/*********************************************************************/
/* 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 N r32
#define X1 r33
#define INCX r34
#define Y1 r35
#define INCY r36
#define PREX r2
#define PREY r3
#define I r14
#define J r15
#define X2 r16
#define Y2 r17
#define INCX2 r18
#define INCY2 r19
#define INCX8 r20
#define INCY8 r21
#define PR r30
#define ARLC r31
#define PREFETCH_SIZE (8 * 16)
PROLOGUE
.prologue
PROFCODE
{ .mmi
.save ar.lc, ARLC
mov ARLC = ar.lc
}
{ .mib
cmp.lt p0, p6 = r0, N
shr I = N, 3
(p6) br.ret.sptk.many b0
}
;;
shl INCX = INCX, ZBASE_SHIFT
shl INCY = INCY, ZBASE_SHIFT
;;
.body
{ .mmi
sub r8 = X1, Y1
mov r9 = 0xf0
mov PR = pr
}
{ .mmi
shladd INCX2 = INCX, 1, r0
shladd INCY2 = INCY, 1, r0
and J = 15, N
}
;;
{ .mmi
shladd INCX8 = INCX, 2, r0
shladd INCY8 = INCY, 2, r0
mov pr.rot = 0
}
{ .mmi
and r8 = r9, r8
cmp.eq p9, p0 = r0, J
adds I = -1, I
}
;;
{ .mmi
adds X2 = 1 * SIZE, X1
adds Y2 = 1 * SIZE, Y1
mov ar.ec = 4
}
{ .mmb
cmp.gt p6, p0 = 127, r8
cmp.eq p16, p0 = r0, r0
(p6) br.cond.dpnt .L20
}
;;
{ .mmi
adds PREX = (PREFETCH_SIZE + 0) * SIZE, X1
adds PREY = (PREFETCH_SIZE + 2) * SIZE, Y1
mov ar.lc = I
}
{ .mib
cmp.eq p8 ,p0 = -1, I
tbit.z p0, p12 = N, 2
(p8) br.cond.dpnt .L15
}
;;
.align 16
.L12:
{ .mmi
(p19) STFD [Y1] = f35
(p19) STFD [Y2] = f39
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p17) LDFD f81 = [X1], INCX
(p17) LDFD f85 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f43
(p19) STFD [Y2] = f47
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p17) LDFD f89 = [X1], INCX
(p17) LDFD f93 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f51
(p19) STFD [Y2] = f55
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f32 = [X1], INCX
(p16) LDFD f36 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f59
(p19) STFD [Y2] = f63
(p19) add Y1 = INCY, Y1
}
{ .mmi
lfetch.fault.nt1 [PREX], INCX8
lfetch.fault.excl.nt1 [PREY], INCY8
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p16) LDFD f40 = [X1], INCX
(p16) LDFD f44 = [X2], INCX
nop __LINE__
}
;;
{ .mmi
(p19) STFD [Y1] = f67
(p19) STFD [Y2] = f71
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f48 = [X1], INCX
(p16) LDFD f52 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f75
(p19) STFD [Y2] = f79
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f56 = [X1], INCX
(p16) LDFD f60 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f83
(p19) STFD [Y2] = f87
(p19) add Y1 = INCY, Y1
}
{ .mmi
lfetch.fault.nt1 [PREX], INCX8
lfetch.fault.excl.nt1 [PREY], INCY8
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f91
(p19) STFD [Y2] = f95
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f64 = [X1], INCX
(p16) LDFD f68 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmb
(p16) LDFD f72 = [X1], INCX
(p16) LDFD f76 = [X2], INCX
br.ctop.sptk.few .L12
}
;;
.align 32
.L15:
{ .mmi
(p12) LDFD f48 = [X1], INCX
(p12) LDFD f49 = [X2], INCX
mov ar.lc = ARLC
}
;;
{ .mmi
(p12) LDFD f50 = [X1], INCX
(p12) LDFD f51 = [X2], INCX
mov pr = PR, -65474
}
;;
{ .mmb
(p12) LDFD f52 = [X1], INCX
(p12) LDFD f53 = [X2], INCX
(p9) br.ret.sptk.many b0
}
;;
{ .mmi
(p12) LDFD f54 = [X1], INCX
(p12) LDFD f55 = [X2], INCX
tbit.z p0, p13 = N, 1
}
;;
{ .mmi
(p13) LDFD f56 = [X1], INCX
(p13) LDFD f57 = [X2], INCX
tbit.z p0, p14 = N, 0
}
;;
{ .mmi
(p13) LDFD f58 = [X1], INCX
(p13) LDFD f59 = [X2], INCX
}
;;
{ .mmi
(p12) STFD [Y1] = f48
(p12) STFD [Y2] = f49
(p12) add Y1 = INCY, Y1
}
{ .mmi
(p14) LDFD f60 = [X1], INCX
(p14) LDFD f61 = [X2], INCX
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f50
(p12) STFD [Y2] = f51
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f52
(p12) STFD [Y2] = f53
