OpenBLAS/kernel/loongarch64/caxpy_lasx.S

708 lines
17 KiB
ArmAsm

/***************************************************************************
Copyright (c) 2023, 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 XX $r5
#define YY $r6
#define ALPHAR $f0
#define ALPHAI $f1
#define X $r7
#define INCX $r8
#define Y $r9
#define INCY $r10
#define I $r12
#define TEMP $r13
#define t1 $r14
#define t2 $r16
#define t3 $r15
#define t4 $r17
#define a1 $f12
#define a2 $f13
#define a3 $f14
#define a4 $f15
#define s1 $f16
#define s2 $f17
#define s3 $f18
#define s4 $f19
#define VX0 $xr8
#define VX1 $xr20
#define VX2 $xr21
#define VX3 $xr22
#define VXAR $xr23
#define VXAI $xr19
#define x1 $xr18
#define x2 $xr17
#define x3 $xr16
#define x4 $xr15
PROLOGUE
bge $r0, N, .L999
li.d TEMP, 1
movgr2fr.d a1, $r0
FFINT a1, a1
CMPEQ $fcc0, ALPHAR, a1
CMPEQ $fcc1, ALPHAI, a1
bceqz $fcc0, .L10
bcnez $fcc1, .L999
.L10:
slli.d TEMP, TEMP, ZBASE_SHIFT
slli.d INCX, INCX, ZBASE_SHIFT
slli.d INCY, INCY, ZBASE_SHIFT
MTG t1, ALPHAR
MTG t2, ALPHAI
#ifdef DOUBLE
xvreplgr2vr.d VXAR, t1
xvreplgr2vr.d VXAI, t2
srai.d I, N, 2
#else
xvreplgr2vr.w VXAR, t1
xvreplgr2vr.w VXAI, t2
srai.d I, N, 3
#endif
bne INCX, TEMP, .L20
bne INCY, TEMP, .L12 // INCX==1 and INCY!=1
b .L11 // INCX==1 and INCY==1
.L20:
bne INCY, TEMP, .L22 // INCX!=1 and INCY!=1
b .L21 // INCX!=1 and INCY==1
.L11:
bge $r0, I, .L997
.align 3
.L111:
#ifdef DOUBLE
xvld VX0, X, 0 * SIZE
xvld VX2, Y, 0 * SIZE
xvld VX1, X, 4 * SIZE
xvld VX3, Y, 4 * SIZE
xvpickev.d x1, VX1, VX0
xvpickod.d x2, VX1, VX0
xvpickev.d x3, VX3, VX2
xvpickod.d x4, VX3, VX2
#else
xvld VX0, X, 0 * SIZE
xvld VX2, Y, 0 * SIZE
xvld VX1, X, 8 * SIZE
xvld VX3, Y, 8 * SIZE
xvpickev.w x1, VX1, VX0
xvpickod.w x2, VX1, VX0
xvpickev.w x3, VX3, VX2
xvpickod.w x4, VX3, VX2
#endif
#if !defined(CONJ)
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmsub.d VX1, VXAR, x1, VX0
xvfmadd.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfadd.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmsub.s VX1, VXAR, x1, VX0
xvfmadd.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfadd.s x4, x4, VX3
#endif
#else
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmadd.d VX1, VXAR, x1, VX0
xvfmsub.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfsub.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmadd.s VX1, VXAR, x1, VX0
xvfmsub.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfsub.s x4, x4, VX3
#endif
#endif
#ifdef DOUBLE
xvilvl.d VX2, x4 ,x3
xvilvh.d VX3, x4, x3
xvst VX2, Y, 0 * SIZE
xvst VX3, Y, 4 * SIZE
addi.d X, X, 8 * SIZE
addi.d Y, Y, 8 * SIZE
#else
xvilvl.w VX2, x4 ,x3
xvilvh.w VX3, x4, x3
xvst VX2, Y, 0 * SIZE
xvst VX3, Y, 8 * SIZE
addi.d X, X, 16 * SIZE
addi.d Y, Y, 16 * SIZE
#endif
addi.d I, I, -1
blt $r0, I, .L111
b .L997
.align 3
.L12: // INCX==1 and INCY!=1
bge $r0, I, .L997
move YY, Y
.align 3
.L121:
#ifdef DOUBLE
xvld VX0, X, 0 * SIZE
xvld VX1, X, 4 * SIZE
