mirrors/gnss-sdr
1
0
Fork 0
mirror of https://github.com/gnss-sdr/gnss-sdr synced 2024-12-14 12:10:34 +00:00
Code Issues Releases Wiki Activity

fixing resampler

Browse Source
This commit is contained in:
Carles Fernandez 2016-04-07 12:42:24 +02:00
parent 58259568d6
commit 9cb60ec948
1 changed files with 105 additions and 113 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

218
src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_xn_resampler2_16ic_xn.h
View File

@ -43,7 +43,7 @@
* *
* <b>Dispatcher Prototype</b> * <b>Dispatcher Prototype</b>
* \code * \code
* void volk_gnsssdr_16ic_xn_resampler2_16ic_xn(lv_16sc_t** result, const lv_16sc_t* local_code, float* rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) * void volk_gnsssdr_16ic_xn_resampler2_16ic_xn(lv_16sc_t** result, const lv_16sc_t* local_code, float* rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
* \endcode * \endcode
* *
* \b Inputs * \b Inputs
@ -52,8 +52,8 @@
* \li code_phase_step_chips: Phase increment per sample [chips/sample]. * \li code_phase_step_chips: Phase increment per sample [chips/sample].
* \li shifts_chips: Vector of floats that defines the spacing (in chips) between the replicas of \p local_code * \li shifts_chips: Vector of floats that defines the spacing (in chips) between the replicas of \p local_code
* \li code_length_chips: Code length in chips. * \li code_length_chips: Code length in chips.
* \li num_out_vectors Number of output vectors. * \li num_out_vectors: Number of output vectors.
* \li num_output_samples: The number of data values to be in the resampled vector. * \li num_points: The number of data values to be in the resampled vector.
* *
* \b Outputs * \b Outputs
* \li result: Pointer to a vector of pointers where the results will be stored. * \li result: Pointer to a vector of pointers where the results will be stored.
@ -70,18 +70,18 @@
#ifdef LV_HAVE_GENERIC #ifdef LV_HAVE_GENERIC
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_generic(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_generic(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
int local_code_chip_index; int local_code_chip_index;
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++) for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{ {
for (int n = 0; n < num_output_samples; n++) for (int n = 0; n < num_points; n++)
{ {
// resample code for current tap // resample code for current tap
local_code_chip_index = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index = local_code_chip_index % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index < 0) local_code_chip_index += code_length_chips; if (local_code_chip_index < 0) local_code_chip_index += code_length_chips;
local_code_chip_index = local_code_chip_index % code_length_chips;
result[current_correlator_tap][n] = local_code[local_code_chip_index]; result[current_correlator_tap][n] = local_code[local_code_chip_index];
} }
} }
@ -92,10 +92,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_generic(lv_16sc_t** r
#ifdef LV_HAVE_SSE4_1 #ifdef LV_HAVE_SSE4_1
#include <smmintrin.h> #include <smmintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse4_1(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse4_1(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int quarterPoints = num_output_samples / 4; const unsigned int quarterPoints = num_points / 4;
const __m128 fours = _mm_set1_ps(4.0f); const __m128 fours = _mm_set1_ps(4.0f);
const __m128 rem_code_phase_chips_reg = _mm_set_ps1(rem_code_phase_chips); const __m128 rem_code_phase_chips_reg = _mm_set_ps1(rem_code_phase_chips);
@ -139,13 +139,13 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse4_1(lv_16sc_t**
} }
indexn = _mm_add_ps(indexn, fours); indexn = _mm_add_ps(indexn, fours);
} }
