/*********************************************************************/
/* 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"
#define M ARG1 /* rdi */
#define X ARG2 /* rsi */
#define INCX ARG3 /* rdx */
#define I %rax
#include "l1param.h"
PROLOGUE
PROFCODE
SAVEREGISTERS
xorps %xmm0, %xmm0
testq M, M
jle .L999
testq INCX, INCX
jle .L999
xorps %xmm1, %xmm1
xorps %xmm2, %xmm2
xorps %xmm3, %xmm3
pcmpeqb %xmm15, %xmm15
psrlq $1, %xmm15
salq $BASE_SHIFT, INCX
xorps %xmm13, %xmm13
cmpq $SIZE, INCX
jne .L20
testq $SIZE, X
je .L05
movsd (X), %xmm0
addq $SIZE, X
andps %xmm15, %xmm0
decq M
jle .L999
ALIGN_3
.L05:
subq $-16 * SIZE, X
movq M, I
sarq $4, I
jle .L12
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
movaps -12 * SIZE(X), %xmm6
movaps -10 * SIZE(X), %xmm7
movaps -8 * SIZE(X), %xmm8
movaps -6 * SIZE(X), %xmm9
movaps -4 * SIZE(X), %xmm10
movaps -2 * SIZE(X), %xmm11
decq I
jle .L11
ALIGN_4
.L10:
#ifdef PREFETCH
PREFETCH (PREFETCHSIZE + 0) - PREOFFSET(X)
#endif
andps %xmm15, %xmm4
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm4, %xmm12
addsd %xmm4, %xmm0
movaps 0 * SIZE(X), %xmm4
andps %xmm15, %xmm5
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm5, %xmm13
addsd %xmm5, %xmm2
movaps 2 * SIZE(X), %xmm5
andps %xmm15, %xmm6
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm6, %xmm12
addsd %xmm6, %xmm0
movaps 4 * SIZE(X), %xmm6
andps %xmm15, %xmm7
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm7, %xmm13
addsd %xmm7, %xmm2
movaps 6 * SIZE(X), %xmm7
#if defined(PREFETCH) && !defined(FETCH128)
PREFETCH (PREFETCHSIZE + 64) - PREOFFSET(X)
#endif
andps %xmm15, %xmm8
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm8, %xmm12
addsd %xmm8, %xmm0
movaps 8 * SIZE(X), %xmm8
andps %xmm15, %xmm9
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm9, %xmm13
addsd %xmm9, %xmm2
movaps 10 * SIZE(X), %xmm9
andps %xmm15, %xmm10
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm10, %xmm12
addsd %xmm10, %xmm0
movaps 12 * SIZE(X), %xmm10
andps %xmm15, %xmm11
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm11, %xmm13
addsd %xmm11, %xmm2
movaps 14 * SIZE(X), %xmm11
subq $-16 * SIZE, X
decq I
jg .L10
ALIGN_4
.L11:
andps %xmm15, %xmm4
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm4, %xmm12
addsd %xmm4, %xmm0
andps %xmm15, %xmm5
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm5, %xmm13
addsd %xmm5, %xmm2
andps %xmm15, %xmm6
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm6, %xmm12
addsd %xmm6, %xmm0
andps %xmm15, %xmm7
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm7, %xmm13
addsd %xmm7, %xmm2
andps %xmm15, %xmm8
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm8, %xmm12
addsd %xmm8, %xmm0
andps %xmm15, %xmm9
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm9, %xmm13
addsd %xmm9, %xmm2
andps %xmm15, %xmm10
addsd %xmm13, %xmm3
pshufd $0x4e, %xmm10, %xmm12
addsd %xmm10, %xmm0
andps %xmm15, %xmm11
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm11, %xmm13
addsd %xmm11, %xmm2
