649 lines
15 KiB
ArmAsm
649 lines
15 KiB
ArmAsm
/***************************************************************************
|
|
Copyright (c) 2020, The OpenBLAS Project
|
|
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.
|
|
3. Neither the name of the OpenBLAS project nor the names of
|
|
its contributors may be used to endorse or promote products
|
|
derived from this software without specific prior written permission.
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "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 OPENBLAS PROJECT 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.
|
|
*****************************************************************************/
|
|
|
|
#define ASSEMBLER
|
|
|
|
#include "common.h"
|
|
|
|
#define M $r4
|
|
#define N $r5
|
|
#define A $r7
|
|
#define LDA $r8
|
|
#define X $r9
|
|
#define INCX $r10
|
|
#define Y $r11
|
|
#define INCY $r6
|
|
#define BUFFER $r17
|
|
|
|
#define YORIG $r18
|
|
#define XX $r12
|
|
#define YY $r13
|
|
#define I $r14
|
|
#define J $r15
|
|
#define AO1 $r23
|
|
#define AO2 $r24
|
|
|
|
#define ALPHA_R $f0
|
|
#define ALPHA_I $f1
|
|
#define a1 $f22
|
|
#define a2 $f8
|
|
#define a3 $f23
|
|
#define a4 $f9
|
|
#define a5 $f10
|
|
#define a6 $f11
|
|
#define a7 $f12
|
|
#define a8 $f13
|
|
#define x1 $f14
|
|
#define x2 $f15
|
|
#define x3 $f16
|
|
#define x4 $f17
|
|
#define y1 $f3
|
|
#define y2 $f4
|
|
#define y3 $f2
|
|
#define y4 $f5
|
|
#define t1 $f6
|
|
#define t2 $f7
|
|
#define t3 $f18
|
|
#define t4 $f19
|
|
#define t5 $f20
|
|
#define t6 $f21
|
|
#define t7 $f24
|
|
#define t8 $f25
|
|
|
|
#if !defined(CONJ) && !defined(XCONJ)
|
|
#define MADD1 MADD
|
|
#define MADD2 MADD
|
|
#define MADD3 NMSUB
|
|
#define MADD4 MADD
|
|
#endif
|
|
#if defined(CONJ) && !defined(XCONJ)
|
|
#define MADD1 MADD
|
|
#define MADD2 MADD
|
|
#define MADD3 MADD
|
|
#define MADD4 NMSUB
|
|
#endif
|
|
#if !defined(CONJ) && defined(XCONJ)
|
|
#define MADD1 MADD
|
|
#define MADD2 NMSUB
|
|
#define MADD3 MADD
|
|
#define MADD4 MADD
|
|
#endif
|
|
#if defined(CONJ) && defined(XCONJ)
|
|
#define MADD1 MADD
|
|
#define MADD2 NMSUB
|
|
#define MADD3 NMSUB
|
|
#define MADD4 NMSUB
|
|
#endif
|
|
|
|
PROLOGUE
|
|
|
|
LDARG INCY, $sp, 0
|
|
LDARG BUFFER, $sp, 8
|
|
#ifndef __64BIT__
|
|
addi.d $sp, $sp, -64
|
|
#else
|
|
addi.d $sp, $sp, -32
|
|
#endif
|
|
SDARG $r23, $sp, 0
|
|
SDARG $r24, $sp, 8
|
|
fst.d $f24, $sp, 16
|
|
fst.d $f25, $sp, 24
|
|
#ifndef __64BIT__
|
|
fst.d $f18, $sp, 32
|
|
fst.d $f19, $sp, 40
|
|
fst.d $f20, $sp, 48
|
