1 /*
2 * FFT/IFFT transforms
3 * AltiVec-enabled
4 * Copyright (c) 2009 Loren Merritt
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
26
27 /**
28 * Do a complex FFT with the parameters defined in ff_fft_init(). The
29 * input data must be permuted before with s->revtab table. No
30 * 1.0/sqrt(n) normalization is done.
31 * AltiVec-enabled
32 * This code assumes that the 'z' pointer is 16 bytes-aligned
33 * It also assumes all FFTComplex are 8 bytes-aligned pair of float
34 */
35
38
39 #if HAVE_GNU_AS
41 {
42 int j, k;
44 int n4 = n >> 2;
45 int n8 = n >> 3;
46 int n32 = n >> 5;
47 const uint16_t *revtabj = s->
revtab;
48 const uint16_t *revtabk = s->
revtab+n4;
53
54 /* pre rotation */
55 k = n32-1;
56 do {
57 vec_f cos,sin,cos0,sin0,cos1,sin1,
re,
im,r0,i0,r1,i1,
a,
b,
c,d;
58 #define CMULA(p,o0,o1,o2,o3)\
59 a = pin[ k*2+p]; /* { z[k].re, z[k].im, z[k+1].re, z[k+1].im } */\
60 b = pin[-k*2-p-1]; /* { z[-k-2].re, z[-k-2].im, z[-k-1].re, z[-k-1].im } */\
61 re = vec_perm(a, b, vcprm(0,2,s0,s2)); /* { z[k].re, z[k+1].re, z[-k-2].re, z[-k-1].re } */\
62 im = vec_perm(a, b, vcprm(s3,s1,3,1)); /* { z[-k-1].im, z[-k-2].im, z[k+1].im, z[k].im } */\
63 cos = vec_perm(cos0, cos1, vcprm(o0,o1,s##o2,s##o3)); /* { cos[k], cos[k+1], cos[-k-2], cos[-k-1] } */\
64 sin = vec_perm(sin0, sin1, vcprm(o0,o1,s##o2,s##o3));\
65 r##p = im*cos - re*sin;\
66 i##p = re*cos + im*sin;
67 #define STORE2(v,dst)\
68 j = dst;\
69 vec_ste(v, 0, output+j*2);\
70 vec_ste(v, 4, output+j*2);
71 #define STORE8(p)\
72 a = vec_perm(r##p, i##p, vcprm(0,s0,0,s0));\
73 b = vec_perm(r##p, i##p, vcprm(1,s1,1,s1));\
74 c = vec_perm(r##p, i##p, vcprm(2,s2,2,s2));\
75 d = vec_perm(r##p, i##p, vcprm(3,s3,3,s3));\
76 STORE2(a, revtabk[ p*2-4]);\
77 STORE2(b, revtabk[ p*2-3]);\
78 STORE2(c, revtabj[-p*2+2]);\
79 STORE2(d, revtabj[-p*2+3]);
80
81 cos0 = tcos[k];
82 sin0 = tsin[k];
83 cos1 = tcos[-k-1];
84 sin1 = tsin[-k-1];
85 CMULA(0, 0,1,2,3);
86 CMULA(1, 2,3,0,1);
87 STORE8(0);
88 STORE8(1);
89 revtabj += 4;
90 revtabk -= 4;
91 k--;
92 } while(k >= 0);
93
95
96 /* post rotation + reordering */
97 j = -n32;
98 k = n32-1;
99 do {
101 #define CMULB(d0,d1,o)\
102 re = pout[o*2];\
103 im = pout[o*2+1];\
104 cos = tcos[o];\
105 sin = tsin[o];\
106 d0 = im*sin - re*cos;\
107 d1 = re*sin + im*cos;
108
109 CMULB(a,b,j);
110 CMULB(c,d,k);
111 pout[2*j] = vec_perm(a, d,
vcprm(0,
s3,1,
s2));
112 pout[2*j+1] = vec_perm(a, d,
vcprm(2,
s1,3,
s0));
113 pout[2*k] = vec_perm(c, b,
vcprm(0,
s3,1,
s2));
114 pout[2*k+1] = vec_perm(c, b,
vcprm(2,
s1,3,
s0));
115 j++;
116 k--;
117 } while(k >= 0);
118 }
119
121 {
122 int k;
124 int n4 = n >> 2;
125 int n16 = n >> 4;
126 vec_u32 sign = {1
U<<31,1
U<<31,1U<<31,1U<<31};
129
130 ff_imdct_half_altivec(s, output+n4, input);
131
132 for (k = 0; k < n16; k++) {
135 p0[-k-1] = vec_perm(a, a,
vcprm(3,2,1,0));
136 p1[k] = vec_perm(b, b,
vcprm(3,2,1,0));
137 }
138 }
139 #endif /* HAVE_GNU_AS */
140
142 {
143 #if HAVE_GNU_AS
148 }
149 #endif
150 }