gr-baz Package
volk_32fc_x2_multiply_32fc_u.h
Go to the documentation of this file.
1 #ifndef INCLUDED_volk_32fc_x2_multiply_32fc_u_H
2 #define INCLUDED_volk_32fc_x2_multiply_32fc_u_H
3 
4 #include <inttypes.h>
5 #include <stdio.h>
6 #include <volk/volk_complex.h>
7 #include <float.h>
8 
9 #ifdef LV_HAVE_SSE3
10 #include <pmmintrin.h>
11  /*!
12  \brief Multiplies the two input complex vectors and stores their results in the third vector
13  \param cVector The vector where the results will be stored
14  \param aVector One of the vectors to be multiplied
15  \param bVector One of the vectors to be multiplied
16  \param num_points The number of complex values in aVector and bVector to be multiplied together and stored into cVector
17  */
18 static inline void volk_32fc_x2_multiply_32fc_u_sse3(lv_32fc_t* cVector, const lv_32fc_t* aVector, const lv_32fc_t* bVector, unsigned int num_points){
19  unsigned int number = 0;
20  const unsigned int halfPoints = num_points / 2;
21 
22  __m128 x, y, yl, yh, z, tmp1, tmp2;
23  lv_32fc_t* c = cVector;
24  const lv_32fc_t* a = aVector;
25  const lv_32fc_t* b = bVector;
26 
27  for(;number < halfPoints; number++){
28 
29  x = _mm_loadu_ps((float*)a); // Load the ar + ai, br + bi as ar,ai,br,bi
30  y = _mm_loadu_ps((float*)b); // Load the cr + ci, dr + di as cr,ci,dr,di
31 
32  yl = _mm_moveldup_ps(y); // Load yl with cr,cr,dr,dr
33  yh = _mm_movehdup_ps(y); // Load yh with ci,ci,di,di
34 
35  tmp1 = _mm_mul_ps(x,yl); // tmp1 = ar*cr,ai*cr,br*dr,bi*dr
36 
37  x = _mm_shuffle_ps(x,x,0xB1); // Re-arrange x to be ai,ar,bi,br
38 
39  tmp2 = _mm_mul_ps(x,yh); // tmp2 = ai*ci,ar*ci,bi*di,br*di
40 
41  z = _mm_addsub_ps(tmp1,tmp2); // ar*cr-ai*ci, ai*cr+ar*ci, br*dr-bi*di, bi*dr+br*di
42 
43  _mm_storeu_ps((float*)c,z); // Store the results back into the C container
44 
45  a += 2;
46  b += 2;
47  c += 2;
48  }
49 
50  if((num_points % 2) != 0) {
51  *c = (*a) * (*b);
52  }
53 }
54 #endif /* LV_HAVE_SSE */
55 
56 #ifdef LV_HAVE_GENERIC
57  /*!
58  \brief Multiplies the two input complex vectors and stores their results in the third vector
59  \param cVector The vector where the results will be stored
60  \param aVector One of the vectors to be multiplied
61  \param bVector One of the vectors to be multiplied
62  \param num_points The number of complex values in aVector and bVector to be multiplied together and stored into cVector
63  */
64 static inline void volk_32fc_x2_multiply_32fc_u_generic(lv_32fc_t* cVector, const lv_32fc_t* aVector, const lv_32fc_t* bVector, unsigned int num_points){
65  lv_32fc_t* cPtr = cVector;
66  const lv_32fc_t* aPtr = aVector;
67  const lv_32fc_t* bPtr= bVector;
68  unsigned int number = 0;
69 
70  for(number = 0; number < num_points; number++){
71  *cPtr++ = (*aPtr++) * (*bPtr++);
72  }
73 }
74 #endif /* LV_HAVE_GENERIC */
75 
76 
77 #endif /* INCLUDED_volk_32fc_x2_multiply_32fc_u_H */