1 | // This file is part of Eigen, a lightweight C++ template library
|
---|
2 | // for linear algebra.
|
---|
3 | //
|
---|
4 | // Copyright (C) 2010 Gael Guennebaud <gael.guennebaud@inria.fr>
|
---|
5 | //
|
---|
6 | // This Source Code Form is subject to the terms of the Mozilla
|
---|
7 | // Public License v. 2.0. If a copy of the MPL was not distributed
|
---|
8 | // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
---|
9 |
|
---|
10 | #ifndef EIGEN_COMPLEX_NEON_H
|
---|
11 | #define EIGEN_COMPLEX_NEON_H
|
---|
12 |
|
---|
13 | namespace Eigen {
|
---|
14 |
|
---|
15 | namespace internal {
|
---|
16 |
|
---|
17 | static uint32x4_t p4ui_CONJ_XOR = EIGEN_INIT_NEON_PACKET4(0x00000000, 0x80000000, 0x00000000, 0x80000000);
|
---|
18 | static uint32x2_t p2ui_CONJ_XOR = EIGEN_INIT_NEON_PACKET2(0x00000000, 0x80000000);
|
---|
19 |
|
---|
20 | //---------- float ----------
|
---|
21 | struct Packet2cf
|
---|
22 | {
|
---|
23 | EIGEN_STRONG_INLINE Packet2cf() {}
|
---|
24 | EIGEN_STRONG_INLINE explicit Packet2cf(const Packet4f& a) : v(a) {}
|
---|
25 | Packet4f v;
|
---|
26 | };
|
---|
27 |
|
---|
28 | template<> struct packet_traits<std::complex<float> > : default_packet_traits
|
---|
29 | {
|
---|
30 | typedef Packet2cf type;
|
---|
31 | enum {
|
---|
32 | Vectorizable = 1,
|
---|
33 | AlignedOnScalar = 1,
|
---|
34 | size = 2,
|
---|
35 |
|
---|
36 | HasAdd = 1,
|
---|
37 | HasSub = 1,
|
---|
38 | HasMul = 1,
|
---|
39 | HasDiv = 1,
|
---|
40 | HasNegate = 1,
|
---|
41 | HasAbs = 0,
|
---|
42 | HasAbs2 = 0,
|
---|
43 | HasMin = 0,
|
---|
44 | HasMax = 0,
|
---|
45 | HasSetLinear = 0
|
---|
46 | };
|
---|
47 | };
|
---|
48 |
|
---|
49 | template<> struct unpacket_traits<Packet2cf> { typedef std::complex<float> type; enum {size=2}; };
|
---|
50 |
|
---|
51 | template<> EIGEN_STRONG_INLINE Packet2cf pset1<Packet2cf>(const std::complex<float>& from)
|
---|
52 | {
|
---|
53 | float32x2_t r64;
|
---|
54 | r64 = vld1_f32((float *)&from);
|
---|
55 |
|
---|
56 | return Packet2cf(vcombine_f32(r64, r64));
|
---|
57 | }
|
---|
58 |
|
---|
59 | template<> EIGEN_STRONG_INLINE Packet2cf padd<Packet2cf>(const Packet2cf& a, const Packet2cf& b) { return Packet2cf(padd<Packet4f>(a.v,b.v)); }
|
---|
60 | template<> EIGEN_STRONG_INLINE Packet2cf psub<Packet2cf>(const Packet2cf& a, const Packet2cf& b) { return Packet2cf(psub<Packet4f>(a.v,b.v)); }
|
---|
61 | template<> EIGEN_STRONG_INLINE Packet2cf pnegate(const Packet2cf& a) { return Packet2cf(pnegate<Packet4f>(a.v)); }
|
---|
62 | template<> EIGEN_STRONG_INLINE Packet2cf pconj(const Packet2cf& a)
|
---|
63 | {
|
---|
64 | Packet4ui b = vreinterpretq_u32_f32(a.v);
|
---|
65 | return Packet2cf(vreinterpretq_f32_u32(veorq_u32(b, p4ui_CONJ_XOR)));
|
---|
66 | }
|
---|
67 |
|
---|
68 | template<> EIGEN_STRONG_INLINE Packet2cf pmul<Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
