From 9c59a184dba820975e5da6fcd5d248aee87f7e2f Mon Sep 17 00:00:00 2001 From: Clyne Sullivan Date: Sun, 2 Feb 2025 11:26:53 -0500 Subject: add l476 implementation --- .../DSP/Source/MatrixFunctions/arm_mat_add_f32.c | 536 +++++++++------------ 1 file changed, 232 insertions(+), 304 deletions(-) (limited to 'Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_add_f32.c') diff --git a/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_add_f32.c b/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_add_f32.c index 879c8f8..7f9ec08 100644 --- a/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_add_f32.c +++ b/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_add_f32.c @@ -1,304 +1,232 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_mat_add_f32.c - * Description: Floating-point matrix addition - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 ARM Limited or its affiliates. All rights reserved. - * - * SPDX-License-Identifier: Apache-2.0 - * - * Licensed under the Apache License, Version 2.0 (the License); you may - * not use this file except in compliance with the License. - * You may obtain a copy of the License at - * - * www.apache.org/licenses/LICENSE-2.0 - * - * Unless required by applicable law or agreed to in writing, software - * distributed under the License is distributed on an AS IS BASIS, WITHOUT - * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. - * See the License for the specific language governing permissions and - * limitations under the License. - */ - -#include "dsp/matrix_functions.h" - -/** - @ingroup groupMatrix - */ - -/** - @defgroup MatrixAdd Matrix Addition - - Adds two matrices. - \image html MatrixAddition.gif "Addition of two 3 x 3 matrices" - - The functions check to make sure that - pSrcA, pSrcB, and pDst have the same - number of rows and columns. - */ - -/** - @addtogroup MatrixAdd - @{ - */ - - -/** - @brief Floating-point matrix addition. - @param[in] pSrcA points to first input matrix structure - @param[in] pSrcB points to second input matrix structure - @param[out] pDst points to output matrix structure - @return execution status - - \ref ARM_MATH_SUCCESS : Operation successful - - \ref ARM_MATH_SIZE_MISMATCH : Matrix size check failed - */ - -#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) -arm_status arm_mat_add_f32( - const arm_matrix_instance_f32 * pSrcA, - const arm_matrix_instance_f32 * pSrcB, - arm_matrix_instance_f32 * pDst) -{ - arm_status status; - uint32_t numSamples; /* total number of elements in the matrix */ - float32_t *pDataA, *pDataB, *pDataDst; - f32x4_t vecA, vecB, vecDst; - float32_t const *pSrcAVec; - float32_t const *pSrcBVec; - uint32_t blkCnt; /* loop counters */ - - pDataA = pSrcA->pData; - pDataB = pSrcB->pData; - pDataDst = pDst->pData; - pSrcAVec = (float32_t const *) pDataA; - pSrcBVec = (float32_t const *) pDataB; - -#ifdef ARM_MATH_MATRIX_CHECK - /* Check for matrix mismatch condition */ - if ((pSrcA->numRows != pSrcB->numRows) || - (pSrcA->numCols != pSrcB->numCols) || - (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols)) - { - /* Set status as ARM_MATH_SIZE_MISMATCH */ - status = ARM_MATH_SIZE_MISMATCH; - } - else -#endif - { - /* - * Total number of samples in the input matrix - */ - numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; - blkCnt = numSamples >> 2; - while (blkCnt > 0U) - { - /* C(m,n) = A(m,n) + B(m,n) */ - /* Add and then store the results in the destination buffer. */ - vecA = vld1q(pSrcAVec); - pSrcAVec += 4; - vecB = vld1q(pSrcBVec); - pSrcBVec += 4; - vecDst = vaddq(vecA, vecB); - vst1q(pDataDst, vecDst); - pDataDst += 4; - /* - * Decrement the blockSize loop counter - */ - blkCnt--; - } - /* - * tail - */ - blkCnt = numSamples & 3; - if (blkCnt > 0U) - { - mve_pred16_t p0 = vctp32q(blkCnt); - vecA = vld1q(pSrcAVec); - vecB = vld1q(pSrcBVec); - vecDst = vaddq_m(vecDst, vecA, vecB, p0); - vstrwq_p(pDataDst, vecDst, p0); - } - /* set status as ARM_MATH_SUCCESS */ - status = ARM_MATH_SUCCESS; - } - return (status); -} -#else -#if defined(ARM_MATH_NEON) -/* - -Neon version is assuming the matrix is small enough. -So no blocking is used for taking into account cache effects. -For big matrix, there exist better libraries for Neon. - -*/ -arm_status arm_mat_add_f32( - const arm_matrix_instance_f32 * pSrcA, - const arm_matrix_instance_f32 * pSrcB, - arm_matrix_instance_f32 * pDst) -{ - float32_t *pIn1 = pSrcA->pData; /* input data matrix pointer A */ - float32_t *pIn2 = pSrcB->pData; /* input data matrix pointer B */ - float32_t *pOut = pDst->pData; /* output data matrix pointer */ - - - uint32_t