|
fst.d $f21, $sp, 56
|
|
#endif
|
|
slli.d LDA, LDA, ZBASE_SHIFT
|
|
slli.d INCX, INCX, ZBASE_SHIFT
|
|
bge $r0, M, .L999
|
|
slli.d INCY, INCY, ZBASE_SHIFT
|
|
bge $r0, N, .L999
|
|
li.d I, 2 * SIZE
|
|
move YORIG, Y
|
|
beq INCY, I, .L10
|
|
srai.d I, M, 2
|
|
move YORIG, BUFFER
|
|
move XX, Y
|
|
move YY, BUFFER
|
|
bge $r0, I, .L05
|
|
.align 3
|
|
|
|
.L02:
|
|
LD a1, XX, 0 * SIZE
|
|
LD a2, XX, 1 * SIZE
|
|
add.d XX, XX, INCY
|
|
LD a3, XX, 0 * SIZE
|
|
LD a4, XX, 1 * SIZE
|
|
add.d XX, XX, INCY
|
|
LD a5, XX, 0 * SIZE
|
|
LD a6, XX, 1 * SIZE
|
|
add.d XX, XX, INCY
|
|
LD a7, XX, 0 * SIZE
|
|
LD a8, XX, 1 * SIZE
|
|
add.d XX, XX, INCY
|
|
addi.d I, I, -1
|
|
addi.d YY, YY, 8 * SIZE
|
|
ST a1, YY, -8 * SIZE
|
|
ST a2, YY, -7 * SIZE
|
|
ST a3, YY, -6 * SIZE
|
|
ST a4, YY, -5 * SIZE
|
|
ST a5, YY, -4 * SIZE
|
|
ST a6, YY, -3 * SIZE
|
|
ST a7, YY, -2 * SIZE
|
|
ST a8, YY, -1 * SIZE
|
|
blt $r0, I, .L02
|
|
.align 3
|
|
|
|
.L05:
|
|
andi I, M, 3
|
|
bge $r0, I, .L10
|
|
.align 3
|
|
|
|
.L06:
|
|
LD a1, XX, 0 * SIZE
|
|
LD a2, XX, 1 * SIZE
|
|
add.d XX, XX, INCY
|
|
addi.d I, I, -1
|
|
ST a1, YY, 0 * SIZE
|
|
ST a2, YY, 1 * SIZE
|
|
addi.d YY, YY, 2 * SIZE
|
|
blt $r0, I, .L06
|
|
.align 3
|
|
|
|
.L10:
|
|
srai.d J, N, 1
|
|
bge $r0, J, .L20
|
|
.align 3
|
|
|
|
.L11:
|
|
LD x1, X, 0 * SIZE
|
|
LD x2, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
LD x3, X, 0 * SIZE
|
|
LD x4, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
MUL a1, ALPHA_R, x1
|
|
move AO1, A
|
|
MUL a2, ALPHA_I, x1
|
|
add.d AO2, A, LDA
|
|
MUL a3, ALPHA_R, x3
|
|
add.d A, AO2, LDA
|
|
MUL a4, ALPHA_I, x3
|
|
#ifndef XCONJ
|
|
NMSUB x1, x2, ALPHA_I, a1
|
|
MADD x2, x2, ALPHA_R, a2
|
|
NMSUB x3, x4, ALPHA_I, a3
|
|
MADD x4, x4, ALPHA_R, a4
|
|
#else
|
|
MADD x1, x2, ALPHA_I, a1
|
|
MSUB x2, x2, ALPHA_R, a2
|
|
MADD x3, x4, ALPHA_I, a3
|
|
MSUB x4, x4, ALPHA_R, a4
|
|
#endif
|
|
srai.d I, M, 2
|
|
move YY, YORIG
|
|
bge $r0, I, .L15
|
|
LD y1, YY, 0 * SIZE
|
|
LD a1, AO1, 0 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a3, AO1, 2 * SIZE
|
|
LD y3, YY, 2 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
LD y4, YY, 3 * SIZE
|
|
LD a4, AO1, 3 * SIZE
|
|
LD a5, AO2, 0 * SIZE
|
|
LD a6, AO2, 1 * SIZE
|
|
LD a7, AO2, 2 * SIZE
|
|
LD a8, AO2, 3 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD y1, YY, 4 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a1, AO1, 4 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
LD y2, YY, 5 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
