Debian Codegraph @

gmp4/mpn/x86_64/core2/divrem_1.asm

raw
1 dnl x86-64 mpn_divrem_1 -- mpn by limb division.
2
3 dnl Copyright 2004, 2005, 2007-2010, 2012 Free Software Foundation, Inc.
4
5 dnl This file is part of the GNU MP Library.
6 dnl
7 dnl The GNU MP Library is free software; you can redistribute it and/or modify
8 dnl it under the terms of either:
9 dnl
10 dnl * the GNU Lesser General Public License as published by the Free
11 dnl Software Foundation; either version 3 of the License, or (at your
12 dnl option) any later version.
13 dnl
14 dnl or
15 dnl
16 dnl * the GNU General Public License as published by the Free Software
17 dnl Foundation; either version 2 of the License, or (at your option) any
18 dnl later version.
19 dnl
20 dnl or both in parallel, as here.
21 dnl
22 dnl The GNU MP Library is distributed in the hope that it will be useful, but
23 dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
24 dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
25 dnl for more details.
26 dnl
27 dnl You should have received copies of the GNU General Public License and the
28 dnl GNU Lesser General Public License along with the GNU MP Library. If not,
29 dnl see https://www.gnu.org/licenses/.
30
31 include(`../config.m4')
32
33
34 C norm unorm frac
35 C AMD K8,K9 15 15 12
36 C AMD K10 15 15 12
37 C Intel P4 44 44 43
38 C Intel core2 24 24 19.5
39 C Intel corei 19 19 18
40 C Intel atom 51 51 36
41 C VIA nano 46 44 22.5
42
43 C mp_limb_t
44 C mpn_divrem_1 (mp_ptr qp, mp_size_t fn,
45 C mp_srcptr np, mp_size_t nn, mp_limb_t d)
46
47 C mp_limb_t
48 C mpn_preinv_divrem_1 (mp_ptr qp, mp_size_t fn,
49 C mp_srcptr np, mp_size_t nn, mp_limb_t d,
50 C mp_limb_t dinv, int cnt)
51
52 C INPUT PARAMETERS
53 define(`qp', `%rdi')
54 define(`fn_param', `%rsi')
55 define(`up_param', `%rdx')
56 define(`un_param', `%rcx')
57 define(`d', `%r8')
58 define(`dinv', `%r9') C only for mpn_preinv_divrem_1
59 C shift passed on stack C only for mpn_preinv_divrem_1
60
61 define(`cnt', `%rcx')
62 define(`up', `%rsi')
63 define(`fn', `%r12')
64 define(`un', `%rbx')
65
66
67 C rax rbx rcx rdx rsi rdi rbp r8 r9 r10 r11 r12 r13 r14 r15
68 C cnt qp d dinv
69
70 ABI_SUPPORT(DOS64)
71 ABI_SUPPORT(STD64)
72
73 IFSTD(`define(`CNTOFF', `40($1)')')
74 IFDOS(`define(`CNTOFF', `104($1)')')
75
76 ASM_START()
77 TEXT
78 ALIGN(16)
79 PROLOGUE(mpn_preinv_divrem_1)
80 FUNC_ENTRY(4)
81 IFDOS(` mov 56(%rsp), %r8 ')
82 IFDOS(` mov 64(%rsp), %r9 ')
83 xor R32(%rax), R32(%rax)
84 push %r13
85 push %r12
86 push %rbp
87 push %rbx
88
89 mov fn_param, fn
90 mov un_param, un
91 add fn_param, un_param
92 mov up_param, up
93
94 lea -8(qp,un_param,8), qp
