1 | /* Function sincos vectorized with AVX-512. Wrapper to AVX2 version. |
2 | Copyright (C) 2014-2016 Free Software Foundation, Inc. |
3 | This file is part of the GNU C Library. |
4 | |
5 | The GNU C Library is free software; you can redistribute it and/or |
6 | modify it under the terms of the GNU Lesser General Public |
7 | License as published by the Free Software Foundation; either |
8 | version 2.1 of the License, or (at your option) any later version. |
9 | |
10 | The GNU C Library is distributed in the hope that it will be useful, |
11 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
13 | Lesser General Public License for more details. |
14 | |
15 | You should have received a copy of the GNU Lesser General Public |
16 | License along with the GNU C Library; if not, see |
17 | <http://www.gnu.org/licenses/>. */ |
18 | |
19 | #include <sysdep.h> |
20 | #include "svml_d_wrapper_impl.h" |
21 | |
22 | .text |
23 | ENTRY (_ZGVeN8vl8l8_sincos) |
24 | WRAPPER_IMPL_AVX512_fFF _ZGVdN4vl8l8_sincos |
25 | END (_ZGVeN8vl8l8_sincos) |
26 | |
27 | /* AVX512 ISA version as wrapper to AVX2 ISA version (for vector |
28 | function declared with #pragma omp declare simd notinbranch). */ |
29 | .macro WRAPPER_IMPL_AVX512_fFF_vvv callee |
30 | #ifndef __ILP32__ |
31 | pushq %rbp |
32 | cfi_adjust_cfa_offset (8) |
33 | cfi_rel_offset (%rbp, 0) |
34 | movq %rsp, %rbp |
35 | cfi_def_cfa_register (%rbp) |
36 | andq $-64, %rsp |
37 | subq $320, %rsp |
38 | /* Encoding for vmovups %zmm0, 256(%rsp). */ |
39 | .byte 0x62 |
40 | .byte 0xf1 |
41 | .byte 0x7c |
42 | .byte 0x48 |
43 | .byte 0x11 |
44 | .byte 0x44 |
45 | .byte 0x24 |
46 | .byte 0x04 |
47 | lea (%rsp), %rdi |
48 | /* Encoding for vmovups %zmm1, 128(%rdi). */ |
49 | .byte 0x62 |
50 | .byte 0xf1 |
51 | .byte 0x7c |
52 | .byte 0x48 |
53 | .byte 0x11 |
54 | .byte 0x4f |
55 | .byte 0x02 |
56 | /* Encoding for vmovups %zmm2, 192(%rdi). */ |
57 | .byte 0x62 |
58 | .byte 0xf1 |
59 | .byte 0x7c |
60 | .byte 0x48 |
61 | .byte 0x11 |
62 | .byte 0x57 |
63 | .byte 0x03 |
64 | lea 64(%rsp), %rsi |
65 | call HIDDEN_JUMPTARGET(\callee) |
66 | vmovdqu 288(%rsp), %ymm0 |
67 | lea 32(%rsp), %rdi |
68 | lea 96(%rsp), %rsi |
69 | call HIDDEN_JUMPTARGET(\callee) |
70 | movq 128(%rsp), %rdx |
71 | movq 192(%rsp), %rsi |
72 | movq 136(%rsp), %r8 |
73 | movq 200(%rsp), %r10 |
74 | movq (%rsp), %rax |
75 | movq 64(%rsp), %rcx |
76 | movq 8(%rsp), %rdi |
77 | movq 72(%rsp), %r9 |
78 | movq %rax, (%rdx) |
79 | movq %rcx, (%rsi) |
80 | movq 144(%rsp), %rax |
81 | movq 208(%rsp), %rcx |
82 | movq %rdi, (%r8) |
83 | movq %r9, (%r10) |
84 | movq 152(%rsp), %rdi |
85 | movq 216(%rsp), %r9 |
86 | movq 16(%rsp), %r11 |
87 | movq 80(%rsp), %rdx |
88 | movq 24(%rsp), %rsi |
89 | movq 88(%rsp), %r8 |
90 | movq %r11, (%rax) |
91 | movq %rdx, (%rcx) |
92 | movq 160(%rsp), %r11 |
93 | movq 224(%rsp), %rdx |
94 | movq %rsi, (%rdi) |
95 | movq %r8, (%r9) |
96 | movq 168(%rsp), %rsi |
97 | movq 232(%rsp), %r8 |
98 | movq 32(%rsp), %r10 |
99 | movq 96(%rsp), %rax |
100 | movq 40(%rsp), %rcx |
101 | movq 104(%rsp), %rdi |
102 | movq %r10, (%r11) |
103 | movq %rax, (%rdx) |
104 | movq 176(%rsp), %r10 |
105 | movq 240(%rsp), %rax |
106 | movq %rcx, (%rsi) |
107 | movq %rdi, (%r8) |
