diff options
| -rw-r--r-- | volk/include/volk/volk_32fc_s32fc_multiply_32fc_u.h | 42 | ||||
| -rw-r--r-- | volk/include/volk/volk_32fc_x2_multiply_32fc_u.h | 77 |
2 files changed, 119 insertions, 0 deletions
diff --git a/volk/include/volk/volk_32fc_s32fc_multiply_32fc_u.h b/volk/include/volk/volk_32fc_s32fc_multiply_32fc_u.h new file mode 100644 index 000000000..201dcf5f6 --- /dev/null +++ b/volk/include/volk/volk_32fc_s32fc_multiply_32fc_u.h @@ -0,0 +1,42 @@ +#ifndef INCLUDED_volk_32fc_s32fc_multiply_32fc_u_H +#define INCLUDED_volk_32fc_s32fc_multiply_32fc_u_H + +#include <inttypes.h> +#include <stdio.h> +#include <volk/volk_complex.h> +#include <float.h> + +#ifdef LV_HAVE_GENERIC + /*! + \brief Multiplies the two input complex vectors and stores their results in the third vector + \param cVector The vector where the results will be stored + \param aVector One of the vectors to be multiplied + \param bVector One of the vectors to be multiplied + \param num_points The number of complex values in aVector and bVector to be multiplied together and stored into cVector + */ +static inline void volk_32fc_s32fc_multiply_32fc_u_generic(lv_32fc_t* cVector, const lv_32fc_t* aVector, const lv_32fc_t scalar, unsigned int num_points){ + lv_32fc_t* cPtr = cVector; + const lv_32fc_t* aPtr = aVector; + unsigned int number = num_points; + + // unwrap loop + while (number >= 8){ + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + *cPtr++ = (*aPtr++) * scalar; + number -= 8; + } + + // clean up any remaining + while (number-- > 0) + *cPtr++ = *aPtr++ * scalar; +} +#endif /* LV_HAVE_GENERIC */ + + +#endif /* INCLUDED_volk_32fc_x2_multiply_32fc_u_H */ diff --git a/volk/include/volk/volk_32fc_x2_multiply_32fc_u.h b/volk/include/volk/volk_32fc_x2_multiply_32fc_u.h new file mode 100644 index 000000000..729c1a4ad --- /dev/null +++ b/volk/include/volk/volk_32fc_x2_multiply_32fc_u.h @@ -0,0 +1,77 @@ +#ifndef INCLUDED_volk_32fc_x2_multiply_32fc_u_H +#define INCLUDED_volk_32fc_x2_multiply_32fc_u_H + +#include <inttypes.h> +#include <stdio.h> +#include <volk/volk_complex.h> +#include <float.h> + +#ifdef LV_HAVE_SSE3 +#include <pmmintrin.h> + /*! + \brief Multiplies the two input complex vectors and stores their results in the third vector + \param cVector The vector where the results will be stored + \param aVector One of the vectors to be multiplied + \param bVector One of the vectors to be multiplied + \param num_points The number of complex values in aVector and bVector to be multiplied together and stored into cVector + */ +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){ + unsigned int number = 0; + const unsigned int halfPoints = num_points / 2; + + __m128 x, y, yl, yh, z, tmp1, tmp2; + lv_32fc_t* c = cVector; + const lv_32fc_t* a = aVector; + const lv_32fc_t* b = bVector; + + for(;number < halfPoints; number++){ + + x = _mm_loadu_ps((float*)a); // Load the ar + ai, br + bi as ar,ai,br,bi + y = _mm_loadu_ps((float*)b); // Load the cr + ci, dr + di as cr,ci,dr,di + + yl = _mm_moveldup_ps(y); // Load yl with cr,cr,dr,dr + yh = _mm_movehdup_ps(y); // Load yh with ci,ci,di,di + + tmp1 = _mm_mul_ps(x,yl); // tmp1 = ar*cr,ai*cr,br*dr,bi*dr + + x = _mm_shuffle_ps(x,x,0xB1); // Re-arrange x to be ai,ar,bi,br + + tmp2 = _mm_mul_ps(x,yh); // tmp2 = ai*ci,ar*ci,bi*di,br*di + + z = _mm_addsub_ps(tmp1,tmp2); // ar*cr-ai*ci, ai*cr+ar*ci, br*dr-bi*di, bi*dr+br*di + + _mm_storeu_ps((float*)c,z); // Store the results back into the C container + + a += 2; + b += 2; + c += 2; + } + + if((num_points % 2) != 0) { + *c = (*a) * (*b); + } +} +#endif /* LV_HAVE_SSE */ + +#ifdef LV_HAVE_GENERIC + /*! + \brief Multiplies the two input complex vectors and stores their results in the third vector + \param cVector The vector where the results will be stored + \param aVector One of the vectors to be multiplied + \param bVector One of the vectors to be multiplied + \param num_points The number of complex values in aVector and bVector to be multiplied together and stored into cVector + */ +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){ + lv_32fc_t* cPtr = cVector; + const lv_32fc_t* aPtr = aVector; + const lv_32fc_t* bPtr= bVector; + unsigned int number = 0; + + for(number = 0; number < num_points; number++){ + *cPtr++ = (*aPtr++) * (*bPtr++); + } +} +#endif /* LV_HAVE_GENERIC */ + + +#endif /* INCLUDED_volk_32fc_x2_multiply_32fc_u_H */ |