mirror of https://github.com/gnss-sdr/gnss-sdr
559 lines
21 KiB
C++
559 lines
21 KiB
C++
/*!
|
|
* \file volk_gnsssdr_16ic_x2_dot_prod_16ic.h
|
|
* \brief VOLK_GNSSSDR kernel: multiplies two 16 bits vectors and accumulates them.
|
|
* \authors <ul>
|
|
* <li> Javier Arribas, 2015. jarribas(at)cttc.es
|
|
* </ul>
|
|
*
|
|
* VOLK_GNSSSDR kernel that multiplies two 16 bits vectors (8 bits the real part
|
|
* and 8 bits the imaginary part) and accumulates them
|
|
*
|
|
* -----------------------------------------------------------------------------
|
|
*
|
|
* GNSS-SDR is a Global Navigation Satellite System software-defined receiver.
|
|
* This file is part of GNSS-SDR.
|
|
*
|
|
* Copyright (C) 2010-2020 (see AUTHORS file for a list of contributors)
|
|
* SPDX-License-Identifier: GPL-3.0-or-later
|
|
*
|
|
* -----------------------------------------------------------------------------
|
|
*/
|
|
|
|
/*!
|
|
* \page volk_gnsssdr_16ic_x2_dot_prod_16ic
|
|
*
|
|
* \b Overview
|
|
*
|
|
* Multiplies two input complex vectors (16-bit integer each component) and accumulates them,
|
|
* storing the result. Results are saturated so never go beyond the limits of the data type.
|
|
*
|
|
* <b>Dispatcher Prototype</b>
|
|
* \code
|
|
* void volk_gnsssdr_16ic_x2_dot_prod_16ic(lv_16sc_t* result, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points);
|
|
* \endcode
|
|
*
|
|
* \b Inputs
|
|
* \li in_a: One of the vectors to be multiplied and accumulated.
|
|
* \li in_b: The other vector to be multiplied and accumulated.
|
|
* \li num_points: Number of complex values to be multiplied together, accumulated and stored into \p result
|
|
*
|
|
* \b Outputs
|
|
* \li result: Value of the accumulated result.
|
|
*
|
|
*/
|
|
|
|
#ifndef INCLUDED_volk_gnsssdr_16ic_x2_dot_prod_16ic_H
|
|
#define INCLUDED_volk_gnsssdr_16ic_x2_dot_prod_16ic_H
|
|
|
|
#include <volk_gnsssdr/saturation_arithmetic.h>
|
|
#include <volk_gnsssdr/volk_gnsssdr_common.h>
|
|
#include <volk_gnsssdr/volk_gnsssdr_complex.h>
|
|
|
|
|
|
#ifdef LV_HAVE_GENERIC
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_generic(lv_16sc_t* result, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
result[0] = lv_cmake((int16_t)0, (int16_t)0);
|
|
unsigned int n;
|
|
for (n = 0; n < num_points; n++)
|
|
{
|
|
lv_16sc_t tmp = in_a[n] * in_b[n];
|
|
result[0] = lv_cmake(sat_adds16i(lv_creal(result[0]), lv_creal(tmp)), sat_adds16i(lv_cimag(result[0]), lv_cimag(tmp)));
|
|
}
|
|
}
|
|
|
|
#endif /* LV_HAVE_GENERIC */
|
|
|
|
|
|
#ifdef LV_HAVE_SSE2
|
|
#include <emmintrin.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_a_sse2(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
lv_16sc_t dotProduct = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
const unsigned int sse_iters = num_points / 4;
|
|
unsigned int number;
|
|
const lv_16sc_t* _in_a = in_a;
|
|
const lv_16sc_t* _in_b = in_b;
|
|
lv_16sc_t* _out = out;
|
|
|
|
if (sse_iters > 0)
|
|
{
|
|
__m128i a, b, c, c_sr, mask_imag, mask_real, real, imag, imag1, imag2, b_sl, a_sl, realcacc, imagcacc;
|
|
__VOLK_ATTR_ALIGNED(16)
|
|
lv_16sc_t dotProductVector[4];
|
|
|
|
realcacc = _mm_setzero_si128();
|
|
imagcacc = _mm_setzero_si128();
|
|
|
|
mask_imag = _mm_set_epi8(0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0);
|
|
mask_real = _mm_set_epi8(0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF);
|
|
|
|
for (number = 0; number < sse_iters; number++)
|
|
{
|
|
// a[127:0]=[a3.i,a3.r,a2.i,a2.r,a1.i,a1.r,a0.i,a0.r]
|
|
a = _mm_load_si128((__m128i*)_in_a); // load (2 byte imag, 2 byte real) x 4 into 128 bits reg
|
|
__VOLK_GNSSSDR_PREFETCH(_in_a + 8);
|
|
b = _mm_load_si128((__m128i*)_in_b);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_b + 8);
|
|
c = _mm_mullo_epi16(a, b); // a3.i*b3.i, a3.r*b3.r, ....
