/*********************************************************************/
/* 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 STACK 16
#define ARGS 16
#define M 4 + STACK + ARGS(%esp)
#define N 8 + STACK + ARGS(%esp)
#define K 12 + STACK + ARGS(%esp)
#define ALPHA_R 16 + STACK + ARGS(%esp)
#define ALPHA_I 24 + STACK + ARGS(%esp)
#define A 32 + STACK + ARGS(%esp)
#define ARG_B 36 + STACK + ARGS(%esp)
#define C 40 + STACK + ARGS(%esp)
#define ARG_LDC 44 + STACK + ARGS(%esp)
#define J 0 + STACK(%esp)
#define BX 4 + STACK(%esp)
#define KK 8 + STACK(%esp)
#define KKK 12 + STACK(%esp)
#define PREFETCH prefetcht0
#define PREFETCHSIZE 84
#define AA %edx
#define BB %ecx
#define CO1 %esi
#define LDC %ebp
#define B %edi
PROLOGUE
subl $ARGS, %esp
pushl %ebp
pushl %edi
pushl %esi
pushl %ebx
PROFCODE
movl ARG_B, B
movl ARG_LDC, LDC
#ifdef TRMMKERNEL
movl OFFSET, %eax
#ifndef LEFT
negl %eax
#endif
movl %eax, KK
#endif
sall $ZBASE_SHIFT, LDC
movl N, %eax
sarl $1, %eax
movl %eax, J
jle .L30
ALIGN_2
.L10:
#if defined(TRMMKERNEL) && defined(LEFT)
movl OFFSET, %eax
movl %eax, KK
#endif
movl K, %eax
sall $BASE_SHIFT + 1, %eax
leal (B, %eax), %eax
movl %eax, BX
movl C, CO1 # coffset = c
leal (, LDC, 2), %eax
addl %eax, C
movl A, AA # aoffset = a
movl M, %ebx
sarl $1, %ebx # i = (m >> 2)
jle .L20
ALIGN_4
.L11:
#if !defined(TRMMKERNEL) || \
(defined(TRMMKERNEL) && defined(LEFT) && defined(TRANSA)) || \
(defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
movl B, BB
#else
movl KK, %eax
leal (, %eax, SIZE), %eax
leal (AA, %eax, 2), AA
leal (B, %eax, 2), BB
#endif
movl BX, %eax
prefetcht0 0 * SIZE(%eax)
subl $-8 * SIZE, BX
movsd 0 * SIZE(AA), %xmm0
xorps %xmm2, %xmm2
xorps %xmm3, %xmm3
xorps %xmm4, %xmm4
prefetcht0 3 * SIZE(CO1)
xorps %xmm5, %xmm5
prefetcht0 3 * SIZE(CO1, LDC)
xorps %xmm6, %xmm6
xorps %xmm7, %xmm7
#ifndef TRMMKERNEL
movl K, %eax
#elif (defined(LEFT) && !defined(TRANSA)) || (!defined(LEFT) && defined(TRANSA))
movl K, %eax
subl KK, %eax
movl %eax, KKK
#else
movl KK, %eax
#ifdef LEFT
addl $2, %eax
#else
addl $2, %eax
#endif
movl %eax, KKK
#endif
sarl $2, %eax
je .L15
ALIGN_4
.L12:
PREFETCH (PREFETCHSIZE + 0) * SIZE(AA)
addsd %xmm2, %xmm6
movsd 1 * SIZE(AA), %xmm2
movaps %xmm0, %xmm1
mulsd 0 * SIZE(BB), %xmm0
addsd %xmm3, %xmm7
mulsd 1 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 2 * SIZE(AA), %xmm0
movaps %xmm2, %xmm3
mulsd 0 * SIZE(BB), %xmm2
addsd %xmm1, %xmm5
mulsd 1 * SIZE(BB), %xmm3
addsd %xmm2, %xmm6
movsd 3 * SIZE(AA), %xmm2
movaps %xmm0, %xmm1
mulsd 2 * SIZE(BB), %xmm0
addsd %xmm3, %xmm7
mulsd 3 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 4 * SIZE(AA), %xmm0
movaps %xmm2, %xmm3
mulsd 2 * SIZE(BB), %xmm2
addsd %xmm1, %xmm5
mulsd 3 * SIZE(BB), %xmm3
addsd %xmm2, %xmm6
movsd 5 * SIZE(AA), %xmm2
movaps %xmm0, %xmm1