LD a3, AO1, 6 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
LD y3, YY, 6 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
LD a2, AO1, 5 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
LD y4, YY, 7 * SIZE
|
|
MADD4 t4, a4, x1, t4
|
|
LD a4, AO1, 7 * SIZE
|
|
MADD1 t1, a5, x3, t1
|
|
MADD2 t2, a5, x4, t2
|
|
LD a5, AO2, 4 * SIZE
|
|
MADD1 t3, a7, x3, t3
|
|
MADD2 t4, a7, x4, t4
|
|
LD a7, AO2, 6 * SIZE
|
|
MADD3 t1, a6, x4, t1
|
|
MADD4 t2, a6, x3, t2
|
|
LD a6, AO2, 5 * SIZE
|
|
MADD3 t3, a8, x4, t3
|
|
addi.d I, I, -1
|
|
MADD4 t4, a8, x3, t4
|
|
LD a8, AO2, 7 * SIZE
|
|
bge $r0, I, .L13
|
|
.align 3
|
|
.L12:
|
|
MADD1 t5, a1, x1, y1
|
|
LD y1, YY, 8 * SIZE
|
|
MADD2 t6, a1, x2, y2
|
|
LD a1, AO1, 8 * SIZE
|
|
MADD1 t7, a3, x1, y3
|
|
LD y2, YY, 9 * SIZE
|
|
MADD2 t8, a3, x2, y4
|
|
LD a3, AO1, 10 * SIZE
|
|
MADD3 t5, a2, x2, t5
|
|
LD y3, YY, 10 * SIZE
|
|
MADD4 t6, a2, x1, t6
|
|
LD a2, AO1, 9 * SIZE
|
|
MADD3 t7, a4, x2, t7
|
|
LD y4, YY, 11 * SIZE
|
|
MADD4 t8, a4, x1, t8
|
|
LD a4, AO1, 11 * SIZE
|
|
MADD1 t5, a5, x3, t5
|
|
ST t1, YY, 0 * SIZE
|
|
MADD2 t6, a5, x4, t6
|
|
LD a5, AO2, 8 * SIZE
|
|
MADD1 t7, a7, x3, t7
|
|
ST t2, YY, 1 * SIZE
|
|
MADD2 t8, a7, x4, t8
|
|
LD a7, AO2, 10 * SIZE
|
|
MADD3 t5, a6, x4, t5
|
|
ST t3, YY, 2 * SIZE
|
|
MADD4 t6, a6, x3, t6
|
|
LD a6, AO2, 9 * SIZE
|
|
MADD3 t7, a8, x4, t7
|
|
ST t4, YY, 3 * SIZE
|
|
MADD4 t8, a8, x3, t8
|
|
LD a8, AO2, 11 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD y1, YY, 12 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a1, AO1, 12 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
LD y2, YY, 13 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
LD a3, AO1, 14 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
LD y3, YY, 14 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
LD a2, AO1, 13 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
LD y4, YY, 15 * SIZE
|
|
MADD4 t4, a4, x1, t4
|
|
LD a4, AO1, 15 * SIZE
|
|
MADD1 t1, a5, x3, t1
|
|
ST t5, YY, 4 * SIZE
|
|
MADD2 t2, a5, x4, t2
|
|
LD a5, AO2, 12 * SIZE
|
|
MADD1 t3, a7, x3, t3
|
|
ST t6, YY, 5 * SIZE
|
|
MADD2 t4, a7, x4, t4
|
|
LD a7, AO2, 14 * SIZE
|
|
MADD3 t1, a6, x4, t1
|
|
ST t7, YY, 6 * SIZE
|
|
MADD4 t2, a6, x3, t2
|
|
LD a6, AO2, 13 * SIZE
|
|
MADD3 t3, a8, x4, t3
|
|
ST t8, YY, 7 * SIZE
|
|
MADD4 t4, a8, x3, t4
|
|
LD a8, AO2, 15 * SIZE
|
|
addi.d I, I, -1
|
|
addi.d YY, YY, 8 * SIZE
|
|
addi.d AO1, AO1, 8 * SIZE
|
|
addi.d AO2, AO2, 8 * SIZE
|
|
blt $r0, I, .L12
|
|
.align 3
|
|
|
|
.L13:
|
|
ST t1, YY, 0 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