|
|
|
|
c_sr = _mm_srli_si128(c, 2); // Shift a right by imm8 bytes while shifting in zeros, and store the results in dst.
|
|
real = _mm_subs_epi16(c, c_sr);
|
|
|
|
b_sl = _mm_slli_si128(b, 2); // b3.r, b2.i ....
|
|
a_sl = _mm_slli_si128(a, 2); // a3.r, a2.i ....
|
|
|
|
imag1 = _mm_mullo_epi16(a, b_sl); // a3.i*b3.r, ....
|
|
imag2 = _mm_mullo_epi16(b, a_sl); // b3.i*a3.r, ....
|
|
|
|
imag = _mm_adds_epi16(imag1, imag2); // with saturation arithmetic!
|
|
|
|
realcacc = _mm_adds_epi16(realcacc, real);
|
|
imagcacc = _mm_adds_epi16(imagcacc, imag);
|
|
|
|
_in_a += 4;
|
|
_in_b += 4;
|
|
}
|
|
|
|
realcacc = _mm_and_si128(realcacc, mask_real);
|
|
imagcacc = _mm_and_si128(imagcacc, mask_imag);
|
|
|
|
a = _mm_or_si128(realcacc, imagcacc);
|
|
|
|
_mm_store_si128((__m128i*)dotProductVector, a); // Store the results back into the dot product vector
|
|
|
|
for (number = 0; number < 4; ++number)
|
|
{
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(dotProductVector[number])), sat_adds16i(lv_cimag(dotProduct), lv_cimag(dotProductVector[number])));
|
|
}
|
|
}
|
|
|
|
for (number = 0; number < (num_points % 4); ++number)
|
|
{
|
|
lv_16sc_t tmp = (*_in_a++) * (*_in_b++);
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(tmp)), sat_adds16i(lv_cimag(dotProduct), lv_cimag(tmp)));
|
|
}
|
|
|
|
*_out = dotProduct;
|
|
}
|
|
|
|
#endif /* LV_HAVE_SSE2 */
|
|
|
|
|
|
#ifdef LV_HAVE_SSE2
|
|
#include <emmintrin.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_u_sse2(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
lv_16sc_t dotProduct = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
const unsigned int sse_iters = num_points / 4;
|
|
|
|
const lv_16sc_t* _in_a = in_a;
|
|
const lv_16sc_t* _in_b = in_b;
|
|
lv_16sc_t* _out = out;
|
|
unsigned int i;
|
|
unsigned int number;
|
|
|
|
if (sse_iters > 0)
|
|
{
|
|
__m128i a, b, c, c_sr, mask_imag, mask_real, real, imag, imag1, imag2, b_sl, a_sl, realcacc, imagcacc, result;
|
|
__VOLK_ATTR_ALIGNED(16)
|
|
lv_16sc_t dotProductVector[4];
|
|
|
|
realcacc = _mm_setzero_si128();
|
|
imagcacc = _mm_setzero_si128();
|
|
|
|
mask_imag = _mm_set_epi8(0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0);
|
|
mask_real = _mm_set_epi8(0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF);
|
|
|
|
for (number = 0; number < sse_iters; number++)
|
|
{
|
|
// std::complex<T> memory structure: real part -> reinterpret_cast<cv T*>(a)[2*i]
|
|
// imaginary part -> reinterpret_cast<cv T*>(a)[2*i + 1]
|
|
// a[127:0]=[a3.i,a3.r,a2.i,a2.r,a1.i,a1.r,a0.i,a0.r]
|
|
a = _mm_loadu_si128((__m128i*)_in_a); // load (2 byte imag, 2 byte real) x 4 into 128 bits reg
|
|
__VOLK_GNSSSDR_PREFETCH(_in_a + 8);
|
|
b = _mm_loadu_si128((__m128i*)_in_b);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_b + 8);
|
|
c = _mm_mullo_epi16(a, b); // a3.i*b3.i, a3.r*b3.r, ....