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.d x3, t1, 0
xvinsgr2vr.d x4, t2, 0
xvinsgr2vr.d x3, t3, 2
xvinsgr2vr.d x4, t4, 2
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
xvinsgr2vr.d x3, t1, 1
xvinsgr2vr.d x4, t2, 1
xvinsgr2vr.d x3, t3, 3
xvinsgr2vr.d x4, t4, 3
add.d Y, Y, INCY
xvpickev.d x1, VX1, VX0
xvpickod.d x2, VX1, VX0
#else
xvld VX0, X, 0 * SIZE
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 0
xvinsgr2vr.w x4, t2, 0
xvinsgr2vr.w x3, t3, 1
xvinsgr2vr.w x4, t4, 1
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 4
xvinsgr2vr.w x4, t2, 4
xvinsgr2vr.w x3, t3, 5
xvinsgr2vr.w x4, t4, 5
xvld VX1, X, 8 * SIZE
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 2
xvinsgr2vr.w x4, t2, 2
xvinsgr2vr.w x3, t3, 3
xvinsgr2vr.w x4, t4, 3
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
xvinsgr2vr.w x3, t1, 6
xvinsgr2vr.w x4, t2, 6
xvinsgr2vr.w x3, t3, 7
xvinsgr2vr.w x4, t4, 7
add.d Y, Y, INCY
xvpickev.w x1, VX1, VX0
xvpickod.w x2, VX1, VX0
#endif
#if !defined(CONJ)
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmsub.d VX1, VXAR, x1, VX0
xvfmadd.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfadd.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmsub.s VX1, VXAR, x1, VX0
xvfmadd.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfadd.s x4, x4, VX3
#endif
#else
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmadd.d VX1, VXAR, x1, VX0
xvfmsub.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfsub.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmadd.s VX1, VXAR, x1, VX0
xvfmsub.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfsub.s x4, x4, VX3
#endif
#endif
#ifdef DOUBLE
xvstelm.d x3, YY, 0 * SIZE, 0
xvstelm.d x4, YY, 1 * SIZE, 0
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 2
xvstelm.d x4, YY, 1 * SIZE, 2
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 1
xvstelm.d x4, YY, 1 * SIZE, 1
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 3
xvstelm.d x4, YY, 1 * SIZE, 3
add.d YY, YY, INCY
addi.d X, X, 8 * SIZE
addi.d I, I, -1
#else
addi.d I, I, -1
xvstelm.w x3, YY, 0 * SIZE, 0
xvstelm.w x4, YY, 1 * SIZE, 0
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 1
xvstelm.w x4, YY, 1 * SIZE, 1
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 4
xvstelm.w x4, YY, 1 * SIZE, 4
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 5
xvstelm.w x4, YY, 1 * SIZE, 5
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 2
xvstelm.w x4, YY, 1 * SIZE, 2
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 3
xvstelm.w x4, YY, 1 * SIZE, 3
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 6
xvstelm.w x4, YY, 1 * SIZE, 6
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 7
xvstelm.w x4, YY, 1 * SIZE, 7
add.d YY, YY, INCY
addi.d X, X, 16 * SIZE
#endif
blt $r0, I, .L121
b .L997
.align 3
.L21:// INCX!=1 and INCY==1
bge $r0, I, .L997
.align 3
.L211:
#ifdef DOUBLE
xvld VX2, Y, 0 * SIZE
xvld VX3, Y, 4 * SIZE
