218 lines
5.3 KiB
ArmAsm
218 lines
5.3 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 $r5
|
|
#define INCX $r6
|
|
#define I $r18
|
|
#define TEMP $r7
|
|
#define a1 $f10
|
|
#define a2 $f11
|
|
#define a3 $f12
|
|
#define a4 $f13
|
|
#define a5 $f14
|
|
#define a6 $f15
|
|
#define a7 $f16
|
|
#define a8 $f17
|
|
#define t1 $f0
|
|
#define t2 $f1
|
|
#define t3 $f2
|
|
#define t4 $f3
|
|
#define t5 $f4
|
|
#define t6 $f5
|
|
#define t7 $f6
|
|
#define t8 $f7
|
|
#define s1 $f22
|
|
#define s2 $f8
|
|
#define s3 $f23
|
|
#define s4 $f9
|
|
#define x1 $r17
|
|
#define x2 $r8
|
|
#define x3 $r9
|
|
#define x4 $r10
|
|
|
|
PROLOGUE
|
|
|
|
#ifdef F_INTERFACE
|
|
LDINT N, 0(N)
|
|
LDINT INCX, 0(INCX)
|
|
#endif
|
|
|
|
li.d x1, 0
|
|
bge $r0, N, .L999
|
|
slli.d INCX, INCX, ZBASE_SHIFT
|
|
bge $r0, INCX, .L999
|
|
LD a1, X, 0 * SIZE
|
|
LD a2, X, 1 * SIZE
|
|
FABS t1, a1
|
|
FABS t2, a2
|
|
ADD s1, t1, t2
|
|
ADD s2, t1, t2
|
|
ADD s3, t1, t2
|
|
ADD s4, t1, t2
|
|
addi.d N, N, -1
|
|
li.d x1, 1
|
|
bge $r0, N, .L999
|
|
add.d X, X, INCX
|
|
li.d x2, 1
|
|
srai.d I, N, 2
|
|
li.d x3, 1
|
|
li.d TEMP, 2
|
|
li.d x4, 1
|
|
bge $r0, I, .L15
|
|
LD a1, X, 0 * SIZE
|
|
LD a2, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
LD a3, X, 0 * SIZE
|
|
LD a4, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
LD a5, X, 0 * SIZE
|
|
LD a6, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
LD a7, X, 0 * SIZE
|
|
LD a8, X, 1 * SIZE
|
|
addi.d I, I, -1
|
|
add.d X, X, INCX
|
|
bge $r0, I, .L13
|
|
.align 3
|
|
|
|
.L12:
|
|
FABS t1, a1
|
|
LD a1, X, 0 * SIZE
|
|
FABS t2, a2
|
|
LD a2, X, 1 * SIZE
|
|
FABS t3, a3
|
|
add.d X, X, INCX
|
|
FABS t4, a4
|
|
FABS t5, a5
|
|
LD a3, X, 0 * SIZE
|
|
FABS t6, a6
|
|
LD a4, X, 1 * SIZE
|
|
FABS t7, a7
|
|
add.d X, X, INCX
|
|
FABS t8, a8
|
|
ADD t1, t1, t2
|
|
LD a5, X, 0 * SIZE
|
|
ADD t3, t3, t4
|
|
LD a6, X, 1 * SIZE
|
|
ADD t5, t5, t6
|
|
add.d X, X, INCX
|
|
ADD t7, t7, t8
|
|
CMPLT $fcc0, s1, t1
|
|
LD a7, X, 0 * SIZE
|
|
CMPLT $fcc1, s2, t3
|
|
LD a8, X, 1 * SIZE
|
|
CMPLT $fcc2, s3, t5
|
|
add.d X, X, INCX
|
|
CMPLT $fcc3, s4, t7
|
|
addi.d I, I, -1
|
|
CMOVT s1, s1, t1, $fcc0
|
|
MOVT(x1, TEMP, $fcc0)
|
|
CMOVT s2, s2, t3, $fcc1
|
|
MOVT(x2, TEMP, $fcc1)
|
|
CMOVT s3, s3, t5, $fcc2
|
|
MOVT(x3, TEMP, $fcc2)
|
|
CMOVT s4, s4, t7, $fcc3
|
|
MOVT(x4, TEMP, $fcc3)
|
|
addi.d TEMP, TEMP, 4
|
|
blt $r0, I, .L12
|
|
.align 3
|
|
|
|
.L13:
|
|
FABS t1, a1
|
|
FABS t2, a2
|
|
FABS t3, a3
|
|
FABS t4, a4
|
|
FABS t5, a5
|
|
FABS t6, a6
|
|
FABS t7, a7
|
|
FABS t8, a8
|
|
ADD t1, t1, t2
|
|
ADD t3, t3, t4
|
|
ADD t5, t5, t6
|
|
ADD t7, t7, t8
|
|
CMPLT $fcc0, s1, t1
|
|
CMPLT $fcc1, s2, t3
|
|
CMPLT $fcc2, s3, t5
|
|
CMPLT $fcc3, s4, t7
|
|
CMOVT s1, s1, t1, $fcc0
|
|
MOVT(x1, TEMP, $fcc0)
|
|
CMOVT s2, s2, t3, $fcc1
|
|
MOVT(x2, TEMP, $fcc1)
|
|
CMOVT s3, s3, t5, $fcc2
|
|
MOVT(x3, TEMP, $fcc2)
|
|
CMOVT s4, s4, t7, $fcc3
|
|
MOVT(x4, TEMP, $fcc3)
|
|
addi.d TEMP, TEMP, 4
|
|
addi.d x2, x2, 1
|
|
addi.d x3, x3, 2
|
|
addi.d x4, x4, 3
|
|
.align 3
|
|
|
|
.L15:
|
|
andi I, N, 3
|
|
bge $r0, I, .L998
|
|
.align 3
|
|
|
|
.L16:
|
|
LD a1, X, 0 * SIZE
|
|
LD a2, X, 1 * SIZE
|
|
add.d X, X, INCX
|
|
FABS t1, a1
|
|
FABS t2, a2
|
|
ADD t1, t1, t2
|
|
addi.d I, I, -1
|
|
CMPLT $fcc0, s1, t1
|
|
CMOVT s1, s1, t1, $fcc0
|
|
MOVT(x1, TEMP, $fcc0)
|
|
addi.d TEMP, TEMP, 1
|
|
blt $r0, I, .L16
|
|
.align 3
|
|
|
|
.L998:
|
|
CMPLT $fcc0, s1, s2
|
|
CMPLT $fcc1, s3, s4
|
|
CMOVT s1, s1, s2, $fcc0
|
|
MOVT(x1, x2, $fcc0)
|
|
CMOVT s3, s3, s4, $fcc1
|
|
MOVT(x3, x4, $fcc1)
|
|
CMPLT $fcc0, s1, s3
|
|
CMOVT s1, s1, s3, $fcc0
|
|
MOVT(x1, x3, $fcc0)
|
|
.align 3
|
|
|
|
.L999:
|
|
move $r4, $r17
|
|
fmov.d $f0, $f22
|
|
jirl $r0, $r1, 0x0
|
|
|
|
EPILOGUE
|