|
|
|
|
c_sr = _mm_srli_si128(c, 2); // Shift a right by imm8 bytes while shifting in zeros, and store the results in dst.
|
|
real = _mm_subs_epi16(c, c_sr);
|
|
|
|
b_sl = _mm_slli_si128(b, 2); // b3.r, b2.i ....
|
|
a_sl = _mm_slli_si128(a, 2); // a3.r, a2.i ....
|
|
|
|
imag1 = _mm_mullo_epi16(a, b_sl); // a3.i*b3.r, ....
|
|
imag2 = _mm_mullo_epi16(b, a_sl); // b3.i*a3.r, ....
|
|
|
|
imag = _mm_adds_epi16(imag1, imag2); // with saturation arithmetic!
|
|
|
|
realcacc = _mm_adds_epi16(realcacc, real);
|
|
imagcacc = _mm_adds_epi16(imagcacc, imag);
|
|
|
|
_in_a += 4;
|
|
_in_b += 4;
|
|
}
|
|
|
|
realcacc = _mm_and_si128(realcacc, mask_real);
|
|
imagcacc = _mm_and_si128(imagcacc, mask_imag);
|
|
|
|
result = _mm_or_si128(realcacc, imagcacc);
|
|
|
|
_mm_storeu_si128((__m128i*)dotProductVector, result); // Store the results back into the dot product vector
|
|
|
|
for (i = 0; i < 4; ++i)
|
|
{
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(dotProductVector[i])), sat_adds16i(lv_cimag(dotProduct), lv_cimag(dotProductVector[i])));
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < (num_points % 4); ++i)
|
|
{
|
|
lv_16sc_t tmp = (*_in_a++) * (*_in_b++);
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(tmp)), sat_adds16i(lv_cimag(dotProduct), lv_cimag(tmp)));
|
|
}
|
|
|
|
*_out = dotProduct;
|
|
}
|
|
#endif /* LV_HAVE_SSE2 */
|
|
|
|
|
|
#ifdef LV_HAVE_AVX2
|
|
#include <immintrin.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_u_axv2(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
lv_16sc_t dotProduct = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
const unsigned int avx_iters = num_points / 8;
|
|
|
|
const lv_16sc_t* _in_a = in_a;
|
|
const lv_16sc_t* _in_b = in_b;
|
|
lv_16sc_t* _out = out;
|
|
unsigned int i;
|
|
unsigned int number;
|
|
|
|
if (avx_iters > 0)
|
|
{
|
|
__m256i a, b, c, c_sr, mask_imag, mask_real, real, imag, imag1, imag2, b_sl, a_sl, realcacc, imagcacc, result;
|
|
__VOLK_ATTR_ALIGNED(32)
|
|
lv_16sc_t dotProductVector[8];
|
|
|
|
realcacc = _mm256_setzero_si256();
|
|
imagcacc = _mm256_setzero_si256();
|
|
|
|
mask_imag = _mm256_set_epi8(0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0);
|
|
mask_real = _mm256_set_epi8(0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF);
|
|
|
|
for (number = 0; number < avx_iters; number++)
|
|
{
|
|
a = _mm256_loadu_si256((__m256i*)_in_a);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_a + 16);
|
|
b = _mm256_loadu_si256((__m256i*)_in_b);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_b + 16);
|
|
c = _mm256_mullo_epi16(a, b);
|
|
|
|
c_sr = _mm256_srli_si256(c, 2); // Shift a right by imm8 bytes while shifting in zeros, and store the results in dst.
