	movsd	 0 * SIZE(Y), %xmm8

	movsd	 1 * SIZE(Y), %xmm9

	addq	 INCY, Y

	movsd	 0 * SIZE(X), %xmm4

	movsd	 1 * SIZE(X), %xmm5

	addq	 INCX, X

	movsd	 0 * SIZE(Y), %xmm10

	movsd	 1 * SIZE(Y), %xmm11

	addq	 INCY, Y

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	mulsd	ALPHA_I, %xmm3

	mulsd	ALPHA_I, %xmm2

	movaps	 %xmm4,  %xmm6

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movsd	 0 * SIZE(X), %xmm0

	movaps	 %xmm5,  %xmm7

	mulsd	ALPHA_R, %xmm5

	ADD2	%xmm1, %xmm9

	movsd	 1 * SIZE(X), %xmm1

	addq	 INCX, X

	decq	%rax

	jle	.L22

	ALIGN_3

.L21:

	mulsd	ALPHA_I, %xmm7

	movsd	 0 * SIZE(Y), %xmm12

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	movsd	 1 * SIZE(Y), %xmm13

	addsd	%xmm2, %xmm9

	addq	 INCY, Y

	addsd	%xmm4, %xmm10

	movsd	 0 * SIZE(X), %xmm4

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	ADD2	%xmm5, %xmm11

	movsd	 1 * SIZE(X), %xmm5

	movaps	 %xmm1,  %xmm3

	addq	 INCX, X

	mulsd	ALPHA_R, %xmm1

	ADD1	%xmm7, %xmm10

	movsd	%xmm8,   0 * SIZE(YY)

	mulsd	ALPHA_I, %xmm3

	addsd	%xmm6, %xmm11

	movsd	%xmm9,   1 * SIZE(YY)

	mulsd	ALPHA_I, %xmm2

	addq	 INCY, YY

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  0 * SIZE(YY)

	mulsd	ALPHA_R, %xmm4

	movsd	 0 * SIZE(Y), %xmm10

	addsd	%xmm0, %xmm12

	movsd	 0 * SIZE(X), %xmm0

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  1 * SIZE(YY)

	addq	 INCY, YY

	mulsd	ALPHA_R, %xmm5

	movsd	 1 * SIZE(Y), %xmm11

	addq	 INCY, Y

	ADD2	%xmm1, %xmm13

	movsd	 1 * SIZE(X), %xmm1

	addq	 INCX, X

	mulsd	ALPHA_I, %xmm7

	movsd	 0 * SIZE(Y), %xmm8

	ADD1	%xmm3, %xmm12

	mulsd	ALPHA_I, %xmm6

	movsd	 1 * SIZE(Y), %xmm9

	addsd	%xmm2, %xmm13

	addq	 INCY, Y

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	addsd	%xmm4, %xmm10

	movsd	 0 * SIZE(X), %xmm4

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD2	%xmm5, %xmm11

	movsd	 1 * SIZE(X), %xmm5

	addq	 INCX, X

	mulsd	ALPHA_I, %xmm3

	movsd	%xmm12,   0 * SIZE(YY)

	ADD1	%xmm7, %xmm10

	mulsd	ALPHA_I, %xmm2

	movsd	%xmm13,   1 * SIZE(YY)

	addsd	%xmm6, %xmm11

	addq	 INCY, YY

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  0 * SIZE(YY)

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movsd	 0 * SIZE(Y), %xmm10

	movsd	 0 * SIZE(X), %xmm0

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  1 * SIZE(YY)

	addq	 INCY, YY

	mulsd	ALPHA_R, %xmm5

	movsd	 1 * SIZE(Y), %xmm11

	addq	 INCY, Y

	ADD2	%xmm1, %xmm9

	movsd	 1 * SIZE(X), %xmm1

	addq	 INCX, X

	decq	%rax

	jg	.L21

	ALIGN_3

.L22:

	mulsd	ALPHA_I, %xmm7

	movsd	 0 * SIZE(Y), %xmm12

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	movsd	 1 * SIZE(Y), %xmm13

	addsd	%xmm2, %xmm9

	addq	 INCY, Y

	addsd	%xmm4, %xmm10

	movsd	 0 * SIZE(X), %xmm4

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	ADD2	%xmm5, %xmm11

	movsd	 1 * SIZE(X), %xmm5

	movaps	 %xmm1,  %xmm3

	addq	 INCX, X

	mulsd	ALPHA_R, %xmm1

	ADD1	%xmm7, %xmm10

	movsd	%xmm8,   0 * SIZE(YY)

