OpenBLAS/kernel/loongarch64/scal.S

331 lines
7.4 KiB
ArmAsm

/***************************************************************************
Copyright (c) 2021, 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 N $r4
#define X $r7
#define INCX $r8
#define I $r17
#define TEMP $r18
#define XX $r5
#define ALPHA $f0
#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 t1 $f14
#define t2 $f15
#define t3 $f16
#define t4 $f17
PROLOGUE
li.d TEMP, SIZE
MTC a1, $r0
slli.d INCX, INCX, BASE_SHIFT
bge $r0, N, .L999
CMPEQ $fcc0, ALPHA, a1
bceqz $fcc0, .L50
srai.d I, N, 3
bne INCX, TEMP, .L20
bge $r0, I, .L15
.align 3
.L12:
ST a1, X, 0 * SIZE
ST a1, X, 1 * SIZE
ST a1, X, 2 * SIZE
ST a1, X, 3 * SIZE
ST a1, X, 4 * SIZE
ST a1, X, 5 * SIZE
ST a1, X, 6 * SIZE
ST a1, X, 7 * SIZE
addi.w I, I, -1
addi.d X, X, 8 * SIZE
blt $r0, I, .L12
.align 3
.L15:
andi I, N, 7
bge $r0, I, .L999
.align 3
.L16:
ST a1, X, 0 * SIZE
addi.d I, I, -1
addi.d X, X, SIZE
blt $r0, I, .L16
move $r4, $r17
fmov.d $f0, $f22
jirl $r0, $r1, 0x0
.align 3
.L20:
srai.d I, N, 3
bge $r0, I, .L25
.align 3
.L22:
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
add.d X, X, INCX
ST a1, X, 0 * SIZE
addi.d I, I, -1
add.d X, X, INCX
blt $r0, I, .L22
.align 3
.L25:
andi I, N, 7
bge $r0, I, .L999
.align 3
.L26:
addi.d I, I, -1
ST a1, X, 0 * SIZE
add.d X, X, INCX
blt $r0, I, .L26
move $r4, $r17
fmov.d $f0, $f22
jirl $r0, $r1, 0x0
.align 3
.L50:
srai.d I, N, 3
bne INCX, TEMP, .L60
addi.d I, I, -1
blt I, $r0, .L55
LD a1, X, 0 * SIZE
LD a2, X, 1 * SIZE
LD a3, X, 2 * SIZE
LD a4, X, 3 * SIZE
LD a5, X, 4 * SIZE
LD a6, X, 5 * SIZE
LD a7, X, 6 * SIZE
LD a8, X, 7 * SIZE
bge $r0, I, .L53
.align 3
.L52:
MUL t1, ALPHA, a1
LD a1, X, 8 * SIZE
MUL t2, ALPHA, a2
LD a2, X, 9 * SIZE
MUL t3, ALPHA, a3
LD a3, X, 10 * SIZE
MUL t4, ALPHA, a4
LD a4, X, 11 * SIZE
ST t1, X, 0 * SIZE
MUL t1, ALPHA, a5
LD a5, X, 12 * SIZE
ST t2, X, 1 * SIZE
MUL t2, ALPHA, a6
LD a6, X, 13 * SIZE
ST t3, X, 2 * SIZE
MUL t3, ALPHA, a7
LD a7, X, 14 * SIZE
ST t4, X, 3 * SIZE
MUL t4, ALPHA, a8
LD a8, X, 15 * SIZE
addi.d I, I, -1
ST t1, X, 4 * SIZE
ST t2, X, 5 * SIZE
ST t3, X, 6 * SIZE
ST t4, X, 7 * SIZE
addi.d X, X, 8 * SIZE
blt $r0, I, .L52
.align 3
.L53:
MUL t1, ALPHA, a1
MUL t2, ALPHA, a2
MUL t3, ALPHA, a3
MUL t4, ALPHA, a4
ST t1, X, 0 * SIZE
MUL t1, ALPHA, a5