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.d x1, t1, 0
xvinsgr2vr.d x2, t2, 0
xvinsgr2vr.d x1, t3, 2
xvinsgr2vr.d x2, t4, 2
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
xvinsgr2vr.d x1, t1, 1
xvinsgr2vr.d x2, t2, 1
xvinsgr2vr.d x1, t3, 3
xvinsgr2vr.d x2, t4, 3
add.d X, X, INCX
xvpickev.d x3, VX3, VX2
xvpickod.d x4, VX3, VX2
#else
xvld VX2, Y, 0 * SIZE
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 0
xvinsgr2vr.w x2, t2, 0
xvinsgr2vr.w x1, t3, 1
xvinsgr2vr.w x2, t4, 1
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 4
xvinsgr2vr.w x2, t2, 4
xvinsgr2vr.w x1, t3, 5
xvinsgr2vr.w x2, t4, 5
xvld VX3, Y, 8 * SIZE
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 2
xvinsgr2vr.w x2, t2, 2
xvinsgr2vr.w x1, t3, 3
xvinsgr2vr.w x2, t4, 3
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 6
xvinsgr2vr.w x2, t2, 6
xvinsgr2vr.w x1, t3, 7
xvinsgr2vr.w x2, t4, 7
xvpickev.w x3, VX3, VX2
xvpickod.w x4, VX3, VX2
#endif
#if !defined(CONJ)
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmsub.d VX1, VXAR, x1, VX0
xvfmadd.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfadd.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmsub.s VX1, VXAR, x1, VX0
xvfmadd.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfadd.s x4, x4, VX3
#endif
#else
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmadd.d VX1, VXAR, x1, VX0
xvfmsub.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfsub.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmadd.s VX1, VXAR, x1, VX0
xvfmsub.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfsub.s x4, x4, VX3
#endif
#endif
#ifdef DOUBLE
xvilvl.d VX2, x4 ,x3
xvilvh.d VX3, x4, x3
addi.d I, I, -1
xvst VX2, Y, 0 * SIZE
xvst VX3, Y, 4 * SIZE
addi.d Y, Y, 8 * SIZE
#else
xvilvl.w VX2, x4 ,x3
xvilvh.w VX3, x4, x3
addi.d I, I, -1
xvst VX2, Y, 0 * SIZE
xvst VX3, Y, 8 * SIZE
addi.d Y, Y, 16 * SIZE
#endif
blt $r0, I, .L211
b .L997
.align 3
.L22:
bge $r0, I, .L997
move YY, Y
.align 3
.L222:
#ifdef DOUBLE
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.d x1, t1, 0
xvinsgr2vr.d x2, t2, 0
xvinsgr2vr.d x1, t3, 1
xvinsgr2vr.d x2, t4, 1
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.d x1, t1, 2
xvinsgr2vr.d x2, t2, 2
xvinsgr2vr.d x1, t3, 3
xvinsgr2vr.d x2, t4, 3
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.d x3, t1, 0
xvinsgr2vr.d x4, t2, 0
xvinsgr2vr.d x3, t3, 1
xvinsgr2vr.d x4, t4, 1
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.d x3, t1, 2
xvinsgr2vr.d x4, t2, 2
xvinsgr2vr.d x3, t3, 3
xvinsgr2vr.d x4, t4, 3
#else
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 0
xvinsgr2vr.w x2, t2, 0
xvinsgr2vr.w x1, t3, 1
xvinsgr2vr.w x2, t4, 1
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 2
xvinsgr2vr.w x2, t2, 2
xvinsgr2vr.w x1, t3, 3
xvinsgr2vr.w x2, t4, 3
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 0
xvinsgr2vr.w x4, t2, 0
xvinsgr2vr.w x3, t3, 1
xvinsgr2vr.w x4, t4, 1
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 2