95
96 mov CNTOFF(%rsp), R8(cnt)
97 shl R8(cnt), d
98 jmp L(ent)
99 EPILOGUE()
100
101 ALIGN(16)
102 PROLOGUE(mpn_divrem_1)
103 FUNC_ENTRY(4)
104 IFDOS(` mov 56(%rsp), %r8 ')
105 xor R32(%rax), R32(%rax)
106 push %r13
107 push %r12
108 push %rbp
109 push %rbx
110
111 mov fn_param, fn
112 mov un_param, un
113 add fn_param, un_param
114 mov up_param, up
115 je L(ret)
116
117 lea -8(qp,un_param,8), qp
118 xor R32(%rbp), R32(%rbp)
119
120 L(unnormalized):
121 test un, un
122 je L(44)
123 mov -8(up,un,8), %rax
124 cmp d, %rax
125 jae L(44)
126 mov %rbp, (qp)
127 mov %rax, %rbp
128 lea -8(qp), qp
129 je L(ret)
130 dec un
131 L(44):
132 bsr d, %rcx
133 not R32(%rcx)
134 sal R8(%rcx), d
135 sal R8(%rcx), %rbp
136
137 push %rcx
138 IFSTD(` push %rdi ')
139 IFSTD(` push %rsi ')
140 push %r8
141 IFSTD(` mov d, %rdi ')
142 IFDOS(` mov d, %rcx ')
143 CALL( mpn_invert_limb)
144 pop %r8
145 IFSTD(` pop %rsi ')
146 IFSTD(` pop %rdi ')
147 pop %rcx
148
149 mov %rax, dinv
150 mov %rbp, %rax
151 test un, un
152 je L(frac)
153 L(ent): mov -8(up,un,8), %rbp
154 shr R8(%rcx), %rax
155 shld R8(%rcx), %rbp, %rax
156 sub $2, un
157 js L(end)
158
159 ALIGN(16)
160 L(top): lea 1(%rax), %r11
161 mul dinv
162 mov (up,un,8), %r10
163 shld R8(%rcx), %r10, %rbp
164 mov %rbp, %r13
165 add %rax, %r13
166 adc %r11, %rdx
167 mov %rdx, %r11
168 imul d, %rdx
169 sub %rdx, %rbp
170 lea (d,%rbp), %rax
171 sub $8, qp
172 cmp %r13, %rbp
173 cmovc %rbp, %rax
174 adc $-1, %r11
175 cmp d, %rax
176 jae L(ufx)
177 L(uok): dec un
178 mov %r11, 8(qp)
179 mov %r10, %rbp
180 jns L(top)
181
182 L(end): lea 1(%rax), %r11
183 sal R8(%rcx), %rbp
184 mul dinv
185 add %rbp, %rax
186 adc %r11, %rdx
187 mov %rax, %r11
188 mov %rdx, %r13
189 imul d, %rdx
190 sub %rdx, %rbp
191 mov d, %rax
192 add %rbp, %rax
193 cmp %r11, %rbp
194 cmovc %rbp, %rax
195 adc $-1, %r13
196 cmp d, %rax
197 jae L(efx)
198 L(eok): mov %r13, (qp)
199 sub $8, qp
200 jmp L(frac)
201
202 L(ufx): sub d, %rax
203 inc %r11
204 jmp L(uok)
205 L(efx): sub d, %rax
206 inc %r13
207 jmp L(eok)
208
209 L(frac):mov d, %rbp
210 neg %rbp
211 jmp L(fent)
212
213 ALIGN(16) C K8-K10 P6-CNR P6-NHM P4
214 L(ftop):mul dinv C 0,12 0,17 0,17
215 add %r11, %rdx C 5 8 10
216 mov %rax, %r11 C 4 8 3
217 mov %rdx, %r13 C 6 9 11
218 imul %rbp, %rdx C 6 9 11
219 mov d, %rax C
220 add %rdx, %rax C 10 14 14
221 cmp %r11, %rdx C 10 14 14
222 cmovc %rdx, %rax C 11 15 15
223 adc $-1, %r13 C
224 mov %r13, (qp) C
225 sub $8, qp C
226 L(fent):lea 1(%rax), %r11 C
227 dec fn C
228 jns L(ftop) C
229
230 shr R8(%rcx), %rax
231 L(ret): pop %rbx
232 pop %rbp
233 pop %r12
234 pop %r13
235 FUNC_EXIT()
236 ret
237 EPILOGUE()