mulsd 4 * SIZE(BB), %xmm0
addsd %xmm3, %xmm7
mulsd 5 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 6 * SIZE(AA), %xmm0
movaps %xmm2, %xmm3
mulsd 4 * SIZE(BB), %xmm2
addsd %xmm1, %xmm5
mulsd 5 * SIZE(BB), %xmm3
addsd %xmm2, %xmm6
movsd 7 * SIZE(AA), %xmm2
movaps %xmm0, %xmm1
mulsd 6 * SIZE(BB), %xmm0
addsd %xmm3, %xmm7
mulsd 7 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 8 * SIZE(AA), %xmm0
movaps %xmm2, %xmm3
mulsd 6 * SIZE(BB), %xmm2
addsd %xmm1, %xmm5
mulsd 7 * SIZE(BB), %xmm3
addl $8 * SIZE, BB
addl $8 * SIZE, AA
decl %eax
jne .L12
ALIGN_4
.L15:
#ifndef TRMMKERNEL
movl K, %eax
#else
movl KKK, %eax
#endif
andl $3, %eax # if (k & 1)
BRANCH
je .L18
ALIGN_3
.L16:
addsd %xmm2, %xmm6
movsd 1 * SIZE(AA), %xmm2
movaps %xmm0, %xmm1
mulsd 0 * SIZE(BB), %xmm0
addsd %xmm3, %xmm7
mulsd 1 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 2 * SIZE(AA), %xmm0
movaps %xmm2, %xmm3
mulsd 0 * SIZE(BB), %xmm2
addsd %xmm1, %xmm5
mulsd 1 * SIZE(BB), %xmm3
addl $2 * SIZE, AA
addl $2 * SIZE, BB
decl %eax
jg .L16
ALIGN_4
.L18:
movsd ALPHA_R, %xmm0
movsd ALPHA_I, %xmm1
addsd %xmm2, %xmm6
addsd %xmm3, %xmm7
movaps %xmm4, %xmm2
mulsd %xmm0, %xmm4
mulsd %xmm1, %xmm2
movaps %xmm6, %xmm3
mulsd %xmm0, %xmm6
mulsd %xmm1, %xmm3
addsd 0 * SIZE(CO1), %xmm4
addsd 1 * SIZE(CO1), %xmm2
addsd 2 * SIZE(CO1), %xmm6
addsd 3 * SIZE(CO1), %xmm3
movlps %xmm4, 0 * SIZE(CO1)
movlps %xmm2, 1 * SIZE(CO1)
movlps %xmm6, 2 * SIZE(CO1)
movlps %xmm3, 3 * SIZE(CO1)
movaps %xmm5, %xmm2
mulsd %xmm0, %xmm5
mulsd %xmm1, %xmm2
movaps %xmm7, %xmm3
mulsd %xmm0, %xmm7
mulsd %xmm1, %xmm3
addsd 0 * SIZE(CO1, LDC), %xmm5
addsd 1 * SIZE(CO1, LDC), %xmm2
addsd 2 * SIZE(CO1, LDC), %xmm7
addsd 3 * SIZE(CO1, LDC), %xmm3
movlps %xmm5, 0 * SIZE(CO1, LDC)
movlps %xmm2, 1 * SIZE(CO1, LDC)
movlps %xmm7, 2 * SIZE(CO1, LDC)
movlps %xmm3, 3 * SIZE(CO1, LDC)
addl $4 * SIZE, CO1
decl %ebx
jg .L11
ALIGN_4
.L20:
movl M, %ebx
testl $1, %ebx
jle .L29
#if !defined(TRMMKERNEL) || \
(defined(TRMMKERNEL) && defined(LEFT) && defined(TRANSA)) || \
(defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
movl B, BB
#else
movl KK, %eax
leal (, %eax, SIZE), %eax
leal (AA, %eax, 1), AA
leal (B, %eax, 2), BB
#endif
movsd 0 * SIZE(AA), %xmm0
xorps %xmm2, %xmm2
xorps %xmm3, %xmm3
xorps %xmm4, %xmm4
xorps %xmm5, %xmm5
#ifndef TRMMKERNEL
movl K, %eax
#elif (defined(LEFT) && !defined(TRANSA)) || (!defined(LEFT) && defined(TRANSA))
movl K, %eax
subl KK, %eax
movl %eax, KKK
#else
movl KK, %eax
#ifdef LEFT
addl $1, %eax
#else
addl $2, %eax
#endif
movl %eax, KKK
#endif
sarl $2, %eax
je .L25
ALIGN_4
.L22:
addsd %xmm2, %xmm4
movsd 0 * SIZE(BB), %xmm2
addsd %xmm3, %xmm5
movsd 1 * SIZE(BB), %xmm3
mulsd %xmm0, %xmm2
PREFETCH (PREFETCHSIZE + 0) * SIZE(AA)
mulsd %xmm0, %xmm3
movsd 1 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 2 * SIZE(BB), %xmm2
