volk_32f_s32f_multiply_32f.h

Go to the documentation of this file.

/* -*- c++ -*- */
/*
 * Copyright 2012, 2014 Free Software Foundation, Inc.
 *
 * This file is part of VOLK
 *
 * SPDX-License-Identifier: LGPL-3.0-or-later
 */
 
#ifndef INCLUDED_volk_32f_s32f_multiply_32f_u_H
#define INCLUDED_volk_32f_s32f_multiply_32f_u_H
 
#include <inttypes.h>
#include <stdio.h>
 
#ifdef LV_HAVE_SSE
#include <xmmintrin.h>
 
static inline void volk_32f_s32f_multiply_32f_u_sse(float* cVector,
                                                    const float* aVector,
                                                    const float scalar,
                                                    unsigned int num_points)
{
    unsigned int number = 0;
    const unsigned int quarterPoints = num_points / 4;
 
    float* cPtr = cVector;
    const float* aPtr = aVector;
 
    __m128 aVal, bVal, cVal;
    bVal = _mm_set_ps1(scalar);
    for (; number < quarterPoints; number++) {
        aVal = _mm_loadu_ps(aPtr);
 
        cVal = _mm_mul_ps(aVal, bVal);
 
        _mm_storeu_ps(cPtr, cVal); // Store the results back into the C container
 
        aPtr += 4;
        cPtr += 4;
    }
 
    number = quarterPoints * 4;
    for (; number < num_points; number++) {
        *cPtr++ = (*aPtr++) * scalar;
    }
}
#endif /* LV_HAVE_SSE */
 
#ifdef LV_HAVE_AVX
#include <immintrin.h>
 
static inline void volk_32f_s32f_multiply_32f_u_avx(float* cVector,
                                                    const float* aVector,
                                                    const float scalar,
                                                    unsigned int num_points)
{
    unsigned int number = 0;
    const unsigned int eighthPoints = num_points / 8;
 
    float* cPtr = cVector;
    const float* aPtr = aVector;
 
    __m256 aVal, bVal, cVal;
    bVal = _mm256_set1_ps(scalar);
    for (; number < eighthPoints; number++) {
 
        aVal = _mm256_loadu_ps(aPtr);
 
        cVal = _mm256_mul_ps(aVal, bVal);
 
        _mm256_storeu_ps(cPtr, cVal); // Store the results back into the C container
 
        aPtr += 8;
        cPtr += 8;
    }
 
    number = eighthPoints * 8;
    for (; number < num_points; number++) {
        *cPtr++ = (*aPtr++) * scalar;
    }
}
#endif /* LV_HAVE_AVX */
 
#ifdef LV_HAVE_GENERIC
 
static inline void volk_32f_s32f_multiply_32f_generic(float* cVector,
                                                      const float* aVector,
                                                      const float scalar,
                                                      unsigned int num_points)
{
    unsigned int number = 0;
    const float* inputPtr = aVector;
    float* outputPtr = cVector;
    for (number = 0; number < num_points; number++) {
        *outputPtr = (*inputPtr) * scalar;
        inputPtr++;
        outputPtr++;
    }
}
#endif /* LV_HAVE_GENERIC */
 
#endif /* INCLUDED_volk_32f_s32f_multiply_32f_u_H */
 
 
#ifndef INCLUDED_volk_32f_s32f_multiply_32f_a_H
#define INCLUDED_volk_32f_s32f_multiply_32f_a_H
 
#include <inttypes.h>
#include <stdio.h>
 
#ifdef LV_HAVE_SSE
#include <xmmintrin.h>
 
static inline void volk_32f_s32f_multiply_32f_a_sse(float* cVector,
                                                    const float* aVector,
                                                    const float scalar,
                                                    unsigned int num_points)
{
    unsigned int number = 0;
    const unsigned int quarterPoints = num_points / 4;
 
    float* cPtr = cVector;
    const float* aPtr = aVector;
 