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f54
(p12) STFD [Y2] = f55
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p13) STFD [Y1] = f56
(p13) STFD [Y2] = f57
(p13) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p13) add Y2 = INCY, Y2
}
;;
{ .mmi
(p13) STFD [Y1] = f58
(p13) STFD [Y2] = f59
(p13) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p13) add Y2 = INCY, Y2
}
;;
{ .mmb
(p14) STFD [Y1] = f60
(p14) STFD [Y2] = f61
br.ret.sptk.many b0
}
;;
.align 16
.L20:
{ .mmi
adds PREX = (PREFETCH_SIZE + 0) * SIZE, X1
adds PREY = (PREFETCH_SIZE + 10) * SIZE, Y1
mov ar.lc = I
}
{ .mib
cmp.eq p8 ,p0 = -1, I
tbit.z p0, p12 = N, 2
(p8) br.cond.dpnt .L25
}
;;
.align 16
.L22:
{ .mmi
(p19) STFD [Y1] = f67
(p19) STFD [Y2] = f71
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p17) LDFD f81 = [X1], INCX
(p17) LDFD f85 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f75
(p19) STFD [Y2] = f79
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p17) LDFD f89 = [X1], INCX
(p17) LDFD f93 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f83
(p19) STFD [Y2] = f87
(p19) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f32 = [X1], INCX
(p16) LDFD f36 = [X2], INCX
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p19) STFD [Y1] = f91
(p19) STFD [Y2] = f95
(p19) add Y1 = INCY, Y1
}
{ .mmi
lfetch.fault.nt1 [PREX], INCX8
lfetch.fault.excl.nt1 [PREY], INCY8
(p19) add Y2 = INCY, Y2
}
;;
{ .mmi
(p16) LDFD f40 = [X1], INCX
(p16) LDFD f44 = [X2], INCX
nop __LINE__
}
;;
{ .mmi
(p18) STFD [Y1] = f34
(p18) STFD [Y2] = f38
(p18) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f48 = [X1], INCX
(p16) LDFD f52 = [X2], INCX
(p18) add Y2 = INCY, Y2
}
;;
{ .mmi
(p18) STFD [Y1] = f42
(p18) STFD [Y2] = f46
(p18) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f56 = [X1], INCX
(p16) LDFD f60 = [X2], INCX
(p18) add Y2 = INCY, Y2
}
;;
{ .mmi
(p18) STFD [Y1] = f50
(p18) STFD [Y2] = f54
(p18) add Y1 = INCY, Y1
}
{ .mmi
lfetch.fault.nt1 [PREX], INCX8
lfetch.fault.excl.nt1 [PREY], INCY8
(p18) add Y2 = INCY, Y2
}
;;
{ .mmi
(p18) STFD [Y1] = f58
(p18) STFD [Y2] = f62
(p18) add Y1 = INCY, Y1
}
{ .mmi
(p16) LDFD f64 = [X1], INCX
(p16) LDFD f68 = [X2], INCX
(p18) add Y2 = INCY, Y2
}
;;
{ .mmb
(p16) LDFD f72 = [X1], INCX
(p16) LDFD f76 = [X2], INCX
br.ctop.sptk.few .L22
}
;;
.align 32
.L25:
{ .mmi
(p12) LDFD f48 = [X1], INCX
(p12) LDFD f49 = [X2], INCX
mov ar.lc = ARLC
}
;;
{ .mmi
(p12) LDFD f50 = [X1], INCX
(p12) LDFD f51 = [X2], INCX
mov pr = PR, -65474
}
;;
{ .mmb
(p12) LDFD f52 = [X1], INCX
(p12) LDFD f53 = [X2], INCX
(p9) br.ret.sptk.many b0
}
;;
{ .mmi
(p12) LDFD f54 = [X1], INCX
(p12) LDFD f55 = [X2], INCX
tbit.z p0, p13 = N, 1
}
;;
{ .mmi
(p13) LDFD f56 = [X1], INCX
(p13) LDFD f57 = [X2], INCX
tbit.z p0, p14 = N, 0
}
;;
{ .mmi
(p13) LDFD f58 = [X1], INCX
(p13) LDFD f59 = [X2], INCX
}
;;
{ .mmi
(p12) STFD [Y1] = f48
(p12) STFD [Y2] = f49
(p12) add Y1 = INCY, Y1
}
{ .mmi
(p14) LDFD f60 = [X1], INCX
(p14) LDFD f61 = [X2], INCX
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f50
(p12) STFD [Y2] = f51
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f52
(p12) STFD [Y2] = f53
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p12) STFD [Y1] = f54
(p12) STFD [Y2] = f55
(p12) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p12) add Y2 = INCY, Y2
}
;;
{ .mmi
(p13) STFD [Y1] = f56
(p13) STFD [Y2] = f57
(p13) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p13) add Y2 = INCY, Y2
}
;;
{ .mmi
(p13) STFD [Y1] = f58
(p13) STFD [Y2] = f59
(p13) add Y1 = INCY, Y1
}
{ .mmi
nop __LINE__
nop __LINE__
(p13) add Y2 = INCY, Y2
}
;;
{ .mmb
(p14) STFD [Y1] = f60
(p14) STFD [Y2] = f61
br.ret.sptk.many b0
}
;;
EPILOGUE