addsd %xmm3, %xmm5
movsd 3 * SIZE(BB), %xmm3
mulsd %xmm0, %xmm2
mulsd %xmm0, %xmm3
movsd 2 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 4 * SIZE(BB), %xmm2
addsd %xmm3, %xmm5
movsd 5 * SIZE(BB), %xmm3
mulsd %xmm0, %xmm2
mulsd %xmm0, %xmm3
movsd 3 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 6 * SIZE(BB), %xmm2
addsd %xmm3, %xmm5
movsd 7 * SIZE(BB), %xmm3
mulsd %xmm0, %xmm2
mulsd %xmm0, %xmm3
movsd 4 * SIZE(AA), %xmm0
addl $4 * SIZE, AA
addl $8 * SIZE, BB
decl %eax
jne .L22
ALIGN_4
.L25:
#ifndef TRMMKERNEL
movl K, %eax
#else
movl KKK, %eax
#endif
andl $3, %eax # if (k & 1)
BRANCH
je .L28
ALIGN_3
.L26:
addsd %xmm2, %xmm4
movsd 0 * SIZE(BB), %xmm2
addsd %xmm3, %xmm5
movsd 1 * SIZE(BB), %xmm3
mulsd %xmm0, %xmm2
mulsd %xmm0, %xmm3
movsd 1 * SIZE(AA), %xmm0
addl $1 * SIZE, AA
addl $2 * SIZE, BB
decl %eax
jg .L26
ALIGN_4
.L28:
movsd ALPHA_R, %xmm0
movsd ALPHA_I, %xmm1
addsd %xmm2, %xmm4
addsd %xmm3, %xmm5
movaps %xmm4, %xmm2
mulsd %xmm0, %xmm4
mulsd %xmm1, %xmm2
movaps %xmm5, %xmm3
mulsd %xmm0, %xmm5
mulsd %xmm1, %xmm3
addsd 0 * SIZE(CO1), %xmm4
addsd 1 * SIZE(CO1), %xmm2
addsd 0 * SIZE(CO1, LDC), %xmm5
addsd 1 * SIZE(CO1, LDC), %xmm3
movlps %xmm4, 0 * SIZE(CO1)
movlps %xmm2, 1 * SIZE(CO1)
movlps %xmm5, 0 * SIZE(CO1, LDC)
movlps %xmm3, 1 * SIZE(CO1, LDC)
ALIGN_4
.L29:
#if defined(TRMMKERNEL) && !defined(LEFT)
addl $2, KK
#endif
movl BB, B
decl J
jg .L10
ALIGN_4
.L30:
testl $1, N
je .L999
#if defined(TRMMKERNEL) && defined(LEFT)
movl OFFSET, %eax
movl %eax, KK
#endif
movl C, CO1
addl LDC, C
movl A, AA
movl M, %ebx
sarl $1, %ebx
jle .L40
ALIGN_4
.L31:
#if !defined(TRMMKERNEL) || \
(defined(TRMMKERNEL) && defined(LEFT) && defined(TRANSA)) || \
(defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
movl B, BB
#else
movl KK, %eax
leal (, %eax, SIZE), %eax
leal (AA, %eax, 2), AA
leal (B, %eax, 1), BB
#endif
movsd 0 * SIZE(BB), %xmm1
xorps %xmm0, %xmm0
prefetcht0 3 * SIZE(CO1)
xorps %xmm2, %xmm2
xorps %xmm4, %xmm4
xorps %xmm6, %xmm6
#ifndef TRMMKERNEL
movl K, %eax
#elif (defined(LEFT) && !defined(TRANSA)) || (!defined(LEFT) && defined(TRANSA))
movl K, %eax
subl KK, %eax
movl %eax, KKK
#else
movl KK, %eax
#ifdef LEFT
addl $2, %eax
#else
addl $1, %eax
#endif
movl %eax, KKK
#endif
sarl $2, %eax
je .L35
ALIGN_4
.L32:
PREFETCH (PREFETCHSIZE + 0) * SIZE(AA)
addsd %xmm0, %xmm4
movsd 0 * SIZE(AA), %xmm0
addsd %xmm2, %xmm6
movsd 1 * SIZE(AA), %xmm2
mulsd %xmm1, %xmm0
mulsd %xmm1, %xmm2
movsd 1 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 2 * SIZE(AA), %xmm0
addsd %xmm2, %xmm6
movsd 3 * SIZE(AA), %xmm2
mulsd %xmm1, %xmm0
mulsd %xmm1, %xmm2
movsd 2 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 4 * SIZE(AA), %xmm0
addsd %xmm2, %xmm6
movsd 5 * SIZE(AA), %xmm2
mulsd %xmm1, %xmm0
mulsd %xmm1, %xmm2