    __m128 aVal, bVal, cVal;
    bVal = _mm_set_ps1(scalar);
    for (; number < quarterPoints; number++) {
        aVal = _mm_load_ps(aPtr);
 
        cVal = _mm_mul_ps(aVal, bVal);
 
        _mm_store_ps(cPtr, cVal); // Store the results back into the C container
 
        aPtr += 4;
        cPtr += 4;
    }
 
    number = quarterPoints * 4;
    for (; number < num_points; number++) {
        *cPtr++ = (*aPtr++) * scalar;
    }
}
#endif /* LV_HAVE_SSE */
 
#ifdef LV_HAVE_AVX
#include <immintrin.h>
 
static inline void volk_32f_s32f_multiply_32f_a_avx(float* cVector,
                                                    const float* aVector,
                                                    const float scalar,
                                                    unsigned int num_points)
{
    unsigned int number = 0;
    const unsigned int eighthPoints = num_points / 8;
 
    float* cPtr = cVector;
    const float* aPtr = aVector;
 
    __m256 aVal, bVal, cVal;
    bVal = _mm256_set1_ps(scalar);
    for (; number < eighthPoints; number++) {
        aVal = _mm256_load_ps(aPtr);
 
        cVal = _mm256_mul_ps(aVal, bVal);
 
        _mm256_store_ps(cPtr, cVal); // Store the results back into the C container
 
        aPtr += 8;
        cPtr += 8;
    }
 
    number = eighthPoints * 8;
    for (; number < num_points; number++) {
        *cPtr++ = (*aPtr++) * scalar;
    }
}
#endif /* LV_HAVE_AVX */
 
#ifdef LV_HAVE_NEON
#include <arm_neon.h>
 
static inline void volk_32f_s32f_multiply_32f_u_neon(float* cVector,
                                                     const float* aVector,
                                                     const float scalar,
                                                     unsigned int num_points)
{
    unsigned int number = 0;
    const float* inputPtr = aVector;
    float* outputPtr = cVector;
    const unsigned int quarterPoints = num_points / 4;
 
    float32x4_t aVal, cVal;
 
    for (number = 0; number < quarterPoints; number++) {
        aVal = vld1q_f32(inputPtr);       // Load into NEON regs
        cVal = vmulq_n_f32(aVal, scalar); // Do the multiply
        vst1q_f32(outputPtr, cVal);       // Store results back to output
        inputPtr += 4;
        outputPtr += 4;
    }
    for (number = quarterPoints * 4; number < num_points; number++) {
        *outputPtr++ = (*inputPtr++) * scalar;
    }
}
#endif /* LV_HAVE_NEON */
 
 
#ifdef LV_HAVE_GENERIC
 
static inline void volk_32f_s32f_multiply_32f_a_generic(float* cVector,
                                                        const float* aVector,
                                                        const float scalar,
                                                        unsigned int num_points)
{
    unsigned int number = 0;
    const float* inputPtr = aVector;
    float* outputPtr = cVector;
    for (number = 0; number < num_points; number++) {
        *outputPtr = (*inputPtr) * scalar;
        inputPtr++;
        outputPtr++;
    }
}
#endif /* LV_HAVE_GENERIC */
 
 
#ifdef LV_HAVE_ORC
 
extern void volk_32f_s32f_multiply_32f_a_orc_impl(float* dst,
                                                  const float* src,
                                                  const float scalar,
                                                  unsigned int num_points);
 
static inline void volk_32f_s32f_multiply_32f_u_orc(float* cVector,
                                                    const float* aVector,
                                                    const float scalar,
                                                    unsigned int num_points)
{
    volk_32f_s32f_multiply_32f_a_orc_impl(cVector, aVector, scalar, num_points);
}
 
#endif /* LV_HAVE_GENERIC */
 
#endif /* INCLUDED_volk_32f_s32f_multiply_32f_a_H */