Chris@10
|
1 /*
|
Chris@10
|
2 * Copyright (c) 2003, 2007-11 Matteo Frigo
|
Chris@10
|
3 * Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology
|
Chris@10
|
4 *
|
Chris@10
|
5 * This program is free software; you can redistribute it and/or modify
|
Chris@10
|
6 * it under the terms of the GNU General Public License as published by
|
Chris@10
|
7 * the Free Software Foundation; either version 2 of the License, or
|
Chris@10
|
8 * (at your option) any later version.
|
Chris@10
|
9 *
|
Chris@10
|
10 * This program is distributed in the hope that it will be useful,
|
Chris@10
|
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
Chris@10
|
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
Chris@10
|
13 * GNU General Public License for more details.
|
Chris@10
|
14 *
|
Chris@10
|
15 * You should have received a copy of the GNU General Public License
|
Chris@10
|
16 * along with this program; if not, write to the Free Software
|
Chris@10
|
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
Chris@10
|
18 *
|
Chris@10
|
19 */
|
Chris@10
|
20
|
Chris@10
|
21 /* This file was automatically generated --- DO NOT EDIT */
|
Chris@10
|
22 /* Generated on Sun Nov 25 07:38:35 EST 2012 */
|
Chris@10
|
23
|
Chris@10
|
24 #include "codelet-dft.h"
|
Chris@10
|
25
|
Chris@10
|
26 #ifdef HAVE_FMA
|
Chris@10
|
27
|
Chris@10
|
28 /* Generated by: ../../../genfft/gen_twiddle_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 2 -name t2fv_2 -include t2f.h */
|
Chris@10
|
29
|
Chris@10
|
30 /*
|
Chris@10
|
31 * This function contains 3 FP additions, 2 FP multiplications,
|
Chris@10
|
32 * (or, 3 additions, 2 multiplications, 0 fused multiply/add),
|
Chris@10
|
33 * 5 stack variables, 0 constants, and 4 memory accesses
|
Chris@10
|
34 */
|
Chris@10
|
35 #include "t2f.h"
|
Chris@10
|
36
|
Chris@10
|
37 static void t2fv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
|
Chris@10
|
38 {
|
Chris@10
|
39 {
|
Chris@10
|
40 INT m;
|
Chris@10
|
41 R *x;
|
Chris@10
|
42 x = ri;
|
Chris@10
|
43 for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(2, rs)) {
|
Chris@10
|
44 V T1, T2, T3;
|
Chris@10
|
45 T1 = LD(&(x[0]), ms, &(x[0]));
|
Chris@10
|
46 T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
|
Chris@10
|
47 T3 = BYTWJ(&(W[0]), T2);
|
Chris@10
|
48 ST(&(x[0]), VADD(T1, T3), ms, &(x[0]));
|
Chris@10
|
49 ST(&(x[WS(rs, 1)]), VSUB(T1, T3), ms, &(x[WS(rs, 1)]));
|
Chris@10
|
50 }
|
Chris@10
|
51 }
|
Chris@10
|
52 VLEAVE();
|
Chris@10
|
53 }
|
Chris@10
|
54
|
Chris@10
|
55 static const tw_instr twinstr[] = {
|
Chris@10
|
56 VTW(0, 1),
|
Chris@10
|
57 {TW_NEXT, VL, 0}
|
Chris@10
|
58 };
|
Chris@10
|
59
|
Chris@10
|
60 static const ct_desc desc = { 2, XSIMD_STRING("t2fv_2"), twinstr, &GENUS, {3, 2, 0, 0}, 0, 0, 0 };
|
Chris@10
|
61
|
Chris@10
|
62 void XSIMD(codelet_t2fv_2) (planner *p) {
|
Chris@10
|
63 X(kdft_dit_register) (p, t2fv_2, &desc);
|
Chris@10
|
64 }
|
Chris@10
|
65 #else /* HAVE_FMA */
|
Chris@10
|
66
|
Chris@10
|
67 /* Generated by: ../../../genfft/gen_twiddle_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 2 -name t2fv_2 -include t2f.h */
|
Chris@10
|
68
|
Chris@10
|
69 /*
|
Chris@10
|
70 * This function contains 3 FP additions, 2 FP multiplications,
|
Chris@10
|
71 * (or, 3 additions, 2 multiplications, 0 fused multiply/add),
|
Chris@10
|
72 * 5 stack variables, 0 constants, and 4 memory accesses
|
Chris@10
|
73 */
|
Chris@10
|
74 #include "t2f.h"
|
Chris@10
|
75
|
Chris@10
|
76 static void t2fv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
|
Chris@10
|
77 {
|
Chris@10
|
78 {
|
Chris@10
|
79 INT m;
|
Chris@10
|
80 R *x;
|
Chris@10
|
81 x = ri;
|
Chris@10
|
82 for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(2, rs)) {
|
Chris@10
|
83 V T1, T3, T2;
|
Chris@10
|
84 T1 = LD(&(x[0]), ms, &(x[0]));
|
Chris@10
|
85 T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
|
Chris@10
|
86 T3 = BYTWJ(&(W[0]), T2);
|
Chris@10
|
87 ST(&(x[WS(rs, 1)]), VSUB(T1, T3), ms, &(x[WS(rs, 1)]));
|
Chris@10
|
88 ST(&(x[0]), VADD(T1, T3), ms, &(x[0]));
|
Chris@10
|
89 }
|
Chris@10
|
90 }
|
Chris@10
|
91 VLEAVE();
|
Chris@10
|
92 }
|
Chris@10
|
93
|
Chris@10
|
94 static const tw_instr twinstr[] = {
|
Chris@10
|
95 VTW(0, 1),
|
Chris@10
|
96 {TW_NEXT, VL, 0}
|
Chris@10
|
97 };
|
Chris@10
|
98
|
Chris@10
|
99 static const ct_desc desc = { 2, XSIMD_STRING("t2fv_2"), twinstr, &GENUS, {3, 2, 0, 0}, 0, 0, 0 };
|
Chris@10
|
100
|
Chris@10
|
101 void XSIMD(codelet_t2fv_2) (planner *p) {
|
Chris@10
|
102 X(kdft_dit_register) (p, t2fv_2, &desc);
|
Chris@10
|
103 }
|
Chris@10
|
104 #endif /* HAVE_FMA */
|