|
|
real = _mm256_subs_epi16(c, c_sr);
|
|
|
|
b_sl = _mm256_slli_si256(b, 2);
|
|
a_sl = _mm256_slli_si256(a, 2);
|
|
|
|
imag1 = _mm256_mullo_epi16(a, b_sl);
|
|
imag2 = _mm256_mullo_epi16(b, a_sl);
|
|
|
|
imag = _mm256_adds_epi16(imag1, imag2); // with saturation arithmetic!
|
|
|
|
realcacc = _mm256_adds_epi16(realcacc, real);
|
|
imagcacc = _mm256_adds_epi16(imagcacc, imag);
|
|
|
|
_in_a += 8;
|
|
_in_b += 8;
|
|
}
|
|
|
|
realcacc = _mm256_and_si256(realcacc, mask_real);
|
|
imagcacc = _mm256_and_si256(imagcacc, mask_imag);
|
|
|
|
result = _mm256_or_si256(realcacc, imagcacc);
|
|
|
|
_mm256_storeu_si256((__m256i*)dotProductVector, result); // Store the results back into the dot product vector
|
|
|
|
for (i = 0; i < 8; ++i)
|
|
{
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(dotProductVector[i])), sat_adds16i(lv_cimag(dotProduct), lv_cimag(dotProductVector[i])));
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < (num_points % 8); ++i)
|
|
{
|
|
lv_16sc_t tmp = (*_in_a++) * (*_in_b++);
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(tmp)), sat_adds16i(lv_cimag(dotProduct), lv_cimag(tmp)));
|
|
}
|
|
|
|
*_out = dotProduct;
|
|
}
|
|
#endif /* LV_HAVE_AVX2 */
|
|
|
|
|
|
#ifdef LV_HAVE_AVX2
|
|
#include <immintrin.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_a_axv2(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
lv_16sc_t dotProduct = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
const unsigned int avx_iters = num_points / 8;
|
|
|
|
const lv_16sc_t* _in_a = in_a;
|
|
const lv_16sc_t* _in_b = in_b;
|
|
lv_16sc_t* _out = out;
|
|
unsigned int i;
|
|
unsigned int number;
|
|
|
|
if (avx_iters > 0)
|
|
{
|
|
__m256i a, b, c, c_sr, mask_imag, mask_real, real, imag, imag1, imag2, b_sl, a_sl, realcacc, imagcacc, result;
|
|
__VOLK_ATTR_ALIGNED(32)
|
|
lv_16sc_t dotProductVector[8];
|
|
|
|
realcacc = _mm256_setzero_si256();
|
|
imagcacc = _mm256_setzero_si256();
|
|
|
|
mask_imag = _mm256_set_epi8(0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0);
|
|
mask_real = _mm256_set_epi8(0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF, 0, 0, 0xFF, 0xFF);
|
|
|
|
for (number = 0; number < avx_iters; number++)
|
|
{
|
|
a = _mm256_load_si256((__m256i*)_in_a);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_a + 16);
|
|
b = _mm256_load_si256((__m256i*)_in_b);
|
|
__VOLK_GNSSSDR_PREFETCH(_in_b + 16);
|
|
c = _mm256_mullo_epi16(a, b);
|
|
|
|
c_sr = _mm256_srli_si256(c, 2); // Shift a right by imm8 bytes while shifting in zeros, and store the results in dst.
