1 | /* Function hypotf vectorized with AVX-512. |
2 | Copyright (C) 2021-2024 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 | https://www.gnu.org/licenses/. */ |
18 | |
19 | /* |
20 | * ALGORITHM DESCRIPTION: |
21 | * |
22 | * HIGH LEVEL OVERVIEW |
23 | * |
24 | * Calculate z = (x*x+y*y) |
25 | * Calculate reciplicle sqrt (z) |
26 | * Calculate make two NR iterations |
27 | * |
28 | * ALGORITHM DETAILS |
29 | * |
30 | * Multiprecision branch for _HA_ only |
31 | * Remove sigm from both arguments |
32 | * Find maximum (_x) and minimum (_y) (by abs value) between arguments |
33 | * Split _x int _a and _b for multiprecision |
34 | * If _x >> _y we will we will not split _y for multiprecision |
35 | * all _y will be put into lower part (_d) and higher part (_c = 0) |
36 | * Fixing _hilo_mask for the case _x >> _y |
37 | * Split _y into _c and _d for multiprecision with fixed mask |
38 | * |
39 | * compute Hi and Lo parts of _z = _x*_x + _y*_y |
40 | * |
41 | * _zHi = _a*_a + _c*_c |
42 | * _zLo = (_x + _a)*_b + _d*_y + _d*_c |
43 | * _z = _zHi + _zLo |
44 | * |
45 | * No multiprecision branch for _LA_ and _EP_ |
46 | * _z = _VARG1 * _VARG1 + _VARG2 * _VARG2 |
47 | * |
48 | * Check _z exponent to be within borders [1E3 ; 60A] else goto Callout |
49 | * |
50 | * Compute resciplicle sqrt s0 ~ 1.0/sqrt(_z), |
51 | * that multiplied by _z, is final result for _EP_ version. |
52 | * |
53 | * First iteration (or zero iteration): |
54 | * s = z * s0 |
55 | * h = .5 * s0 |
56 | * d = s * h - .5 |
57 | * |
58 | * Second iteration: |
59 | * h = d * h + h |
60 | * s = s * d + s |
61 | * d = s * s - z (in multiprecision for _HA_) |
62 | * |
63 | * result = s - h * d |
64 | * |
65 | * EP version of the function can be implemented as y[i]=sqrt(a[i]^2+b[i]^2) |
66 | * with all intermediate operations done in target precision for i=1, .., n. |
67 | * It can return result y[i]=0 in case a[i]^2 and b[i]^2 underflow in target |
68 | * precision (for some i). It can return result y[i]=NAN in case |
69 | * a[i]^2+b[i]^2 overflow in target precision, for some i. It can return |
70 | * result y[i]=NAN in case a[i] or b[i] is infinite, for some i. |
71 | * |
72 | * |
73 | */ |
74 | |
75 | /* Offsets for data table __svml_shypot_data_internal |
76 | */ |
77 | #define _sAbsMask 0 |
78 | #define _sHalf 64 |
79 | #define _iExpBound 128 |
80 | |
81 | #include <sysdep.h> |
82 | |
83 | .section .text.evex512, "ax" , @progbits |
84 | ENTRY(_ZGVeN16vv_hypotf_skx) |
85 | pushq %rbp |
86 | cfi_def_cfa_offset(16) |
87 | movq %rsp, %rbp |
88 | cfi_def_cfa(6, 16) |
89 | cfi_offset(6, -16) |
90 | andq $-64, %rsp |
91 | subq $256, %rsp |
92 | vgetexpps {sae}, %zmm0, %zmm2 |
93 | vgetexpps {sae}, %zmm1, %zmm3 |
94 | vmovups _sHalf+__svml_shypot_data_internal(%rip), %zmm6 |
95 | vmaxps {sae}, %zmm3, %zmm2, %zmm4 |
96 | vmulps {rn-sae}, %zmm0, %zmm0, %zmm2 |
97 | vandps _sAbsMask+__svml_shypot_data_internal(%rip), %zmm4, %zmm5 |
98 | vfmadd231ps {rn-sae}, %zmm1, %zmm1, %zmm2 |
99 | vpcmpd $5, _iExpBound+__svml_shypot_data_internal(%rip), %zmm5, %k0 |
100 | vrsqrt14ps %zmm2, %zmm7 |
101 | kmovw %k0, %edx |
102 | vmulps {rn-sae}, %zmm7, %zmm2, %zmm9 |
103 | vmulps {rn-sae}, %zmm7, %zmm6, %zmm8 |
104 | vfnmadd231ps {rn-sae}, %zmm9, %zmm9, %zmm2 |
105 | vfmadd213ps {rn-sae}, %zmm9, %zmm8, %zmm2 |
106 | |
107 | /* |
108 | * VSCALEF( S, _VRES1, _VRES1, sExp ); |
109 | * The end of implementation |
110 | */ |
111 | testl %edx, %edx |
112 | |
113 | /* Go to special inputs processing branch */ |
114 | jne L(SPECIAL_VALUES_BRANCH) |
115 | # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 zmm2 |
116 | |
117 | /* Restore registers |
118 | * and exit the function |
119 | */ |