for(unsigned int n = quarterPoints * 4; n < num_output_samples; n++) for(unsigned int n = quarterPoints * 4; n < num_points; n++)
{ {
// resample code for current tap // resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips; if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_]; _result[current_correlator_tap][n] = local_code[local_code_chip_index_];
} }
} }
@ -156,10 +156,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse4_1(lv_16sc_t**
#ifdef LV_HAVE_SSE4_1 #ifdef LV_HAVE_SSE4_1
#include <smmintrin.h> #include <smmintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse4_1(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse4_1(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int quarterPoints = num_output_samples / 4; const unsigned int quarterPoints = num_points / 4;
const __m128 fours = _mm_set1_ps(4.0f); const __m128 fours = _mm_set1_ps(4.0f);
const __m128 rem_code_phase_chips_reg = _mm_set_ps1(rem_code_phase_chips); const __m128 rem_code_phase_chips_reg = _mm_set_ps1(rem_code_phase_chips);
@ -203,13 +203,13 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse4_1(lv_16sc_t**
} }
indexn = _mm_add_ps(indexn, fours); indexn = _mm_add_ps(indexn, fours);
} }
for(unsigned int n = quarterPoints * 4; n < num_output_samples; n++) for(unsigned int n = quarterPoints * 4; n < num_points; n++)
{ {
// resample code for current tap // resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips; if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_]; _result[current_correlator_tap][n] = local_code[local_code_chip_index_];
} }
} }
@ -220,10 +220,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse4_1(lv_16sc_t**
#ifdef LV_HAVE_SSE3 #ifdef LV_HAVE_SSE3
#include <pmmintrin.h> #include <pmmintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse3(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse3(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int quarterPoints = num_output_samples / 4; const unsigned int quarterPoints = num_points / 4;
const __m128 ones = _mm_set1_ps(1.0f); const __m128 ones = _mm_set1_ps(1.0f);
const __m128 fours = _mm_set1_ps(4.0f); const __m128 fours = _mm_set1_ps(4.0f);
@ -246,7 +246,6 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse3(lv_16sc_t** re
__m128 indexn = _mm_set_ps(3.0f, 2.0f, 1.0f, 0.0f); __m128 indexn = _mm_set_ps(3.0f, 2.0f, 1.0f, 0.0f);
for(unsigned int n = 0; n < quarterPoints; n++) for(unsigned int n = 0; n < quarterPoints; n++)
{ {
//__builtin_prefetch(&_result[current_correlator_tap][4 * n] + 8, 1, 0);
aux = _mm_mul_ps(code_phase_step_chips_reg, indexn); aux = _mm_mul_ps(code_phase_step_chips_reg, indexn);
aux = _mm_add_ps(aux, aux2); aux = _mm_add_ps(aux, aux2);
// floor // floor
@ -272,14 +271,13 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse3(lv_16sc_t** re
} }
indexn = _mm_add_ps(indexn, fours); indexn = _mm_add_ps(indexn, fours);
} }
for(unsigned int n = quarterPoints * 4; n < num_output_samples; n++) for(unsigned int n = quarterPoints * 4; n < num_points; n++)
{ {
//__builtin_prefetch(&_result[current_correlator_tap][n] + 2, 1, 0);
// resample code for current tap // resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips; if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_]; _result[current_correlator_tap][n] = local_code[local_code_chip_index_];
} }
} }
@ -290,10 +288,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_sse3(lv_16sc_t** re
#ifdef LV_HAVE_SSE3 #ifdef LV_HAVE_SSE3
#include <pmmintrin.h> #include <pmmintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse3(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse3(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int quarterPoints = num_output_samples / 4; const unsigned int quarterPoints = num_points / 4;
const __m128 ones = _mm_set1_ps(1.0f); const __m128 ones = _mm_set1_ps(1.0f);
const __m128 fours = _mm_set1_ps(4.0f); const __m128 fours = _mm_set1_ps(4.0f);