|
|
real = _mm256_subs_epi16(c, c_sr);
|
|
|
|
b_sl = _mm256_slli_si256(b, 2);
|
|
a_sl = _mm256_slli_si256(a, 2);
|
|
|
|
imag1 = _mm256_mullo_epi16(a, b_sl);
|
|
imag2 = _mm256_mullo_epi16(b, a_sl);
|
|
|
|
imag = _mm256_adds_epi16(imag1, imag2); // with saturation arithmetic!
|
|
|
|
realcacc = _mm256_adds_epi16(realcacc, real);
|
|
imagcacc = _mm256_adds_epi16(imagcacc, imag);
|
|
|
|
_in_a += 8;
|
|
_in_b += 8;
|
|
}
|
|
|
|
realcacc = _mm256_and_si256(realcacc, mask_real);
|
|
imagcacc = _mm256_and_si256(imagcacc, mask_imag);
|
|
|
|
result = _mm256_or_si256(realcacc, imagcacc);
|
|
|
|
_mm256_store_si256((__m256i*)dotProductVector, result); // Store the results back into the dot product vector
|
|
|
|
for (i = 0; i < 8; ++i)
|
|
{
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(dotProductVector[i])), sat_adds16i(lv_cimag(dotProduct), lv_cimag(dotProductVector[i])));
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < (num_points % 8); ++i)
|
|
{
|
|
lv_16sc_t tmp = (*_in_a++) * (*_in_b++);
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(tmp)), sat_adds16i(lv_cimag(dotProduct), lv_cimag(tmp)));
|
|
}
|
|
|
|
*_out = dotProduct;
|
|
}
|
|
#endif /* LV_HAVE_AVX2 */
|
|
|
|
|
|
#ifdef LV_HAVE_NEON
|
|
#include <arm_neon.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_neon(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
unsigned int quarter_points = num_points / 4;
|
|
unsigned int number;
|
|
|
|
lv_16sc_t* a_ptr = (lv_16sc_t*)in_a;
|
|
lv_16sc_t* b_ptr = (lv_16sc_t*)in_b;
|
|
*out = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
if (quarter_points > 0)
|
|
{
|
|
// for 2-lane vectors, 1st lane holds the real part,
|
|
// 2nd lane holds the imaginary part
|
|
int16x4x2_t a_val, b_val, c_val, accumulator;
|
|
int16x4x2_t tmp_real, tmp_imag;
|
|
__VOLK_ATTR_ALIGNED(16)
|
|
lv_16sc_t accum_result[4];
|
|
accumulator.val[0] = vdup_n_s16(0);
|
|
accumulator.val[1] = vdup_n_s16(0);
|
|
lv_16sc_t dotProduct = lv_cmake((int16_t)0, (int16_t)0);
|
|
|
|
for (number = 0; number < quarter_points; ++number)
|
|
{
|
|
a_val = vld2_s16((int16_t*)a_ptr); // a0r|a1r|a2r|a3r || a0i|a1i|a2i|a3i
|
|
b_val = vld2_s16((int16_t*)b_ptr); // b0r|b1r|b2r|b3r || b0i|b1i|b2i|b3i
|
|
__VOLK_GNSSSDR_PREFETCH(a_ptr + 8);
|
|
__VOLK_GNSSSDR_PREFETCH(b_ptr + 8);
|
|
|
|
// multiply the real*real and imag*imag to get real result
|
|
// a0r*b0r|a1r*b1r|a2r*b2r|a3r*b3r
|
|
tmp_real.val[0] = vmul_s16(a_val.val[0], b_val.val[0]);