ST t2, X, 1 * SIZE
MUL t2, ALPHA, a6
ST t3, X, 2 * SIZE
MUL t3, ALPHA, a7
ST t4, X, 3 * SIZE
MUL t4, ALPHA, a8
ST t1, X, 4 * SIZE
ST t2, X, 5 * SIZE
ST t3, X, 6 * SIZE
ST t4, X, 7 * SIZE
addi.d X, X, 8 * SIZE
.align 3
.L55:
andi I, N, 7
bge $r0, I, .L999
.align 3
.L56:
LD a1, X, 0 * SIZE
MUL t1, ALPHA, a1
addi.d X, X, SIZE
addi.d I, I, -1
ST t1, X, -1 * SIZE
blt $r0, I, .L56
move $r4, $r17
fmov.d $f0, $f22
jirl $r0, $r1, 0x0
.align 3
.L60:
srai.d I, N, 3
move XX, X
addi.d I, I, -1
blt I, $r0, .L65
LD a1, X, 0 * SIZE
add.d X, X, INCX
LD a2, X, 0 * SIZE
add.d X, X, INCX
LD a3, X, 0 * SIZE
add.d X, X, INCX
LD a4, X, 0 * SIZE
add.d X, X, INCX
LD a5, X, 0 * SIZE
add.d X, X, INCX
LD a6, X, 0 * SIZE
add.d X, X, INCX
LD a7, X, 0 * SIZE
add.d X, X, INCX
LD a8, X, 0 * SIZE
add.d X, X, INCX
bge $r0, I, .L63
.align 3
.L62:
MUL t1, ALPHA, a1
LD a1, X, 0 * SIZE
add.d X, X, INCX
MUL t2, ALPHA, a2
LD a2, X, 0 * SIZE
add.d X, X, INCX
MUL t3, ALPHA, a3
LD a3, X, 0 * SIZE
add.d X, X, INCX
MUL t4, ALPHA, a4
LD a4, X, 0 * SIZE
add.d X, X, INCX
ST t1, XX, 0 * SIZE
add.d XX, XX, INCX
ST t2, XX, 0 * SIZE
add.d XX, XX, INCX
ST t3, XX, 0 * SIZE
add.d XX, XX, INCX
ST t4, XX, 0 * SIZE
add.d XX, XX, INCX
MUL t1, ALPHA, a5
LD a5, X, 0 * SIZE
add.d X, X, INCX
MUL t2, ALPHA, a6
LD a6, X, 0 * SIZE
add.d X, X, INCX
MUL t3, ALPHA, a7
LD a7, X, 0 * SIZE
add.d X, X, INCX
MUL t4, ALPHA, a8
LD a8, X, 0 * SIZE
add.d X, X, INCX
ST t1, XX, 0 * SIZE
add.d XX, XX, INCX
ST t2, XX, 0 * SIZE
add.d XX, XX, INCX
ST t3, XX, 0 * SIZE
add.d XX, XX, INCX
ST t4, XX, 0 * SIZE
addi.d I, I, -1
add.d XX, XX, INCX
blt $r0, I, .L62
.align 3
.L63:
MUL t1, ALPHA, a1
MUL t2, ALPHA, a2
MUL t3, ALPHA, a3
MUL t4, ALPHA, a4
ST t1, XX, 0 * SIZE
add.d XX, XX, INCX
ST t2, XX, 0 * SIZE
add.d XX, XX, INCX
ST t3, XX, 0 * SIZE
add.d XX, XX, INCX
ST t4, XX, 0 * SIZE
add.d XX, XX, INCX
MUL t1, ALPHA, a5
MUL t2, ALPHA, a6
MUL t3, ALPHA, a7
MUL t4, ALPHA, a8
ST t1, XX, 0 * SIZE
add.d XX, XX, INCX
ST t2, XX, 0 * SIZE
add.d XX, XX, INCX
ST t3, XX, 0 * SIZE
add.d XX, XX, INCX
ST t4, XX, 0 * SIZE
add.d XX, XX, INCX
.align 3
.L65:
andi I, N, 7
bge $r0, I, .L999
.align 3
.L66:
LD a1, X, 0 * SIZE
MUL t1, ALPHA, a1
addi.d I, I, -1
ST t1, X, 0 * SIZE
add.d X, X, INCX
blt $r0, I, .L66
.align 3
.L999:
move $r4, $r17
fmov.d $f0, $f22
jirl $r0, $r1, 0x0
EPILOGUE