Annotation of OpenXM_contrib/gmp/mpn/pa32/hppa1_1/sqr_diagonal.asm, Revision 1.1.1.1
1.1 ohara 1: dnl HP-PA 1.1 32-bit mpn_sqr_diagonal.
2:
3: dnl Copyright 2001, 2002 Free Software Foundation, Inc.
4:
5: dnl This file is part of the GNU MP Library.
6:
7: dnl The GNU MP Library is free software; you can redistribute it and/or modify
8: dnl it under the terms of the GNU Lesser General Public License as published
9: dnl by the Free Software Foundation; either version 2.1 of the License, or (at
10: dnl your option) any later version.
11:
12: dnl The GNU MP Library is distributed in the hope that it will be useful, but
13: dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
14: dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
15: dnl License for more details.
16:
17: dnl You should have received a copy of the GNU Lesser General Public License
18: dnl along with the GNU MP Library; see the file COPYING.LIB. If not, write to
19: dnl the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
20: dnl MA 02111-1307, USA.
21:
22: include(`../config.m4')
23:
24: C This code runs at 6 cycles/limb on the PA7100 and 2.5 cycles/limb on PA8x00.
25: C 2-way unrolling wouldn't help the PA7100; it could however bring times down
26: C to 2.0 cycles/limb for the PA8x00.
27:
28: C INPUT PARAMETERS
29: define(`rp',`%r26')
30: define(`up',`%r25')
31: define(`n',`%r24')
32:
33: ASM_START()
34: PROLOGUE(mpn_sqr_diagonal)
35: ldo 4(rp),rp
36: fldws,ma 4(up),%fr4r
37: addib,= -1,n,L(exit)
38: xmpyu %fr4r,%fr4r,%fr5
39:
40: .label L(loop)
41: fldws,ma 4(up),%fr4r
42: fstws %fr5r,-4(rp)
43: fstws,ma %fr5l,8(rp)
44: addib,<> -1,n,L(loop)
45: xmpyu %fr4r,%fr4r,%fr5
46:
47: .label L(exit)
48: fstws %fr5r,-4(rp)
49: bv 0(%r2)
50: fstws %fr5l,0(rp)
51: EPILOGUE(mpn_sqr_diagonal)
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>