@ -316,8 +314,6 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse3(lv_16sc_t** re
__m128 indexn = _mm_set_ps(3.0f, 2.0f, 1.0f, 0.0f); __m128 indexn = _mm_set_ps(3.0f, 2.0f, 1.0f, 0.0f);
for(unsigned int n = 0; n < quarterPoints; n++) for(unsigned int n = 0; n < quarterPoints; n++)
{ {
//__builtin_prefetch(&_result[current_correlator_tap][4 * n + 3], 1, 0);
//__builtin_prefetch(&local_code_chip_index[4]);
aux = _mm_mul_ps(code_phase_step_chips_reg, indexn); aux = _mm_mul_ps(code_phase_step_chips_reg, indexn);
aux = _mm_add_ps(aux, aux2); aux = _mm_add_ps(aux, aux2);
// floor // floor
@ -343,14 +339,13 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse3(lv_16sc_t** re
} }
indexn = _mm_add_ps(indexn, fours); indexn = _mm_add_ps(indexn, fours);
} }
for(unsigned int n = quarterPoints * 4; n < num_output_samples; n++) for(unsigned int n = quarterPoints * 4; n < num_points; n++)
{ {
//__builtin_prefetch(&_result[current_correlator_tap][n], 1, 0);
// resample code for current tap // resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips; if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_]; _result[current_correlator_tap][n] = local_code[local_code_chip_index_];
} }
} }
@ -361,10 +356,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_sse3(lv_16sc_t** re
#ifdef LV_HAVE_AVX #ifdef LV_HAVE_AVX
#include <immintrin.h> #include <immintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_avx(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_avx(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int avx_iters = num_output_samples / 8; const unsigned int avx_iters = num_points / 8;
const __m256 eights = _mm256_set1_ps(8.0f); const __m256 eights = _mm256_set1_ps(8.0f);
const __m256 rem_code_phase_chips_reg = _mm256_set1_ps(rem_code_phase_chips); const __m256 rem_code_phase_chips_reg = _mm256_set1_ps(rem_code_phase_chips);
@ -375,87 +370,16 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_a_avx(lv_16sc_t** res
const __m256 zeros = _mm256_setzero_ps(); const __m256 zeros = _mm256_setzero_ps();
const __m256 code_length_chips_reg_f = _mm256_set1_ps((float)code_length_chips); const __m256 code_length_chips_reg_f = _mm256_set1_ps((float)code_length_chips);
const __m256 n0 = _mm256_set_ps(7.0f, 6.0f, 5.0f, 4.0f, 3.0f, 2.0f, 1.0f, 0.0f);
__m256i local_code_chip_index_reg, i; __m256i local_code_chip_index_reg, i;
__m256 aux, aux2, shifts_chips_reg, c, cTrunc, base, negatives; __m256 aux, aux2, aux3, shifts_chips_reg, c, cTrunc, base, negatives, indexn;
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++) for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{ {
shifts_chips_reg = _mm256_set1_ps((float)shifts_chips[current_correlator_tap]); shifts_chips_reg = _mm256_set1_ps((float)shifts_chips[current_correlator_tap]);
aux2 = _mm256_sub_ps(shifts_chips_reg, rem_code_phase_chips_reg); aux2 = _mm256_sub_ps(shifts_chips_reg, rem_code_phase_chips_reg);
__m256 indexn = _mm256_set_ps(7.0f, 6.0f, 5.0f, 4.0f, 3.0f, 2.0f, 1.0f, 0.0f); indexn = n0;
for(unsigned int n = 0; n < avx_iters; n++)
{
__builtin_prefetch(&_result[current_correlator_tap][8 * n + 7], 1, 0);
__builtin_prefetch(&local_code_chip_index[8]);
aux = _mm256_mul_ps(code_phase_step_chips_reg, indexn);
aux = _mm256_add_ps(aux, aux2);
// floor
aux = _mm256_floor_ps(aux);
// fmod
c = _mm256_div_ps(aux, code_length_chips_reg_f);
i = _mm256_cvttps_epi32(c);
cTrunc = _mm256_cvtepi32_ps(i);
base = _mm256_mul_ps(cTrunc, code_length_chips_reg_f);
aux = _mm256_sub_ps(aux, base);
negatives = _mm256_cmp_ps(aux, zeros, 0x01);
aux2 = _mm256_and_ps(code_length_chips_reg_f, negatives);
local_code_chip_index_reg = _mm256_cvtps_epi32(_mm256_add_ps(aux, aux2));
_mm256_store_si256((__m256i*)local_code_chip_index, local_code_chip_index_reg);
for(unsigned int k = 0; k < 8; ++k)
{
_result[current_correlator_tap][n * 8 + k] = local_code[local_code_chip_index[k]];