120 | |
121 | L(EXIT): |
122 | vmovaps %zmm2, %zmm0 |
123 | movq %rbp, %rsp |
124 | popq %rbp |
125 | cfi_def_cfa(7, 8) |
126 | cfi_restore(6) |
127 | ret |
128 | cfi_def_cfa(6, 16) |
129 | cfi_offset(6, -16) |
130 | |
131 | /* Branch to process |
132 | * special inputs |
133 | */ |
134 | |
135 | L(SPECIAL_VALUES_BRANCH): |
136 | vmovups %zmm0, 64(%rsp) |
137 | vmovups %zmm1, 128(%rsp) |
138 | vmovups %zmm2, 192(%rsp) |
139 | # LOE rbx r12 r13 r14 r15 edx zmm2 |
140 | |
141 | xorl %eax, %eax |
142 | # LOE rbx r12 r13 r14 r15 eax edx |
143 | |
144 | vzeroupper |
145 | movq %r12, 16(%rsp) |
146 | /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ |
147 | .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 |
148 | movl %eax, %r12d |
149 | movq %r13, 8(%rsp) |
150 | /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ |
151 | .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 |
152 | movl %edx, %r13d |
153 | movq %r14, (%rsp) |
154 | /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ |
155 | .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 |
156 | # LOE rbx r15 r12d r13d |
157 | |
158 | /* Range mask |
159 | * bits check |
160 | */ |
161 | |
162 | L(RANGEMASK_CHECK): |
163 | btl %r12d, %r13d |
164 | |
165 | /* Call scalar math function */ |
166 | jc L(SCALAR_MATH_CALL) |
167 | # LOE rbx r15 r12d r13d |
168 | |
169 | /* Special inputs |
170 | * processing loop |
171 | */ |
172 | |
173 | L(SPECIAL_VALUES_LOOP): |
174 | incl %r12d |
175 | cmpl $16, %r12d |
176 | |
177 | /* Check bits in range mask */ |
178 | jl L(RANGEMASK_CHECK) |
179 | # LOE rbx r15 r12d r13d |
180 | |
181 | movq 16(%rsp), %r12 |
182 | cfi_restore(12) |
183 | movq 8(%rsp), %r13 |
184 | cfi_restore(13) |
185 | movq (%rsp), %r14 |
186 | cfi_restore(14) |
187 | vmovups 192(%rsp), %zmm2 |
188 | |
189 | /* Go to exit */ |
190 | jmp L(EXIT) |
191 | /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ |
192 | .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 |
193 | /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ |
194 | .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 |
195 | /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ |
196 | .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 |
197 | # LOE rbx r12 r13 r14 r15 zmm2 |
198 | |
199 | /* Scalar math function call |
200 | * to process special input |
201 | */ |
202 | |
203 | L(SCALAR_MATH_CALL): |
204 | movl %r12d, %r14d |
205 | vmovss 64(%rsp, %r14, 4), %xmm0 |
206 | vmovss 128(%rsp, %r14, 4), %xmm1 |
207 | call hypotf@PLT |
208 | # LOE rbx r14 r15 r12d r13d xmm0 |
209 | |
210 | vmovss %xmm0, 192(%rsp, %r14, 4) |
211 | |
212 | /* Process special inputs in loop */ |
213 | jmp L(SPECIAL_VALUES_LOOP) |
214 | # LOE rbx r15 r12d r13d |
215 | END(_ZGVeN16vv_hypotf_skx) |
216 | |
217 | .section .rodata, "a" |
218 | .align 64 |
219 | |
220 | #ifdef __svml_shypot_data_internal_typedef |
221 | typedef unsigned int VUINT32; |
222 | typedef struct { |
223 | __declspec(align(64)) VUINT32 _sAbsMask[16][1]; |
224 | __declspec(align(64)) VUINT32 _sHalf[16][1]; |
225 | __declspec(align(64)) VUINT32 _iExpBound[16][1]; |
226 | } __svml_shypot_data_internal; |
227 | #endif |
228 | __svml_shypot_data_internal: |
229 | .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* _sAbsMask */ |
230 | .align 64 |
231 | .long 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000 /* _sHalf */ |
232 | /* fma based algorithm*/ |
233 | .align 64 |
234 | .long 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000, 0x427C0000 /* _iExpBound */ |
235 | .align 64 |
236 | .type __svml_shypot_data_internal, @object |
237 | .size __svml_shypot_data_internal, .-__svml_shypot_data_internal |
238 | |