movsd 3 * SIZE(BB), %xmm1
addsd %xmm0, %xmm4
movsd 6 * SIZE(AA), %xmm0
addsd %xmm2, %xmm6
movsd 7 * SIZE(AA), %xmm2
mulsd %xmm1, %xmm0
mulsd %xmm1, %xmm2
movsd 4 * SIZE(BB), %xmm1
addl $8 * SIZE, AA
addl $4 * SIZE, BB
decl %eax
jne .L32
ALIGN_4
.L35:
#ifndef TRMMKERNEL
movl K, %eax
#else
movl KKK, %eax
#endif
andl $3, %eax # if (k & 1)
BRANCH
je .L38
ALIGN_3
.L36:
addsd %xmm0, %xmm4
movsd 0 * SIZE(AA), %xmm0
addsd %xmm2, %xmm6
movsd 1 * SIZE(AA), %xmm2
mulsd %xmm1, %xmm0
mulsd %xmm1, %xmm2
movsd 1 * SIZE(BB), %xmm1
addl $2 * SIZE, AA
addl $1 * SIZE, BB
decl %eax
jg .L36
ALIGN_4
.L38:
addsd %xmm0, %xmm4
addsd %xmm2, %xmm6
movsd ALPHA_R, %xmm0
movsd ALPHA_I, %xmm1
movaps %xmm4, %xmm2
mulsd %xmm0, %xmm4
mulsd %xmm1, %xmm2
movaps %xmm6, %xmm3
mulsd %xmm0, %xmm6
mulsd %xmm1, %xmm3
addsd 0 * SIZE(CO1), %xmm4
addsd 1 * SIZE(CO1), %xmm2
addsd 2 * SIZE(CO1), %xmm6
addsd 3 * SIZE(CO1), %xmm3
movlps %xmm4, 0 * SIZE(CO1)
movlps %xmm2, 1 * SIZE(CO1)
movlps %xmm6, 2 * SIZE(CO1)
movlps %xmm3, 3 * SIZE(CO1)
addl $4 * SIZE, CO1
decl %ebx
jg .L31
ALIGN_4
.L40:
movl M, %ebx
testl $1, %ebx
jle .L999
#if !defined(TRMMKERNEL) || \
(defined(TRMMKERNEL) && defined(LEFT) && defined(TRANSA)) || \
(defined(TRMMKERNEL) && !defined(LEFT) && !defined(TRANSA))
movl B, BB
#else
movl KK, %eax
leal (, %eax, SIZE), %eax
leal (AA, %eax, 1), AA
leal (B, %eax, 1), BB
#endif
movsd 0 * SIZE(AA), %xmm0
xorps %xmm4, %xmm4
movsd 0 * SIZE(BB), %xmm2
xorps %xmm5, %xmm5
#ifndef TRMMKERNEL
movl K, %eax
#elif (defined(LEFT) && !defined(TRANSA)) || (!defined(LEFT) && defined(TRANSA))
movl K, %eax
subl KK, %eax
movl %eax, KKK
#else
movl KK, %eax
#ifdef LEFT
addl $1, %eax
#else
addl $1, %eax
#endif
movl %eax, KKK
#endif
sarl $2, %eax
je .L45
ALIGN_4
.L42:
mulsd %xmm0, %xmm2
movsd 1 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 1 * SIZE(BB), %xmm2
mulsd %xmm0, %xmm2
movsd 2 * SIZE(AA), %xmm0
addsd %xmm2, %xmm5
movsd 2 * SIZE(BB), %xmm2
mulsd %xmm0, %xmm2
movsd 3 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 3 * SIZE(BB), %xmm2
mulsd %xmm0, %xmm2
movsd 4 * SIZE(AA), %xmm0
addsd %xmm2, %xmm5
movsd 4 * SIZE(BB), %xmm2
addl $4 * SIZE, AA
addl $4 * SIZE, BB
decl %eax
jne .L42
ALIGN_4
.L45:
#ifndef TRMMKERNEL
movl K, %eax
#else
movl KKK, %eax
#endif
andl $3, %eax # if (k & 1)
BRANCH
je .L48
ALIGN_3
.L46:
mulsd %xmm0, %xmm2
movsd 1 * SIZE(AA), %xmm0
addsd %xmm2, %xmm4
movsd 1 * SIZE(BB), %xmm2
addl $1 * SIZE, AA
addl $1 * SIZE, BB
decl %eax
jg .L46
ALIGN_4
.L48:
addsd %xmm5, %xmm4
movsd ALPHA_R, %xmm0
movsd ALPHA_I, %xmm1
movaps %xmm4, %xmm2
mulsd %xmm0, %xmm4
mulsd %xmm1, %xmm2
addsd 0 * SIZE(CO1), %xmm4
addsd 1 * SIZE(CO1), %xmm2
movlps %xmm4, 0 * SIZE(CO1)
movlps %xmm2, 1 * SIZE(CO1)
ALIGN_4
.L999:
popl %ebx
popl %esi
popl %edi
popl %ebp
addl $ARGS, %esp
ret
EPILOGUE