}
indexn = _mm256_add_ps(indexn, eights);
}
}
_mm256_zeroupper();
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{
for(unsigned int n = avx_iters * 8; n < num_output_samples; n++)
{
// resample code for current tap
__builtin_prefetch(&_result[current_correlator_tap][n], 1, 0);
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_];
}
}
}
#endif
#ifdef LV_HAVE_AVX
#include <immintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_avx(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples)
{
lv_16sc_t** _result = result;
const unsigned int avx_iters = num_output_samples / 8;
const __m256 eights = _mm256_set1_ps(8.0f);
const __m256 rem_code_phase_chips_reg = _mm256_set1_ps(rem_code_phase_chips);
const __m256 code_phase_step_chips_reg = _mm256_set1_ps(code_phase_step_chips);
__VOLK_ATTR_ALIGNED(32) int local_code_chip_index[8];
int local_code_chip_index_;
const __m256 zeros = _mm256_setzero_ps();
const __m256 code_length_chips_reg_f = _mm256_set1_ps((float)code_length_chips);
__m256i local_code_chip_index_reg, i;
__m256 aux, aux2, shifts_chips_reg, c, cTrunc, base, negatives;
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{
shifts_chips_reg = _mm256_set1_ps((float)shifts_chips[current_correlator_tap]);
aux2 = _mm256_sub_ps(shifts_chips_reg, rem_code_phase_chips_reg);
__m256 indexn = _mm256_set_ps(7.0f, 6.0f, 5.0f, 4.0f, 3.0f, 2.0f, 1.0f, 0.0f);
for(unsigned int n = 0; n < avx_iters; n++) for(unsigned int n = 0; n < avx_iters; n++)
{ {
aux = _mm256_mul_ps(code_phase_step_chips_reg, indexn); aux = _mm256_mul_ps(code_phase_step_chips_reg, indexn);
@ -463,16 +387,15 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_avx(lv_16sc_t** res
// floor // floor
aux = _mm256_floor_ps(aux); aux = _mm256_floor_ps(aux);
negatives = _mm256_cmp_ps(aux, zeros, 0x01);
aux3 = _mm256_and_ps(code_length_chips_reg_f, negatives);
aux = _mm256_add_ps(aux, aux3);
// fmod // fmod
c = _mm256_div_ps(aux, code_length_chips_reg_f); c = _mm256_div_ps(aux, code_length_chips_reg_f);
i = _mm256_cvttps_epi32(c); i = _mm256_cvttps_epi32(c);
cTrunc = _mm256_cvtepi32_ps(i); cTrunc = _mm256_cvtepi32_ps(i);
base = _mm256_mul_ps(cTrunc, code_length_chips_reg_f); base = _mm256_mul_ps(cTrunc, code_length_chips_reg_f);
aux = _mm256_sub_ps(aux, base); local_code_chip_index_reg = _mm256_cvttps_epi32(_mm256_sub_ps(aux, base));
negatives = _mm256_cmp_ps(aux, zeros, 0x01);
aux2 = _mm256_and_ps(code_length_chips_reg_f, negatives);
local_code_chip_index_reg = _mm256_cvtps_epi32(_mm256_add_ps(aux, aux2));
_mm256_store_si256((__m256i*)local_code_chip_index, local_code_chip_index_reg); _mm256_store_si256((__m256i*)local_code_chip_index, local_code_chip_index_reg);
for(unsigned int k = 0; k < 8; ++k) for(unsigned int k = 0; k < 8; ++k)
@ -485,13 +408,82 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_avx(lv_16sc_t** res
_mm256_zeroupper(); _mm256_zeroupper();
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++) for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{ {
for(unsigned int n = avx_iters * 8; n < num_output_samples; n++) for(unsigned int n = avx_iters * 8; n < num_points; n++)
{ {
// resample code for current tap // resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips); local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
//Take into account that in multitap correlators, the shifts can be negative! //Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips; if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_];
}
}
}
#endif
#ifdef LV_HAVE_AVX
#include <immintrin.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_avx(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{
lv_16sc_t** _result = result;
const unsigned int avx_iters = num_points / 8;
const __m256 eights = _mm256_set1_ps(8.0f);
const __m256 rem_code_phase_chips_reg = _mm256_set1_ps(rem_code_phase_chips);
