mirrors/gnss-sdr
1
0
Fork 0
mirror of https://github.com/gnss-sdr/gnss-sdr synced 2025-02-22 22:10:19 +00:00
Code Issues Releases Wiki Activity

adding neon implementation

Browse Source
This commit is contained in:
Carles Fernandez 2016-01-25 18:13:54 +01:00
parent 3306c21cf8
commit ba8f0e86b2
1 changed files with 32 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_convert_32fc.h
View File

@ -107,4 +107,36 @@ static inline void volk_gnsssdr_16ic_convert_32fc_u_sse2(lv_32fc_t* outputVector
// a = _mm_load_si128((__m128i*)_in); //load (2 byte imag, 2 byte real) x 4 into 128 bits reg
//use _mm_cvtepi16_epi32 !!!!
#ifdef LV_HAVE_NEON
#include <arm_neon.h>
static inline void volk_gnsssdr_16ic_convert_32fc_neon(lv_32fc_t* outputVector, const lv_16sc_t* inputVector, unsigned int num_points)
{
const unsigned int sse_iters = num_points / 2;
const lv_16sc_t* _in = inputVector;
lv_32fc_t* _out = outputVector;
int16x4_t a16x4;
int32x4_t a32x4;
float32x4_t f32x4;
for(unsigned int number = 0; number < sse_iters; number++)
{
a16x4 = vld1_s16((const int16_t*)_in);
__builtin_prefetch(_in + 4);
a32x4 = vmovl_s16(a16x4);
f32x4 = vcvtq_f32_s32(a32x4);
vst1q_f32((float32_t*)_out, f32x4);
_in += 2;
_out += 2;
}
for (unsigned int i = 0; i < (num_points % 2); ++i)
{
*_out++ = lv_cmake((float)lv_creal(*_in), (float)lv_cimag(*_in));
_in++;
}
}
#endif /* LV_HAVE_NEON */
#endif /* INCLUDED_volk_gnsssdr_32fc_convert_16ic_H */