addsd %xmm13, %xmm3
subq $-16 * SIZE, X
ALIGN_3
.L12:
andq $15, M
jle .L998
testq $8, M
je .L13
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
movaps -12 * SIZE(X), %xmm6
movaps -10 * SIZE(X), %xmm7
addq $8 * SIZE, X
andps %xmm15, %xmm4
pshufd $0x4e, %xmm4, %xmm12
addsd %xmm4, %xmm0
andps %xmm15, %xmm5
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm5, %xmm13
addsd %xmm5, %xmm2
addsd %xmm13, %xmm3
andps %xmm15, %xmm6
pshufd $0x4e, %xmm6, %xmm12
addsd %xmm6, %xmm0
andps %xmm15, %xmm7
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm7, %xmm13
addsd %xmm7, %xmm2
addsd %xmm13, %xmm3
ALIGN_3
.L13:
testq $4, M
je .L14
movaps -16 * SIZE(X), %xmm4
movaps -14 * SIZE(X), %xmm5
addq $4 * SIZE, X
andps %xmm15, %xmm4
pshufd $0x4e, %xmm4, %xmm12
addsd %xmm4, %xmm0
andps %xmm15, %xmm5
addsd %xmm12, %xmm1
pshufd $0x4e, %xmm5, %xmm13
addsd %xmm5, %xmm2
addsd %xmm13, %xmm3
ALIGN_3
.L14:
testq $2, M
je .L15
movaps -16 * SIZE(X), %xmm4
addq $2 * SIZE, X
andps %xmm15, %xmm4
pshufd $0x4e, %xmm4, %xmm5
addsd %xmm4, %xmm2
addsd %xmm5, %xmm3
ALIGN_3
.L15:
testq $1, M
je .L998
movsd -16 * SIZE(X), %xmm4
andps %xmm15, %xmm4
addsd %xmm4, %xmm0
jmp .L998
ALIGN_3
.L20:
movq M, I
sarq $3, I
jle .L25
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
movsd (X), %xmm9
addq INCX, X
movsd (X), %xmm10
addq INCX, X
movsd (X), %xmm11
decq I
jle .L23
ALIGN_4
.L22:
andps %xmm15, %xmm4
addq INCX, X
addsd %xmm4, %xmm0
movsd (X), %xmm4
andps %xmm15, %xmm5
addq INCX, X
addsd %xmm5, %xmm1
movsd (X), %xmm5
andps %xmm15, %xmm6
addq INCX, X
addsd %xmm6, %xmm2
movsd (X), %xmm6
andps %xmm15, %xmm7
addq INCX, X
addsd %xmm7, %xmm3
movsd (X), %xmm7
andps %xmm15, %xmm8
addq INCX, X
addsd %xmm8, %xmm0
movsd (X), %xmm8
andps %xmm15, %xmm9
addq INCX, X
addsd %xmm9, %xmm1
movsd (X), %xmm9
andps %xmm15, %xmm10
addq INCX, X
addsd %xmm10, %xmm2
movsd (X), %xmm10
andps %xmm15, %xmm11
addq INCX, X
addsd %xmm11, %xmm3
movsd (X), %xmm11
decq I
jg .L22
ALIGN_4
.L23:
andps %xmm15, %xmm4
addq INCX, X
addsd %xmm4, %xmm0
andps %xmm15, %xmm5
addsd %xmm5, %xmm1
andps %xmm15, %xmm6
addsd %xmm6, %xmm2
andps %xmm15, %xmm7
addsd %xmm7, %xmm3
andps %xmm15, %xmm8
addsd %xmm8, %xmm0
andps %xmm15, %xmm9
addsd %xmm9, %xmm1
andps %xmm15, %xmm10
addsd %xmm10, %xmm2
andps %xmm15, %xmm11
addsd %xmm11, %xmm3
ALIGN_3
.L25:
andq $7, M
jle .L998
testq $4, M
je .L26
movsd (X), %xmm4
addq INCX, X
movsd (X), %xmm5
addq INCX, X
movsd (X), %xmm6
andps %xmm15, %xmm4
addsd %xmm4, %xmm0
addq INCX, X
movsd (X), %xmm7
andps %xmm15, %xmm5
addsd %xmm5, %xmm1
addq INCX, X
andps %xmm15, %xmm6
addsd %xmm6, %xmm2
andps %xmm15, %xmm7
addsd %xmm7, %xmm3
ALIGN_3
.L26:
testq $2, M
je .L27
movsd (X), %xmm4
addq INCX, X
movsd (X), %xmm5
addq INCX, X
andps %xmm15, %xmm4
andps %xmm15, %xmm5
addsd %xmm4, %xmm0
addsd %xmm5, %xmm1
ALIGN_3
.L27:
testq $1, M
je .L998
movsd (X), %xmm4
andps %xmm15, %xmm4
addsd %xmm4, %xmm0
ALIGN_3
.L998:
addsd %xmm1, %xmm0
addsd %xmm3, %xmm2
addsd %xmm2, %xmm0
ALIGN_4
.L999:
RESTOREREGISTERS
ret
EPILOGUE