69 | {
|
---|
70 | Packet4f v1, v2;
|
---|
71 |
|
---|
72 | // Get the real values of a | a1_re | a1_re | a2_re | a2_re |
|
---|
73 | v1 = vcombine_f32(vdup_lane_f32(vget_low_f32(a.v), 0), vdup_lane_f32(vget_high_f32(a.v), 0));
|
---|
74 | // Get the real values of a | a1_im | a1_im | a2_im | a2_im |
|
---|
75 | v2 = vcombine_f32(vdup_lane_f32(vget_low_f32(a.v), 1), vdup_lane_f32(vget_high_f32(a.v), 1));
|
---|
76 | // Multiply the real a with b
|
---|
77 | v1 = vmulq_f32(v1, b.v);
|
---|
78 | // Multiply the imag a with b
|
---|
79 | v2 = vmulq_f32(v2, b.v);
|
---|
80 | // Conjugate v2
|
---|
81 | v2 = vreinterpretq_f32_u32(veorq_u32(vreinterpretq_u32_f32(v2), p4ui_CONJ_XOR));
|
---|
82 | // Swap real/imag elements in v2.
|
---|
83 | v2 = vrev64q_f32(v2);
|
---|
84 | // Add and return the result
|
---|
85 | return Packet2cf(vaddq_f32(v1, v2));
|
---|
86 | }
|
---|
87 |
|
---|
88 | template<> EIGEN_STRONG_INLINE Packet2cf pand <Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
89 | {
|
---|
90 | return Packet2cf(vreinterpretq_f32_u32(vorrq_u32(vreinterpretq_u32_f32(a.v),vreinterpretq_u32_f32(b.v))));
|
---|
91 | }
|
---|
92 | template<> EIGEN_STRONG_INLINE Packet2cf por <Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
93 | {
|
---|
94 | return Packet2cf(vreinterpretq_f32_u32(vorrq_u32(vreinterpretq_u32_f32(a.v),vreinterpretq_u32_f32(b.v))));
|
---|
95 | }
|
---|
96 | template<> EIGEN_STRONG_INLINE Packet2cf pxor <Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
97 | {
|
---|
98 | return Packet2cf(vreinterpretq_f32_u32(veorq_u32(vreinterpretq_u32_f32(a.v),vreinterpretq_u32_f32(b.v))));
|
---|
99 | }
|
---|
100 | template<> EIGEN_STRONG_INLINE Packet2cf pandnot<Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
101 | {
|
---|
102 | return Packet2cf(vreinterpretq_f32_u32(vbicq_u32(vreinterpretq_u32_f32(a.v),vreinterpretq_u32_f32(b.v))));
|
---|
103 | }
|
---|
104 |
|
---|
105 | template<> EIGEN_STRONG_INLINE Packet2cf pload<Packet2cf>(const std::complex<float>* from) { EIGEN_DEBUG_ALIGNED_LOAD return Packet2cf(pload<Packet4f>((const float*)from)); }
|
---|
106 | template<> EIGEN_STRONG_INLINE Packet2cf ploadu<Packet2cf>(const std::complex<float>* from) { EIGEN_DEBUG_UNALIGNED_LOAD return Packet2cf(ploadu<Packet4f>((const float*)from)); }
|
---|
107 |
|
---|
108 | template<> EIGEN_STRONG_INLINE Packet2cf ploaddup<Packet2cf>(const std::complex<float>* from) { return pset1<Packet2cf>(*from); }
|
---|
109 |
|
---|
110 | template<> EIGEN_STRONG_INLINE void pstore <std::complex<float> >(std::complex<float> * to, const Packet2cf& from) { EIGEN_DEBUG_ALIGNED_STORE pstore((float*)to, from.v); }
|
---|
111 | template<> EIGEN_STRONG_INLINE void pstoreu<std::complex<float> >(std::complex<float> * to, const Packet2cf& from) { EIGEN_DEBUG_UNALIGNED_STORE pstoreu((float*)to, from.v); }