const __m256 code_phase_step_chips_reg = _mm256_set1_ps(code_phase_step_chips);
__VOLK_ATTR_ALIGNED(32) int local_code_chip_index[8];
int local_code_chip_index_;
const __m256 zeros = _mm256_setzero_ps();
const __m256 code_length_chips_reg_f = _mm256_set1_ps((float)code_length_chips);
const __m256 n0 = _mm256_set_ps(7.0f, 6.0f, 5.0f, 4.0f, 3.0f, 2.0f, 1.0f, 0.0f);
__m256i local_code_chip_index_reg, i;
__m256 aux, aux2, aux3, shifts_chips_reg, c, cTrunc, base, negatives, indexn;
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{
shifts_chips_reg = _mm256_set1_ps((float)shifts_chips[current_correlator_tap]);
aux2 = _mm256_sub_ps(shifts_chips_reg, rem_code_phase_chips_reg);
indexn = n0;
for(unsigned int n = 0; n < avx_iters; n++)
{
aux = _mm256_mul_ps(code_phase_step_chips_reg, indexn);
aux = _mm256_add_ps(aux, aux2);
// floor
aux = _mm256_floor_ps(aux);
negatives = _mm256_cmp_ps(aux, zeros, 0x01);
aux3 = _mm256_and_ps(code_length_chips_reg_f, negatives);
aux = _mm256_add_ps(aux, aux3);
// fmod
c = _mm256_div_ps(aux, code_length_chips_reg_f);
i = _mm256_cvttps_epi32(c);
cTrunc = _mm256_cvtepi32_ps(i);
base = _mm256_mul_ps(cTrunc, code_length_chips_reg_f);
local_code_chip_index_reg = _mm256_cvttps_epi32(_mm256_sub_ps(aux, base));
_mm256_store_si256((__m256i*)local_code_chip_index, local_code_chip_index_reg);
for(unsigned int k = 0; k < 8; ++k)
{
_result[current_correlator_tap][n * 8 + k] = local_code[local_code_chip_index[k]];
}
indexn = _mm256_add_ps(indexn, eights);
}
}
_mm256_zeroupper();
for (int current_correlator_tap = 0; current_correlator_tap < num_out_vectors; current_correlator_tap++)
{
for(unsigned int n = avx_iters * 8; n < num_points; n++)
{
// resample code for current tap
local_code_chip_index_ = (int)floor(code_phase_step_chips * (float)n + shifts_chips[current_correlator_tap] - rem_code_phase_chips);
//Take into account that in multitap correlators, the shifts can be negative!
if (local_code_chip_index_ < 0) local_code_chip_index_ += code_length_chips;
local_code_chip_index_ = local_code_chip_index_ % code_length_chips;
_result[current_correlator_tap][n] = local_code[local_code_chip_index_]; _result[current_correlator_tap][n] = local_code[local_code_chip_index_];
} }
} }
@ -502,10 +494,10 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_u_avx(lv_16sc_t** res
#ifdef LV_HAVE_NEON #ifdef LV_HAVE_NEON
#include <arm_neon.h> #include <arm_neon.h>
static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_neon(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_output_samples) static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_neon(lv_16sc_t** result, const lv_16sc_t* local_code, float rem_code_phase_chips, float code_phase_step_chips, float* shifts_chips, unsigned int code_length_chips, int num_out_vectors, unsigned int num_points)
{ {
lv_16sc_t** _result = result; lv_16sc_t** _result = result;
const unsigned int neon_iters = num_output_samples / 4; const unsigned int neon_iters = num_points / 4;
const int32x4_t ones = vdupq_n_s32(1); const int32x4_t ones = vdupq_n_s32(1);
const float32x4_t fours = vdupq_n_f32(4.0f); const float32x4_t fours = vdupq_n_f32(4.0f);
const float32x4_t rem_code_phase_chips_reg = vdupq_n_f32(rem_code_phase_chips); const float32x4_t rem_code_phase_chips_reg = vdupq_n_f32(rem_code_phase_chips);
@ -562,7 +554,7 @@ static inline void volk_gnsssdr_16ic_xn_resampler2_16ic_xn_neon(lv_16sc_t** resu
} }
indexn = vaddq_f32(indexn, fours); indexn = vaddq_f32(indexn, fours);
} }
for(unsigned int n = neon_iters * 4; n < num_output_samples; n++) for(unsigned int n = neon_iters * 4; n < num_points; n++)
{ {
__builtin_prefetch(&_result[current_correlator_tap][n], 1, 0); __builtin_prefetch(&_result[current_correlator_tap][n], 1, 0);
// resample code for current tap // resample code for current tap