|
|
// a0i*b0i|a1i*b1i|a2i*b2i|a3i*b3i
|
|
tmp_real.val[1] = vmul_s16(a_val.val[1], b_val.val[1]);
|
|
|
|
// Multiply cross terms to get the imaginary result
|
|
// a0r*b0i|a1r*b1i|a2r*b2i|a3r*b3i
|
|
tmp_imag.val[0] = vmul_s16(a_val.val[0], b_val.val[1]);
|
|
// a0i*b0r|a1i*b1r|a2i*b2r|a3i*b3r
|
|
tmp_imag.val[1] = vmul_s16(a_val.val[1], b_val.val[0]);
|
|
|
|
c_val.val[0] = vqsub_s16(tmp_real.val[0], tmp_real.val[1]);
|
|
c_val.val[1] = vqadd_s16(tmp_imag.val[0], tmp_imag.val[1]);
|
|
|
|
accumulator.val[0] = vqadd_s16(accumulator.val[0], c_val.val[0]);
|
|
accumulator.val[1] = vqadd_s16(accumulator.val[1], c_val.val[1]);
|
|
|
|
a_ptr += 4;
|
|
b_ptr += 4;
|
|
}
|
|
|
|
vst2_s16((int16_t*)accum_result, accumulator);
|
|
for (number = 0; number < 4; ++number)
|
|
{
|
|
dotProduct = lv_cmake(sat_adds16i(lv_creal(dotProduct), lv_creal(accum_result[number])), sat_adds16i(lv_cimag(dotProduct), lv_cimag(accum_result[number])));
|
|
}
|
|
|
|
*out = dotProduct;
|
|
}
|
|
|
|
// tail case
|
|
for (number = quarter_points * 4; number < num_points; ++number)
|
|
{
|
|
*out += (*a_ptr++) * (*b_ptr++);
|
|
}
|
|
}
|
|
|
|
#endif /* LV_HAVE_NEON */
|
|
|
|
|
|
#ifdef LV_HAVE_NEON
|
|
#include <arm_neon.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_neon_vma(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
unsigned int quarter_points = num_points / 4;
|
|
unsigned int number;
|
|
|
|
lv_16sc_t* a_ptr = (lv_16sc_t*)in_a;
|
|
lv_16sc_t* b_ptr = (lv_16sc_t*)in_b;
|
|
// for 2-lane vectors, 1st lane holds the real part,
|
|
// 2nd lane holds the imaginary part
|
|
int16x4x2_t a_val, b_val, accumulator;
|
|
int16x4x2_t tmp;
|
|
__VOLK_ATTR_ALIGNED(16)
|
|
lv_16sc_t accum_result[4];
|
|
accumulator.val[0] = vdup_n_s16(0);
|
|
accumulator.val[1] = vdup_n_s16(0);
|
|
|
|
for (number = 0; number < quarter_points; ++number)
|
|
{
|
|
a_val = vld2_s16((int16_t*)a_ptr); // a0r|a1r|a2r|a3r || a0i|a1i|a2i|a3i
|
|
b_val = vld2_s16((int16_t*)b_ptr); // b0r|b1r|b2r|b3r || b0i|b1i|b2i|b3i
|
|
__VOLK_GNSSSDR_PREFETCH(a_ptr + 8);
|
|
__VOLK_GNSSSDR_PREFETCH(b_ptr + 8);
|
|
|
|
tmp.val[0] = vmul_s16(a_val.val[0], b_val.val[0]);
|
|
tmp.val[1] = vmul_s16(a_val.val[1], b_val.val[0]);
|
|
|
|
// use multiply accumulate/subtract to get result
|
|
tmp.val[0] = vmls_s16(tmp.val[0], a_val.val[1], b_val.val[1]);
|
|
tmp.val[1] = vmla_s16(tmp.val[1], a_val.val[0], b_val.val[1]);
|
|
|
|
accumulator.val[0] = vqadd_s16(accumulator.val[0], tmp.val[0]);