ST t2, YY, 1 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
ST t3, YY, 2 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
ST t4, YY, 3 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
MADD3 t1, a2, x2, t1
|
|
MADD4 t2, a2, x1, t2
|
|
MADD3 t3, a4, x2, t3
|
|
MADD4 t4, a4, x1, t4
|
|
MADD1 t1, a5, x3, t1
|
|
MADD2 t2, a5, x4, t2
|
|
MADD1 t3, a7, x3, t3
|
|
MADD2 t4, a7, x4, t4
|
|
MADD3 t1, a6, x4, t1
|
|
addi.d AO1, AO1, 8 * SIZE
|
|
MADD4 t2, a6, x3, t2
|
|
addi.d AO2, AO2, 8 * SIZE
|
|
MADD3 t3, a8, x4, t3
|
|
addi.d YY, YY, 8 * SIZE
|
|
MADD4 t4, a8, x3, t4
|
|
ST t1, YY, -4 * SIZE
|
|
ST t2, YY, -3 * SIZE
|
|
ST t3, YY, -2 * SIZE
|
|
ST t4, YY, -1 * SIZE
|
|
.align 3
|
|
|
|
.L15:
|
|
andi I, M, 2
|
|
bge $r0, I, .L16
|
|
LD a1, AO1, 0 * SIZE
|
|
LD y1, YY, 0 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a3, AO1, 2 * SIZE
|
|
LD y3, YY, 2 * SIZE
|
|
LD a4, AO1, 3 * SIZE
|
|
LD y4, YY, 3 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD a5, AO2, 0 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a6, AO2, 1 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
LD a7, AO2, 2 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
LD a8, AO2, 3 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
MADD4 t2, a2, x1, t2
|
|
MADD3 t3, a4, x2, t3
|
|
MADD4 t4, a4, x1, t4
|
|
MADD1 t1, a5, x3, t1
|
|
MADD2 t2, a5, x4, t2
|
|
MADD1 t3, a7, x3, t3
|
|
MADD2 t4, a7, x4, t4
|
|
MADD3 t1, a6, x4, t1
|
|
addi.d YY, YY, 4 * SIZE
|
|
MADD4 t2, a6, x3, t2
|
|
addi.d AO1, AO1, 4 * SIZE
|
|
MADD3 t3, a8, x4, t3
|
|
addi.d AO2, AO2, 4 * SIZE
|
|
MADD4 t4, a8, x3, t4
|
|
ST t1, YY, -4 * SIZE
|
|
ST t2, YY, -3 * SIZE
|
|
ST t3, YY, -2 * SIZE
|
|
ST t4, YY, -1 * SIZE
|
|
.align 3
|
|
|
|
.L16:
|
|
andi I, M, 1
|
|
bge $r0, I, .L19
|
|
LD y1, YY, 0 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a1, AO1, 0 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD a5, AO2, 0 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a6, AO2, 1 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
MADD4 t2, a2, x1, t2
|
|
MADD1 t1, a5, x3, t1
|
|
MADD2 t2, a5, x4, t2
|
|
MADD3 t1, a6, x4, t1
|
|
MADD4 t2, a6, x3, t2
|
|
ST t1, YY, 0 * SIZE
|
|
ST t2, YY, 1 * SIZE
|
|
.align 3
|
|
|
|
.L19:
|
|
addi.d J, J, -1
|
|
blt $r0, J, .L11
|
|
.align 3
|
|
|
|
.L20:
|
|
andi J, N, 1
|
|
bge $r0, J, .L900
|
|
LD x1, X, 0 * SIZE
|
|
LD x2, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
MUL a1, ALPHA_R, x1
|
|
move AO1, A
|
|
MUL a2, ALPHA_I, x1
|
|
#ifndef XCONJ
|
|
