diff options
Diffstat (limited to 'Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c')
| -rw-r--r-- | Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c | 132 |
1 files changed, 89 insertions, 43 deletions
diff --git a/Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c b/Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c index 1426735b9..a4825a531 100644 --- a/Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c +++ b/Drivers/CMSIS/DSP/Source/StatisticsFunctions/arm_power_f32.c @@ -3,13 +3,13 @@ * Title: arm_power_f32.c * Description: Sum of the squares of the elements of a floating-point vector * - * $Date: 27. January 2017 - * $Revision: V.1.5.1 + * $Date: 18. March 2019 + * $Revision: V1.6.0 * * Target Processor: Cortex-M cores * -------------------------------------------------------------------- */ /* - * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved. + * Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved. * * SPDX-License-Identifier: Apache-2.0 * @@ -29,40 +29,37 @@ #include "arm_math.h" /** - * @ingroup groupStats + @ingroup groupStats */ /** - * @defgroup power Power - * - * Calculates the sum of the squares of the elements in the input vector. - * The underlying algorithm is used: - * - * <pre> - * Result = pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + pSrc[2] * pSrc[2] + ... + pSrc[blockSize-1] * pSrc[blockSize-1]; - * </pre> - * - * There are separate functions for floating point, Q31, Q15, and Q7 data types. + @defgroup power Power + + Calculates the sum of the squares of the elements in the input vector. + The underlying algorithm is used: + + <pre> + Result = pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + pSrc[2] * pSrc[2] + ... + pSrc[blockSize-1] * pSrc[blockSize-1]; + </pre> + + There are separate functions for floating point, Q31, Q15, and Q7 data types. */ /** - * @addtogroup power - * @{ + @addtogroup power + @{ */ - /** - * @brief Sum of the squares of the elements of a floating-point vector. - * @param[in] *pSrc points to the input vector - * @param[in] blockSize length of the input vector - * @param[out] *pResult sum of the squares value returned here - * @return none. - * + @brief Sum of the squares of the elements of a floating-point vector. + @param[in] pSrc points to the input vector + @param[in] blockSize number of samples in input vector + @param[out] pResult sum of the squares value returned here + @return none */ - - +#if defined(ARM_MATH_NEON) void arm_power_f32( - float32_t * pSrc, + const float32_t * pSrc, uint32_t blockSize, float32_t * pResult) { @@ -70,60 +67,109 @@ void arm_power_f32( float32_t in; /* Temporary variable to store input value */ uint32_t blkCnt; /* loop counter */ -#if defined (ARM_MATH_DSP) - /* Run the below code for Cortex-M4 and Cortex-M3 */ + float32x4_t sumV = vdupq_n_f32(0.0f); /* Temporary result storage */ + float32x2_t sumV2; + float32x4_t inV; - /*loop Unrolling */ blkCnt = blockSize >> 2U; - /* First part of the processing with loop unrolling. Compute 4 outputs at a time. + /* Compute 4 outputs at a time. ** a second loop below computes the remaining 1 to 3 samples. */ while (blkCnt > 0U) { /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ /* Compute Power and then store the result in a temporary variable, sum. */ + inV = vld1q_f32(pSrc); + sumV = vmlaq_f32(sumV, inV, inV); + pSrc += 4; + + /* Decrement the loop counter */ + blkCnt--; + } + sumV2 = vpadd_f32(vget_low_f32(sumV),vget_high_f32(sumV)); + sum = sumV2[0] + sumV2[1]; + + /* If the blockSize is not a multiple of 4, compute any remaining output samples here. + ** No loop unrolling is used. */ + blkCnt = blockSize % 0x4U; + + while (blkCnt > 0U) + { + /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ + /* compute power and then store the result in a temporary variable, sum. */ + in = *pSrc++; + sum += in * in; + + /* Decrement the loop counter */ + blkCnt--; + } + + /* Store the result to the destination */ + *pResult = sum; +} +#else +void arm_power_f32( + const float32_t * pSrc, + uint32_t blockSize, + float32_t * pResult) +{ + uint32_t blkCnt; /* Loop counter */ + float32_t sum = 0.0f; /* Temporary result storage */ + float32_t in; /* Temporary variable to store input value */ + +#if defined (ARM_MATH_LOOPUNROLL) + + /* Loop unrolling: Compute 4 outputs at a time */ + blkCnt = blockSize >> 2U; + + while (blkCnt > 0U) + { + /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */ + + /* Compute Power and store result in a temporary variable, sum. */ in = *pSrc++; sum += in * in; + in = *pSrc++; sum += in * in; + in = *pSrc++; sum += in * in; + in = *pSrc++; sum += in * in; - /* Decrement the loop counter */ + /* Decrement loop counter */ blkCnt--; } - /* If the blockSize is not a multiple of 4, compute any remaining output samples here. - ** No loop unrolling is used. */ + /* Loop unrolling: Compute remaining outputs */ blkCnt = blockSize % 0x4U; - #else - /* Run the below code for Cortex-M0 */ - /* Loop over blockSize number of values */ + /* Initialize blkCnt with number of samples */ blkCnt = blockSize; -#endif /* #if defined (ARM_MATH_DSP) */ - +#endif /* #if defined (ARM_MATH_LOOPUNROLL) */ while (blkCnt > 0U) { - /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ - /* compute power and then store the result in a temporary variable, sum. */ + /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */ + + /* Compute Power and store result in a temporary variable, sum. */ in = *pSrc++; sum += in * in; - /* Decrement the loop counter */ + /* Decrement loop counter */ blkCnt--; } - /* Store the result to the destination */ + /* Store result to destination */ *pResult = sum; } +#endif /* #if defined(ARM_MATH_NEON) */ /** - * @} end of power group + @} end of power group */ |