|
---|
112 |
|
---|
113 | template<> EIGEN_STRONG_INLINE void prefetch<std::complex<float> >(const std::complex<float> * addr) { EIGEN_ARM_PREFETCH((float *)addr); }
|
---|
114 |
|
---|
115 | template<> EIGEN_STRONG_INLINE std::complex<float> pfirst<Packet2cf>(const Packet2cf& a)
|
---|
116 | {
|
---|
117 | std::complex<float> EIGEN_ALIGN16 x[2];
|
---|
118 | vst1q_f32((float *)x, a.v);
|
---|
119 | return x[0];
|
---|
120 | }
|
---|
121 |
|
---|
122 | template<> EIGEN_STRONG_INLINE Packet2cf preverse(const Packet2cf& a)
|
---|
123 | {
|
---|
124 | float32x2_t a_lo, a_hi;
|
---|
125 | Packet4f a_r128;
|
---|
126 |
|
---|
127 | a_lo = vget_low_f32(a.v);
|
---|
128 | a_hi = vget_high_f32(a.v);
|
---|
129 | a_r128 = vcombine_f32(a_hi, a_lo);
|
---|
130 |
|
---|
131 | return Packet2cf(a_r128);
|
---|
132 | }
|
---|
133 |
|
---|
134 | template<> EIGEN_STRONG_INLINE Packet2cf pcplxflip<Packet2cf>(const Packet2cf& a)
|
---|
135 | {
|
---|
136 | return Packet2cf(vrev64q_f32(a.v));
|
---|
137 | }
|
---|
138 |
|
---|
139 | template<> EIGEN_STRONG_INLINE std::complex<float> predux<Packet2cf>(const Packet2cf& a)
|
---|
140 | {
|
---|
141 | float32x2_t a1, a2;
|
---|
142 | std::complex<float> s;
|
---|
143 |
|
---|
144 | a1 = vget_low_f32(a.v);
|
---|
145 | a2 = vget_high_f32(a.v);
|
---|
146 | a2 = vadd_f32(a1, a2);
|
---|
147 | vst1_f32((float *)&s, a2);
|
---|
148 |
|
---|
149 | return s;
|
---|
150 | }
|
---|
151 |
|
---|
152 | template<> EIGEN_STRONG_INLINE Packet2cf preduxp<Packet2cf>(const Packet2cf* vecs)
|
---|
153 | {
|
---|
154 | Packet4f sum1, sum2, sum;
|
---|
155 |
|
---|
156 | // Add the first two 64-bit float32x2_t of vecs[0]
|
---|
157 | sum1 = vcombine_f32(vget_low_f32(vecs[0].v), vget_low_f32(vecs[1].v));
|
---|
158 | sum2 = vcombine_f32(vget_high_f32(vecs[0].v), vget_high_f32(vecs[1].v));
|
---|
159 | sum = vaddq_f32(sum1, sum2);
|
---|
160 |
|
---|
161 | return Packet2cf(sum);
|
---|
162 | }
|
---|
163 |
|
---|
164 | template<> EIGEN_STRONG_INLINE std::complex<float> predux_mul<Packet2cf>(const Packet2cf& a)
|
---|
165 | {
|
---|
166 | float32x2_t a1, a2, v1, v2, prod;
|
---|
167 | std::complex<float> s;
|
---|
168 |
|
---|
169 | a1 = vget_low_f32(a.v);
|
---|
170 | a2 = vget_high_f32(a.v);
|
---|
171 | // Get the real values of a | a1_re | a1_re | a2_re | a2_re |
|
---|
172 | v1 = vdup_lane_f32(a1, 0);
|
---|
173 | // Get the real values of a | a1_im | a1_im | a2_im | a2_im |
|
---|
174 | v2 = vdup_lane_f32(a1, 1);
|
---|
175 | // Multiply the real a with b
|
---|
176 | v1 = vmul_f32(v1, a2);
|
---|
177 | // Multiply the imag a with b
|
---|
178 | v2 = vmul_f32(v2, a2);
|
---|
179 | // Conjugate v2
|
---|
180 | v2 = vreinterpret_f32_u32(veor_u32(vreinterpret_u32_f32(v2), p2ui_CONJ_XOR));
|
---|
181 | // Swap real/imag elements in v2.
