beta-0.89.2
[luatex.git] / source / libs / gmp / gmp-src / mpn / power / sub_n.asm
blob390c802d8bca5bc32f74e169cc1a3e2dabbc945a
1 dnl IBM POWER mpn_sub_n -- Subtract two limb vectors of equal, non-zero
2 dnl length.
4 dnl Copyright 1992, 1994-1996, 1999-2001, 2005 Free Software Foundation, Inc.
6 dnl This file is part of the GNU MP Library.
7 dnl
8 dnl The GNU MP Library is free software; you can redistribute it and/or modify
9 dnl it under the terms of either:
10 dnl
11 dnl * the GNU Lesser General Public License as published by the Free
12 dnl Software Foundation; either version 3 of the License, or (at your
13 dnl option) any later version.
14 dnl
15 dnl or
16 dnl
17 dnl * the GNU General Public License as published by the Free Software
18 dnl Foundation; either version 2 of the License, or (at your option) any
19 dnl later version.
20 dnl
21 dnl or both in parallel, as here.
22 dnl
23 dnl The GNU MP Library is distributed in the hope that it will be useful, but
24 dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
25 dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
26 dnl for more details.
27 dnl
28 dnl You should have received copies of the GNU General Public License and the
29 dnl GNU Lesser General Public License along with the GNU MP Library. If not,
30 dnl see https://www.gnu.org/licenses/.
33 dnl INPUT PARAMETERS
34 dnl res_ptr r3
35 dnl s1_ptr r4
36 dnl s2_ptr r5
37 dnl size r6
39 include(`../config.m4')
41 ASM_START()
42 PROLOGUE(mpn_sub_n)
43 andil. 10,6,1 C odd or even number of limbs?
44 l 8,0(4) C load least significant s1 limb
45 l 0,0(5) C load least significant s2 limb
46 cal 3,-4(3) C offset res_ptr, it's updated before it's used
47 sri 10,6,1 C count for unrolled loop
48 sf 7,0,8 C subtract least significant limbs, set cy
49 mtctr 10 C copy count into CTR
50 beq 0,Leven C branch if even # of limbs (# of limbs >= 2)
52 C We have an odd # of limbs. Add the first limbs separately.
53 cmpi 1,10,0 C is count for unrolled loop zero?
54 bc 4,6,L1 C bne cr1,L1 (misassembled by gas)
55 st 7,4(3)
56 sfe 3,0,0 C load !cy into ...
57 sfi 3,3,0 C ... return value register
58 br C return
60 C We added least significant limbs. Now reload the next limbs to enter loop.
61 L1: lu 8,4(4) C load s1 limb and update s1_ptr
62 lu 0,4(5) C load s2 limb and update s2_ptr
63 stu 7,4(3)
64 sfe 7,0,8 C subtract limbs, set cy
65 Leven: lu 9,4(4) C load s1 limb and update s1_ptr
66 lu 10,4(5) C load s2 limb and update s2_ptr
67 bdz Lend C If done, skip loop
69 Loop: lu 8,4(4) C load s1 limb and update s1_ptr
70 lu 0,4(5) C load s2 limb and update s2_ptr
71 sfe 11,10,9 C subtract previous limbs with cy, set cy
72 stu 7,4(3) C
73 lu 9,4(4) C load s1 limb and update s1_ptr
74 lu 10,4(5) C load s2 limb and update s2_ptr
75 sfe 7,0,8 C subtract previous limbs with cy, set cy
76 stu 11,4(3) C
77 bdn Loop C decrement CTR and loop back
79 Lend: sfe 11,10,9 C subtract limbs with cy, set cy
80 st 7,4(3) C
81 st 11,8(3) C
82 sfe 3,0,0 C load !cy into ...
83 sfi 3,3,0 C ... return value register
85 EPILOGUE(mpn_sub_n)