NMSUB x1, x2, ALPHA_I, a1
|
|
MADD x2, x2, ALPHA_R, a2
|
|
#else
|
|
MADD x1, x2, ALPHA_I, a1
|
|
MSUB x2, x2, ALPHA_R, a2
|
|
#endif
|
|
srai.d I, M, 2
|
|
move YY, YORIG
|
|
bge $r0, I, .L25
|
|
LD y1, YY, 0 * SIZE
|
|
LD a1, AO1, 0 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a3, AO1, 2 * SIZE
|
|
LD y3, YY, 2 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
LD y4, YY, 3 * SIZE
|
|
LD a4, AO1, 3 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD y1, YY, 4 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a1, AO1, 4 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
LD y2, YY, 5 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
LD a3, AO1, 6 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
LD y3, YY, 6 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
LD a2, AO1, 5 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
LD y4, YY, 7 * SIZE
|
|
MADD4 t4, a4, x1, t4
|
|
addi.d I, I, -1
|
|
LD a4, AO1, 7 * SIZE
|
|
bge $r0, I, .L23
|
|
.align 3
|
|
.L22:
|
|
MADD1 t5, a1, x1, y1
|
|
LD y1, YY, 8 * SIZE
|
|
MADD2 t6, a1, x2, y2
|
|
LD a1, AO1, 8 * SIZE
|
|
MADD1 t7, a3, x1, y3
|
|
LD y2, YY, 9 * SIZE
|
|
MADD2 t8, a3, x2, y4
|
|
LD a3, AO1, 10 * SIZE
|
|
MADD3 t5, a2, x2, t5
|
|
LD y3, YY, 10 * SIZE
|
|
MADD4 t6, a2, x1, t6
|
|
LD a2, AO1, 9 * SIZE
|
|
MADD3 t7, a4, x2, t7
|
|
LD y4, YY, 11 * SIZE
|
|
MADD4 t8, a4, x1, t8
|
|
LD a4, AO1, 11 * SIZE
|
|
ST t1, YY, 0 * SIZE
|
|
ST t2, YY, 1 * SIZE
|
|
ST t3, YY, 2 * SIZE
|
|
ST t4, YY, 3 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
LD y1, YY, 12 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
LD a1, AO1, 12 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
LD y2, YY, 13 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
LD a3, AO1, 14 * SIZE
|
|
MADD3 t1, a2, x2, t1
|
|
LD y3, YY, 14 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
LD a2, AO1, 13 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
LD y4, YY, 15 * SIZE
|
|
MADD4 t4, a4, x1, t4
|
|
LD a4, AO1, 15 * SIZE
|
|
ST t5, YY, 4 * SIZE
|
|
ST t6, YY, 5 * SIZE
|
|
ST t7, YY, 6 * SIZE
|
|
ST t8, YY, 7 * SIZE
|
|
addi.d I, I, -1
|
|
addi.d YY, YY, 8 * SIZE
|
|
addi.d AO1, AO1, 8 * SIZE
|
|
blt $r0, I, .L22
|
|
.align 3
|
|
|
|
.L23:
|
|
ST t1, YY, 0 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
ST t2, YY, 1 * SIZE
|
|
MADD2 t2, a1, x2, y2
|
|
ST t3, YY, 2 * SIZE
|
|
MADD1 t3, a3, x1, y3
|
|
ST t4, YY, 3 * SIZE
|
|
MADD2 t4, a3, x2, y4
|
|
MADD3 t1, a2, x2, t1
|
|
addi.d AO1, AO1, 8 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
