add_n.asm revision 1.1.1.1 1 dnl SPARC v9 mpn_add_n for T3/T4.
2
3 dnl Contributed to the GNU project by David Miller.
4
5 dnl Copyright 2013 Free Software Foundation, Inc.
6
7 dnl This file is part of the GNU MP Library.
8 dnl
9 dnl The GNU MP Library is free software; you can redistribute it and/or modify
10 dnl it under the terms of either:
11 dnl
12 dnl * the GNU Lesser General Public License as published by the Free
13 dnl Software Foundation; either version 3 of the License, or (at your
14 dnl option) any later version.
15 dnl
16 dnl or
17 dnl
18 dnl * the GNU General Public License as published by the Free Software
19 dnl Foundation; either version 2 of the License, or (at your option) any
20 dnl later version.
21 dnl
22 dnl or both in parallel, as here.
23 dnl
24 dnl The GNU MP Library is distributed in the hope that it will be useful, but
25 dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
26 dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
27 dnl for more details.
28 dnl
29 dnl You should have received copies of the GNU General Public License and the
30 dnl GNU Lesser General Public License along with the GNU MP Library. If not,
31 dnl see https://www.gnu.org/licenses/.
32
33 include(`../config.m4')
34
35 C cycles/limb
36 C UltraSPARC T3: 8
37 C UltraSPARC T4: 3
38
39 C INPUT PARAMETERS
40 define(`rp', `%i0')
41 define(`up', `%i1')
42 define(`vp', `%i2')
43 define(`n', `%i3')
44 define(`cy', `%i4')
45
46 define(`u0_off', `%l2')
47 define(`u1_off', `%l3')
48 define(`loop_n', `%l6')
49 define(`tmp', `%l7')
50
51 ASM_START()
52 REGISTER(%g2,#scratch)
53 REGISTER(%g3,#scratch)
54 PROLOGUE(mpn_add_nc)
55 save %sp, -176, %sp
56 b,a L(ent)
57 EPILOGUE()
58 PROLOGUE(mpn_add_n)
59 save %sp, -176, %sp
60
61 mov 0, cy
62 L(ent):
63 subcc n, 1, n
64 be L(final_one)
65 cmp %g0, cy
66
67 ldx [up + 0], %o4
68 sllx n, 3, tmp
69
70 ldx [vp + 0], %o5
71 add up, tmp, u0_off
72
73 ldx [up + 8], %g5
74 neg tmp, loop_n
75
76 ldx [vp + 8], %g1
77 add u0_off, 8, u1_off
78
79 sub loop_n, -(2 * 8), loop_n
80
81 brgez,pn loop_n, L(loop_tail)
82 add vp, (2 * 8), vp
83
84 b,a L(top)
85 ALIGN(16)
86 L(top):
87 addxccc(%o4, %o5, tmp)
88 ldx [vp + 0], %o5
89
90 add rp, (2 * 8), rp
91 ldx [loop_n + u0_off], %o4
92
93 add vp, (2 * 8), vp
94 stx tmp, [rp - 16]
95
96 addxccc(%g1, %g5, tmp)
97 ldx [vp - 8], %g1
98
99 ldx [loop_n + u1_off], %g5
100 sub loop_n, -(2 * 8), loop_n
101
102 brlz loop_n, L(top)
103 stx tmp, [rp - 8]
104
105 L(loop_tail):
106 addxccc(%o4, %o5, %g3)
107 add loop_n, u0_off, up
108
109 addxccc(%g1, %g5, %g5)
110 stx %g3, [rp + 0]
111
112 brgz,pt loop_n, L(done)
113 stx %g5, [rp + 8]
114
115 add rp, (2 * 8), rp
116 L(final_one):
117 ldx [up+0], %o4
118 ldx [vp+0], %o5
119 addxccc(%o4, %o5, %g3)
120 stx %g3, [rp+0]
121
122 L(done):
123 addxc(%g0, %g0, %i0)
124 ret
125 restore
126 EPILOGUE()
127