Mercurial > hg > sv-dependency-builds
comparison src/fftw-3.3.5/rdft/simd/common/hc2cbdftv_2.c @ 42:2cd0e3b3e1fd
Current fftw source
author | Chris Cannam |
---|---|
date | Tue, 18 Oct 2016 13:40:26 +0100 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
41:481f5f8c5634 | 42:2cd0e3b3e1fd |
---|---|
1 /* | |
2 * Copyright (c) 2003, 2007-14 Matteo Frigo | |
3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology | |
4 * | |
5 * This program is free software; you can redistribute it and/or modify | |
6 * it under the terms of the GNU General Public License as published by | |
7 * the Free Software Foundation; either version 2 of the License, or | |
8 * (at your option) any later version. | |
9 * | |
10 * This program 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 | |
13 * GNU General Public License for more details. | |
14 * | |
15 * You should have received a copy of the GNU General Public License | |
16 * along with this program; if not, write to the Free Software | |
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
18 * | |
19 */ | |
20 | |
21 /* This file was automatically generated --- DO NOT EDIT */ | |
22 /* Generated on Sat Jul 30 16:52:42 EDT 2016 */ | |
23 | |
24 #include "codelet-rdft.h" | |
25 | |
26 #ifdef HAVE_FMA | |
27 | |
28 /* Generated by: ../../../genfft/gen_hc2cdft_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 2 -dif -sign 1 -name hc2cbdftv_2 -include hc2cbv.h */ | |
29 | |
30 /* | |
31 * This function contains 5 FP additions, 4 FP multiplications, | |
32 * (or, 3 additions, 2 multiplications, 2 fused multiply/add), | |
33 * 8 stack variables, 0 constants, and 4 memory accesses | |
34 */ | |
35 #include "hc2cbv.h" | |
36 | |
37 static void hc2cbdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) | |
38 { | |
39 { | |
40 INT m; | |
41 for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 2)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(8, rs)) { | |
42 V T2, T3, T1, T5, T4, T7, T6; | |
43 T2 = LD(&(Rp[0]), ms, &(Rp[0])); | |
44 T3 = LD(&(Rm[0]), -ms, &(Rm[0])); | |
45 T1 = LDW(&(W[0])); | |
46 T5 = VFMACONJ(T3, T2); | |
47 T4 = VZMULI(T1, VFNMSCONJ(T3, T2)); | |
48 T7 = VCONJ(VSUB(T5, T4)); | |
49 T6 = VADD(T4, T5); | |
50 ST(&(Rm[0]), T7, -ms, &(Rm[0])); | |
51 ST(&(Rp[0]), T6, ms, &(Rp[0])); | |
52 } | |
53 } | |
54 VLEAVE(); | |
55 } | |
56 | |
57 static const tw_instr twinstr[] = { | |
58 VTW(1, 1), | |
59 {TW_NEXT, VL, 0} | |
60 }; | |
61 | |
62 static const hc2c_desc desc = { 2, XSIMD_STRING("hc2cbdftv_2"), twinstr, &GENUS, {3, 2, 2, 0} }; | |
63 | |
64 void XSIMD(codelet_hc2cbdftv_2) (planner *p) { | |
65 X(khc2c_register) (p, hc2cbdftv_2, &desc, HC2C_VIA_DFT); | |
66 } | |
67 #else /* HAVE_FMA */ | |
68 | |
69 /* Generated by: ../../../genfft/gen_hc2cdft_c.native -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 2 -dif -sign 1 -name hc2cbdftv_2 -include hc2cbv.h */ | |
70 | |
71 /* | |
72 * This function contains 5 FP additions, 2 FP multiplications, | |
73 * (or, 5 additions, 2 multiplications, 0 fused multiply/add), | |
74 * 9 stack variables, 0 constants, and 4 memory accesses | |
75 */ | |
76 #include "hc2cbv.h" | |
77 | |
78 static void hc2cbdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) | |
79 { | |
80 { | |
81 INT m; | |
82 for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 2)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(8, rs)) { | |
83 V T6, T5, T2, T4, T3, T1, T7, T8; | |
84 T2 = LD(&(Rp[0]), ms, &(Rp[0])); | |
85 T3 = LD(&(Rm[0]), -ms, &(Rm[0])); | |
86 T4 = VCONJ(T3); | |
87 T6 = VADD(T2, T4); | |
88 T1 = LDW(&(W[0])); | |
89 T5 = VZMULI(T1, VSUB(T2, T4)); | |
90 T7 = VADD(T5, T6); | |
91 ST(&(Rp[0]), T7, ms, &(Rp[0])); | |
92 T8 = VCONJ(VSUB(T6, T5)); | |
93 ST(&(Rm[0]), T8, -ms, &(Rm[0])); | |
94 } | |
95 } | |
96 VLEAVE(); | |
97 } | |
98 | |
99 static const tw_instr twinstr[] = { | |
100 VTW(1, 1), | |
101 {TW_NEXT, VL, 0} | |
102 }; | |
103 | |
104 static const hc2c_desc desc = { 2, XSIMD_STRING("hc2cbdftv_2"), twinstr, &GENUS, {5, 2, 0, 0} }; | |
105 | |
106 void XSIMD(codelet_hc2cbdftv_2) (planner *p) { | |
107 X(khc2c_register) (p, hc2cbdftv_2, &desc, HC2C_VIA_DFT); | |
108 } | |
109 #endif /* HAVE_FMA */ |