xvinsgr2vr.w x4, t2, 2
xvinsgr2vr.w x3, t3, 3
xvinsgr2vr.w x4, t4, 3
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 4
xvinsgr2vr.w x2, t2, 4
xvinsgr2vr.w x1, t3, 5
xvinsgr2vr.w x2, t4, 5
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
xvinsgr2vr.w x1, t1, 6
xvinsgr2vr.w x2, t2, 6
xvinsgr2vr.w x1, t3, 7
xvinsgr2vr.w x2, t4, 7
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 4
xvinsgr2vr.w x4, t2, 4
xvinsgr2vr.w x3, t3, 5
xvinsgr2vr.w x4, t4, 5
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
xvinsgr2vr.w x3, t1, 6
xvinsgr2vr.w x4, t2, 6
xvinsgr2vr.w x3, t3, 7
xvinsgr2vr.w x4, t4, 7
#endif
#if !defined(CONJ)
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmsub.d VX1, VXAR, x1, VX0
xvfmadd.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfadd.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmsub.s VX1, VXAR, x1, VX0
xvfmadd.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfadd.s x4, x4, VX3
#endif
#else
#ifdef DOUBLE
xvfmul.d VX0, VXAI, x2
xvfmul.d VX2, VXAI, x1
xvfmadd.d VX1, VXAR, x1, VX0
xvfmsub.d VX3, x2, VXAR, VX2
xvfadd.d x3, x3, VX1
xvfsub.d x4, x4, VX3
#else
xvfmul.s VX0, VXAI, x2
xvfmul.s VX2, VXAI, x1
xvfmadd.s VX1, VXAR, x1, VX0
xvfmsub.s VX3, x2, VXAR, VX2
xvfadd.s x3, x3, VX1
xvfsub.s x4, x4, VX3
#endif
#endif
addi.d I, I, -1
#ifdef DOUBLE
xvstelm.d x3, YY, 0 * SIZE, 0
xvstelm.d x4, YY, 1 * SIZE, 0
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 1
xvstelm.d x4, YY, 1 * SIZE, 1
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 2
xvstelm.d x4, YY, 1 * SIZE, 2
add.d YY, YY, INCY
xvstelm.d x3, YY, 0 * SIZE, 3
xvstelm.d x4, YY, 1 * SIZE, 3
#else
xvstelm.w x3, YY, 0 * SIZE, 0
xvstelm.w x4, YY, 1 * SIZE, 0
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 1
xvstelm.w x4, YY, 1 * SIZE, 1
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 2
xvstelm.w x4, YY, 1 * SIZE, 2
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 3
xvstelm.w x4, YY, 1 * SIZE, 3
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 4
xvstelm.w x4, YY, 1 * SIZE, 4
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 5
xvstelm.w x4, YY, 1 * SIZE, 5
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 6
xvstelm.w x4, YY, 1 * SIZE, 6
add.d YY, YY, INCY
xvstelm.w x3, YY, 0 * SIZE, 7
xvstelm.w x4, YY, 1 * SIZE, 7
#endif
add.d YY, YY, INCY
blt $r0, I, .L222
.align 3
.L997:
#ifdef DOUBLE
andi I, N, 3
#else
andi I, N, 7
#endif
bge $r0, I, .L999
.align 3
.L998:
LD a1, X, 0 * SIZE
LD a2, X, 1 * SIZE
LD a3, Y, 0 * SIZE
LD a4, Y, 1 * SIZE
addi.d I, I, -1
#if !defined(CONJ)
MUL s1, ALPHAI, a2
MUL s2, ALPHAI, a1
MSUB s3, ALPHAR, a1, s1
MADD s4, a2, ALPHAR, s2
ADD s3, s3, a3
ADD s4, s4, a4
#else
MUL s1, ALPHAI, a2
MUL s2, ALPHAI, a1
MADD s3, ALPHAR, a1, s1
MSUB s4, a2, ALPHAR, s2
ADD s3, s3, a3
SUB s4, a4, s4
#endif
ST s3, Y, 0 * SIZE
ST s4, Y, 1 * SIZE
add.d X, X, INCX
add.d Y, Y, INCY
blt $r0, I, .L998
.align 3
.L999:
move $r4, $r12
jirl $r0, $r1, 0x0
.align 3
EPILOGUE