108 | movq 184(%rsp), %rcx |
109 | movq 248(%rsp), %rdi |
110 | movq 48(%rsp), %r9 |
111 | movq 112(%rsp), %r11 |
112 | movq 56(%rsp), %rdx |
113 | movq 120(%rsp), %rsi |
114 | movq %r9, (%r10) |
115 | movq %r11, (%rax) |
116 | movq %rdx, (%rcx) |
117 | movq %rsi, (%rdi) |
118 | movq %rbp, %rsp |
119 | cfi_def_cfa_register (%rsp) |
120 | popq %rbp |
121 | cfi_adjust_cfa_offset (-8) |
122 | cfi_restore (%rbp) |
123 | ret |
124 | #else |
125 | leal 8(%rsp), %r10d |
126 | .cfi_def_cfa 10, 0 |
127 | andl $-64, %esp |
128 | pushq -8(%r10d) |
129 | pushq %rbp |
130 | .cfi_escape 0x10,0x6,0x2,0x76,0 |
131 | movl %esp, %ebp |
132 | pushq %r12 |
133 | leal -112(%rbp), %esi |
134 | pushq %r10 |
135 | .cfi_escape 0xf,0x3,0x76,0x70,0x6 |
136 | .cfi_escape 0x10,0xc,0x2,0x76,0x78 |
137 | leal -176(%rbp), %edi |
138 | movq %rsi, %r12 |
139 | pushq %rbx |
140 | .cfi_escape 0x10,0x3,0x2,0x76,0x68 |
141 | movq %rdi, %rbx |
142 | subl $280, %esp |
143 | vmovdqa %ymm1, -208(%ebp) |
144 | vmovdqa %ymm2, -240(%ebp) |
145 | /* Encoding for vmovapd %zmm0, -304(%ebp). */ |
146 | .byte 0x67 |
147 | .byte 0x62 |
148 | .byte 0xf1 |
149 | .byte 0xfd |
150 | .byte 0x48 |
151 | .byte 0x29 |
152 | .byte 0x85 |
153 | .byte 0xd0 |
154 | .byte 0xfe |
155 | .byte 0xff |
156 | .byte 0xff |
157 | call HIDDEN_JUMPTARGET(\callee) |
158 | leal 32(%r12), %esi |
159 | vmovupd -272(%ebp), %ymm0 |
160 | leal 32(%rbx), %edi |
161 | call HIDDEN_JUMPTARGET(\callee) |
162 | movl -208(%ebp), %eax |
163 | vmovsd -176(%ebp), %xmm0 |
164 | vmovsd %xmm0, (%eax) |
165 | movl -204(%ebp), %eax |
166 | vmovsd -168(%ebp), %xmm0 |
167 | vmovsd %xmm0, (%eax) |
168 | movl -200(%ebp), %eax |
169 | vmovsd -160(%ebp), %xmm0 |
170 | vmovsd %xmm0, (%eax) |
171 | movl -196(%ebp), %eax |
172 | vmovsd -152(%ebp), %xmm0 |
173 | vmovsd %xmm0, (%eax) |
174 | movl -192(%ebp), %eax |
175 | vmovsd -144(%ebp), %xmm0 |
176 | vmovsd %xmm0, (%eax) |
177 | movl -188(%ebp), %eax |
178 | vmovsd -136(%ebp), %xmm0 |
179 | vmovsd %xmm0, (%eax) |
180 | movl -184(%ebp), %eax |
181 | vmovsd -128(%ebp), %xmm0 |
182 | vmovsd %xmm0, (%eax) |
183 | movl -180(%ebp), %eax |
184 | vmovsd -120(%ebp), %xmm0 |
185 | vmovsd %xmm0, (%eax) |
186 | movl -240(%ebp), %eax |
187 | vmovsd -112(%ebp), %xmm0 |
188 | vmovsd %xmm0, (%eax) |
189 | movl -236(%ebp), %eax |
190 | vmovsd -104(%ebp), %xmm0 |
191 | vmovsd %xmm0, (%eax) |
192 | movl -232(%ebp), %eax |
193 | vmovsd -96(%ebp), %xmm0 |
194 | vmovsd %xmm0, (%eax) |
195 | movl -228(%ebp), %eax |
196 | vmovsd -88(%ebp), %xmm0 |
197 | vmovsd %xmm0, (%eax) |
198 | movl -224(%ebp), %eax |
199 | vmovsd -80(%ebp), %xmm0 |
200 | vmovsd %xmm0, (%eax) |
201 | movl -220(%ebp), %eax |
202 | vmovsd -72(%ebp), %xmm0 |
203 | vmovsd %xmm0, (%eax) |
204 | movl -216(%ebp), %eax |
205 | vmovsd -64(%ebp), %xmm0 |
206 | vmovsd %xmm0, (%eax) |
207 | movl -212(%ebp), %eax |
208 | vmovsd -56(%ebp), %xmm0 |
209 | vmovsd %xmm0, (%eax) |
210 | addl $280, %esp |
211 | popq %rbx |
212 | popq %r10 |
213 | .cfi_def_cfa 10, 0 |
214 | popq %r12 |
215 | popq %rbp |
216 | leal -8(%r10), %esp |
217 | .cfi_def_cfa 7, 8 |
218 | ret |
219 | #endif |
220 | .endm |
221 | |
222 | ENTRY (_ZGVeN8vvv_sincos) |
223 | WRAPPER_IMPL_AVX512_fFF_vvv _ZGVdN4vl8l8_sincos |
224 | END (_ZGVeN8vvv_sincos) |
225 | |