addi.d YY, YY, 8 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
MADD4 t4, a4, x1, t4
|
|
ST t1, YY, -4 * SIZE
|
|
ST t2, YY, -3 * SIZE
|
|
ST t3, YY, -2 * SIZE
|
|
ST t4, YY, -1 * SIZE
|
|
.align 3
|
|
|
|
.L25:
|
|
andi I, M, 2
|
|
bge $r0, I, .L26
|
|
LD a1, AO1, 0 * SIZE
|
|
LD y1, YY, 0 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a3, AO1, 2 * SIZE
|
|
LD y3, YY, 2 * SIZE
|
|
LD a4, AO1, 3 * SIZE
|
|
LD y4, YY, 3 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
MADD2 t2, a1, x2, y2
|
|
MADD1 t3, a3, x1, y3
|
|
MADD2 t4, a3, x2, y4
|
|
MADD3 t1, a2, x2, t1
|
|
addi.d YY, YY, 4 * SIZE
|
|
MADD4 t2, a2, x1, t2
|
|
addi.d AO1, AO1, 4 * SIZE
|
|
MADD3 t3, a4, x2, t3
|
|
MADD4 t4, a4, x1, t4
|
|
ST t1, YY, -4 * SIZE
|
|
ST t2, YY, -3 * SIZE
|
|
ST t3, YY, -2 * SIZE
|
|
ST t4, YY, -1 * SIZE
|
|
.align 3
|
|
|
|
.L26:
|
|
andi I, M, 1
|
|
bge $r0, I, .L900
|
|
LD y1, YY, 0 * SIZE
|
|
LD y2, YY, 1 * SIZE
|
|
LD a1, AO1, 0 * SIZE
|
|
LD a2, AO1, 1 * SIZE
|
|
MADD1 t1, a1, x1, y1
|
|
MADD2 t2, a1, x2, y2
|
|
MADD3 t1, a2, x2, t1
|
|
MADD4 t2, a2, x1, t2
|
|
ST t1, YY, 0 * SIZE
|
|
ST t2, YY, 1 * SIZE
|
|
.align 3
|
|
|
|
.L900:
|
|
li.d YORIG, 2 * SIZE
|
|
srai.d I, M, 2
|
|
beq INCY, YORIG, .L999
|
|
move XX, BUFFER
|
|
bge $r0, I, .L905
|
|
.align 3
|
|
|
|
.L902:
|
|
LD a1, XX, 0 * SIZE
|
|
LD a2, XX, 1 * SIZE
|
|
LD a3, XX, 2 * SIZE
|
|
LD a4, XX, 3 * SIZE
|
|
LD a5, XX, 4 * SIZE
|
|
LD a6, XX, 5 * SIZE
|
|
LD a7, XX, 6 * SIZE
|
|
LD a8, XX, 7 * SIZE
|
|
addi.d I, I, -1
|
|
ST a1, Y, 0 * SIZE
|
|
ST a2, Y, 1 * SIZE
|
|
add.d Y, Y, INCY
|
|
ST a3, Y, 0 * SIZE
|
|
ST a4, Y, 1 * SIZE
|
|
add.d Y, Y, INCY
|
|
ST a5, Y, 0 * SIZE
|
|
ST a6, Y, 1 * SIZE
|
|
add.d Y, Y, INCY
|
|
ST a7, Y, 0 * SIZE
|
|
ST a8, Y, 1 * SIZE
|
|
add.d Y, Y, INCY
|
|
addi.d XX, XX, 8 * SIZE
|
|
blt $r0, I, .L902
|
|
.align 3
|
|
|
|
.L905:
|
|
andi I, M, 3
|
|
bge $r0, I, .L999
|
|
.align 3
|
|
|
|
.L906:
|
|
LD a1, XX, 0 * SIZE
|
|
LD a2, XX, 1 * SIZE
|
|
addi.d XX, XX, 2 * SIZE
|
|
addi.d I, I, -1
|
|
ST a1, Y, 0 * SIZE
|
|
ST a2, Y, 1 * SIZE
|
|
add.d Y, Y, INCY
|
|
blt $r0, I, .L906
|
|
.align 3
|
|
|
|
.L999:
|
|
LDARG $r23, $sp, 0
|
|
LDARG $r24, $sp, 8
|
|
fld.d $f24, $sp, 16
|
|
fld.d $f25, $sp, 24
|
|
#ifndef __64BIT__
|
|
fld.d $f18, $sp, 32
|
|
fld.d $f19, $sp, 40
|
|
fld.d $f20, $sp, 48
|
|
fld.d $f21, $sp, 56
|
|
#endif
|
|
#ifdef __64BIT__
|
|
addi.d $sp, $sp, 32
|
|
#else
|
|
addi.d $sp, $sp, 64
|
|
#endif
|
|
move $r4, $r17
|
|
fmov.d $f0, $f22
|
|
jirl $r0, $r1, 0x0
|
|
|
|
EPILOGUE
|