|
---|
182 | v2 = vrev64_f32(v2);
|
---|
183 | // Add v1, v2
|
---|
184 | prod = vadd_f32(v1, v2);
|
---|
185 |
|
---|
186 | vst1_f32((float *)&s, prod);
|
---|
187 |
|
---|
188 | return s;
|
---|
189 | }
|
---|
190 |
|
---|
191 | template<int Offset>
|
---|
192 | struct palign_impl<Offset,Packet2cf>
|
---|
193 | {
|
---|
194 | EIGEN_STRONG_INLINE static void run(Packet2cf& first, const Packet2cf& second)
|
---|
195 | {
|
---|
196 | if (Offset==1)
|
---|
197 | {
|
---|
198 | first.v = vextq_f32(first.v, second.v, 2);
|
---|
199 | }
|
---|
200 | }
|
---|
201 | };
|
---|
202 |
|
---|
203 | template<> struct conj_helper<Packet2cf, Packet2cf, false,true>
|
---|
204 | {
|
---|
205 | EIGEN_STRONG_INLINE Packet2cf pmadd(const Packet2cf& x, const Packet2cf& y, const Packet2cf& c) const
|
---|
206 | { return padd(pmul(x,y),c); }
|
---|
207 |
|
---|
208 | EIGEN_STRONG_INLINE Packet2cf pmul(const Packet2cf& a, const Packet2cf& b) const
|
---|
209 | {
|
---|
210 | return internal::pmul(a, pconj(b));
|
---|
211 | }
|
---|
212 | };
|
---|
213 |
|
---|
214 | template<> struct conj_helper<Packet2cf, Packet2cf, true,false>
|
---|
215 | {
|
---|
216 | EIGEN_STRONG_INLINE Packet2cf pmadd(const Packet2cf& x, const Packet2cf& y, const Packet2cf& c) const
|
---|
217 | { return padd(pmul(x,y),c); }
|
---|
218 |
|
---|
219 | EIGEN_STRONG_INLINE Packet2cf pmul(const Packet2cf& a, const Packet2cf& b) const
|
---|
220 | {
|
---|
221 | return internal::pmul(pconj(a), b);
|
---|
222 | }
|
---|
223 | };
|
---|
224 |
|
---|
225 | template<> struct conj_helper<Packet2cf, Packet2cf, true,true>
|
---|
226 | {
|
---|
227 | EIGEN_STRONG_INLINE Packet2cf pmadd(const Packet2cf& x, const Packet2cf& y, const Packet2cf& c) const
|
---|
228 | { return padd(pmul(x,y),c); }
|
---|
229 |
|
---|
230 | EIGEN_STRONG_INLINE Packet2cf pmul(const Packet2cf& a, const Packet2cf& b) const
|
---|
231 | {
|
---|
232 | return pconj(internal::pmul(a, b));
|
---|
233 | }
|
---|
234 | };
|
---|
235 |
|
---|
236 | template<> EIGEN_STRONG_INLINE Packet2cf pdiv<Packet2cf>(const Packet2cf& a, const Packet2cf& b)
|
---|
237 | {
|
---|
238 | // TODO optimize it for AltiVec
|
---|
239 | Packet2cf res = conj_helper<Packet2cf,Packet2cf,false,true>().pmul(a,b);
|
---|
240 | Packet4f s, rev_s;
|
---|
241 |
|
---|
242 | // this computes the norm
|
---|
243 | s = vmulq_f32(b.v, b.v);
|
---|
244 | rev_s = vrev64q_f32(s);
|
---|
245 |
|
---|
246 | return Packet2cf(pdiv(res.v, vaddq_f32(s,rev_s)));
|
---|
247 | }
|
---|
248 |
|
---|
249 | } // end namespace internal
|
---|
250 |
|
---|
251 | } // end namespace Eigen
|
---|
252 |
|
---|
253 | #endif // EIGEN_COMPLEX_NEON_H
|
---|