	mulsd	ALPHA_I, %xmm3

	addsd	%xmm6, %xmm11

	movsd	%xmm9,   1 * SIZE(YY)

	mulsd	ALPHA_I, %xmm2

	addq	 INCY, YY

	movaps	 %xmm4,  %xmm6

	movsd	%xmm10,  0 * SIZE(YY)

	mulsd	ALPHA_R, %xmm4

	movsd	 0 * SIZE(Y), %xmm10

	addsd	%xmm0, %xmm12

	movaps	 %xmm5,  %xmm7

	movsd	%xmm11,  1 * SIZE(YY)

	mulsd	ALPHA_R, %xmm5

	addq	 INCY, YY

	ADD2	%xmm1, %xmm13

	movsd	 1 * SIZE(Y), %xmm11

	mulsd	ALPHA_I, %xmm7

	addq	 INCY, Y

	ADD1	%xmm3, %xmm12

	mulsd	ALPHA_I, %xmm6

	addsd	%xmm2, %xmm13

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	addsd	%xmm4, %xmm10

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	ADD2	%xmm5, %xmm11

	mulsd	ALPHA_I, %xmm3

	ADD1	%xmm7, %xmm10

	addsd	%xmm6, %xmm11

	mulsd	ALPHA_I, %xmm2

	movsd	%xmm12,  0 * SIZE(YY)

	movsd	%xmm13,  1 * SIZE(YY)

	addq	 INCY, YY

	movsd	%xmm10,  0 * SIZE(YY)

	movsd	%xmm11,  1 * SIZE(YY)

	addq	 INCY, YY

	ALIGN_3

.L25:

	movq	M,  %rax

	andq	$2, %rax

	jle	.L27

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	addq	 INCX, X

	movsd	 0 * SIZE(X), %xmm4

	movsd	 1 * SIZE(X), %xmm5

	addq	 INCX, X

	movaps	 %xmm0,  %xmm2

	movsd	 0 * SIZE(Y), %xmm8

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	movsd	 1 * SIZE(Y), %xmm9

	addq	 INCY, Y

	mulsd	ALPHA_R, %xmm1

	movsd	 0 * SIZE(Y), %xmm10

	mulsd	ALPHA_I, %xmm3

	movsd	 1 * SIZE(Y), %xmm11

	mulsd	ALPHA_I, %xmm2

	addq	 INCY, Y

	movaps	 %xmm4,  %xmm6

	mulsd	ALPHA_R, %xmm4

	addsd	%xmm0, %xmm8

	movaps	 %xmm5,  %xmm7

	mulsd	ALPHA_R, %xmm5

	ADD2	%xmm1, %xmm9

	mulsd	ALPHA_I, %xmm7

	ADD1	%xmm3, %xmm8

	mulsd	ALPHA_I, %xmm6

	addsd	%xmm2, %xmm9

	addsd	%xmm4, %xmm10

	movsd	%xmm8,   0 * SIZE(YY)

	ADD2	%xmm5, %xmm11

	movsd	%xmm9,   1 * SIZE(YY)

	ADD1	%xmm7, %xmm10

	addq	 INCY, YY

	addsd	%xmm6, %xmm11

	movsd	%xmm10,  0 * SIZE(YY)

	movsd	%xmm11,  1 * SIZE(YY)

	addq	 INCY, YY

	ALIGN_3

.L27:

	movq	M,  %rax

	andq	$1, %rax

	jle	.L999

	movsd	 0 * SIZE(X), %xmm0

	movsd	 1 * SIZE(X), %xmm1

	movsd	 0 * SIZE(Y), %xmm8

	movsd	 1 * SIZE(Y), %xmm9

	movaps	 %xmm0,  %xmm2

	mulsd	ALPHA_R, %xmm0

	movaps	 %xmm1,  %xmm3

	mulsd	ALPHA_R, %xmm1

	mulsd	ALPHA_I, %xmm3

	mulsd	ALPHA_I, %xmm2

	addsd	%xmm0, %xmm8

	ADD2	%xmm1, %xmm9

	ADD1	%xmm3, %xmm8

	addsd	%xmm2, %xmm9

	movsd	%xmm8,   0 * SIZE(YY)

	movsd	%xmm9,   1 * SIZE(YY)

	ALIGN_3

.L999:

	xorq	%rax, %rax

	RESTOREREGISTERS

	ret

	EPILOGUE