|
|
accumulator.val[1] = vqadd_s16(accumulator.val[1], tmp.val[1]);
|
|
|
|
a_ptr += 4;
|
|
b_ptr += 4;
|
|
}
|
|
|
|
vst2_s16((int16_t*)accum_result, accumulator);
|
|
*out = accum_result[0] + accum_result[1] + accum_result[2] + accum_result[3];
|
|
|
|
// tail case
|
|
for (number = quarter_points * 4; number < num_points; ++number)
|
|
{
|
|
*out += (*a_ptr++) * (*b_ptr++);
|
|
}
|
|
}
|
|
|
|
#endif /* LV_HAVE_NEON */
|
|
|
|
|
|
#ifdef LV_HAVE_NEON
|
|
#include <arm_neon.h>
|
|
|
|
static inline void volk_gnsssdr_16ic_x2_dot_prod_16ic_neon_optvma(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
|
|
{
|
|
unsigned int quarter_points = num_points / 4;
|
|
unsigned int number;
|
|
|
|
lv_16sc_t* a_ptr = (lv_16sc_t*)in_a;
|
|
lv_16sc_t* b_ptr = (lv_16sc_t*)in_b;
|
|
// for 2-lane vectors, 1st lane holds the real part,
|
|
// 2nd lane holds the imaginary part
|
|
int16x4x2_t a_val, b_val, accumulator1, accumulator2;
|
|
|
|
__VOLK_ATTR_ALIGNED(16)
|
|
lv_16sc_t accum_result[4];
|
|
accumulator1.val[0] = vdup_n_s16(0);
|
|
accumulator1.val[1] = vdup_n_s16(0);
|
|
accumulator2.val[0] = vdup_n_s16(0);
|
|
accumulator2.val[1] = vdup_n_s16(0);
|
|
|
|
for (number = 0; number < quarter_points; ++number)
|
|
{
|
|
a_val = vld2_s16((int16_t*)a_ptr); // a0r|a1r|a2r|a3r || a0i|a1i|a2i|a3i
|
|
b_val = vld2_s16((int16_t*)b_ptr); // b0r|b1r|b2r|b3r || b0i|b1i|b2i|b3i
|
|
__VOLK_GNSSSDR_PREFETCH(a_ptr + 8);
|
|
__VOLK_GNSSSDR_PREFETCH(b_ptr + 8);
|
|
|
|
// use 2 accumulators to remove inter-instruction data dependencies
|
|
accumulator1.val[0] = vmla_s16(accumulator1.val[0], a_val.val[0], b_val.val[0]);
|
|
accumulator2.val[0] = vmls_s16(accumulator2.val[0], a_val.val[1], b_val.val[1]);
|
|
accumulator1.val[1] = vmla_s16(accumulator1.val[1], a_val.val[0], b_val.val[1]);
|
|
accumulator2.val[1] = vmla_s16(accumulator2.val[1], a_val.val[1], b_val.val[0]);
|
|
|
|
a_ptr += 4;
|
|
b_ptr += 4;
|
|
}
|
|
|
|
accumulator1.val[0] = vqadd_s16(accumulator1.val[0], accumulator2.val[0]);
|
|
accumulator1.val[1] = vqadd_s16(accumulator1.val[1], accumulator2.val[1]);
|
|
|
|
vst2_s16((int16_t*)accum_result, accumulator1);
|
|
*out = accum_result[0] + accum_result[1] + accum_result[2] + accum_result[3];
|
|
|
|
// tail case
|
|
for (number = quarter_points * 4; number < num_points; ++number)
|
|
{
|
|
*out += (*a_ptr++) * (*b_ptr++);
|
|
}
|
|
}
|
|
|
|
#endif /* LV_HAVE_NEON */
|
|
|
|
#endif /* INCLUDED_volk_gnsssdr_16ic_x2_dot_prod_16ic_H */
|