numSamples; /* total number of elements in the matrix */ - uint32_t blkCnt; /* loop counters */ - arm_status status; /* status of matrix addition */ - -#ifdef ARM_MATH_MATRIX_CHECK - /* Check for matrix mismatch condition */ - if ((pSrcA->numRows != pSrcB->numRows) || - (pSrcA->numCols != pSrcB->numCols) || - (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols)) - { - /* Set status as ARM_MATH_SIZE_MISMATCH */ - status = ARM_MATH_SIZE_MISMATCH; - } - else -#endif - { - float32x4_t vec1; - float32x4_t vec2; - float32x4_t res; - - /* Total number of samples in the input matrix */ - numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; - - blkCnt = numSamples >> 2U; - - /* Compute 4 outputs at a time. - ** a second loop below computes the remaining 1 to 3 samples. */ - while (blkCnt > 0U) - { - /* C(m,n) = A(m,n) + B(m,n) */ - /* Add and then store the results in the destination buffer. */ - vec1 = vld1q_f32(pIn1); - vec2 = vld1q_f32(pIn2); - res = vaddq_f32(vec1, vec2); - vst1q_f32(pOut, res); - - /* update pointers to process next samples */ - pIn1 += 4U; - pIn2 += 4U; - pOut += 4U; - /* Decrement the loop counter */ - blkCnt--; - } - - /* If the numSamples is not a multiple of 4, compute any remaining output samples here. - ** No loop unrolling is used. */ - blkCnt = numSamples % 0x4U; - - while (blkCnt > 0U) - { - /* C(m,n) = A(m,n) + B(m,n) */ - /* Add and then store the results in the destination buffer. */ - *pOut++ = (*pIn1++) + (*pIn2++); - - /* Decrement the loop counter */ - blkCnt--; - } - - /* set status as ARM_MATH_SUCCESS */ - status = ARM_MATH_SUCCESS; - } - - /* Return to application */ - return (status); -} -#else -arm_status arm_mat_add_f32( - const arm_matrix_instance_f32 * pSrcA, - const arm_matrix_instance_f32 * pSrcB, - arm_matrix_instance_f32 * pDst) -{ - float32_t *pInA = pSrcA->pData; /* input data matrix pointer A */ - float32_t *pInB = pSrcB->pData; /* input data matrix pointer B */ - float32_t *pOut = pDst->pData; /* output data matrix pointer */ - - uint32_t numSamples; /* total number of elements in the matrix */ - uint32_t blkCnt; /* loop counters */ - arm_status status; /* status of matrix addition */ - -#ifdef ARM_MATH_MATRIX_CHECK - - /* Check for matrix mismatch condition */ - if ((pSrcA->numRows != pSrcB->numRows) || - (pSrcA->numCols != pSrcB->numCols) || - (pSrcA->numRows != pDst->numRows) || - (pSrcA->numCols != pDst->numCols) ) - { - /* Set status as ARM_MATH_SIZE_MISMATCH */ - status = ARM_MATH_SIZE_MISMATCH; - } - else - -#endif /* #ifdef ARM_MATH_MATRIX_CHECK */ - - { - /* Total number of samples in input matrix */ - numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; - -#if defined (ARM_MATH_LOOPUNROLL) - - /* Loop unrolling: Compute 4 outputs at a time */ - blkCnt = numSamples >> 2U; - - while (blkCnt > 0U) - { - /* C(m,n) = A(m,n) + B(m,n) */ - - /* Add and store result in destination buffer. */ - *pOut++ = *pInA++ + *pInB++; - - *pOut++ = *pInA++ + *pInB++; - - *pOut++ = *pInA++ + *pInB++; - - *pOut++ = *pInA++ + *pInB++; - - /* Decrement loop counter */ - blkCnt--; - } - - /* Loop unrolling: Compute remaining outputs */ - blkCnt = numSamples % 0x4U; - -#else - - /* Initialize blkCnt with number of samples */ - blkCnt = numSamples; - -#endif /* #if defined (ARM_MATH_LOOPUNROLL) */ - - while (blkCnt > 0U) - { - /* C(m,n) = A(m,n) + B(m,n) */ - - /* Add and store result in destination buffer. */ - *pOut++ = *pInA++ + *pInB++; - - /* Decrement loop counter */ - blkCnt--; - } - - /* Set status as ARM_MATH_SUCCESS */ - status = ARM_MATH_SUCCESS; - } - - /* Return to application */ - return (status); -} -#endif /* #if defined(ARM_MATH_NEON) */ -#endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */ - -/** - @} end of MatrixAdd group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_mat_add_f32.c + * Description: Floating-point matrix addition + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 ARM Limited or its affiliates. All rights reserved. + * + * SPDX-License-Identifier: Apache-2.0 + * + * Licensed under the Apache License, Version 2.0 (the License); you may + * not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an AS IS BASIS, WITHOUT + * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "arm_math.h" + +/** + @ingroup groupMatrix + */ + +/** + @defgroup MatrixAdd Matrix Addition + + Adds two matrices. + \image html MatrixAddition.gif "Addition of two 3 x 3 matrices" + + The functions check to make sure that + pSrcA, pSrcB, and pDst have the same + number of rows and columns. + */ + +/** + @addtogroup MatrixAdd + @{ + */ + + +/** + @brief Floating-point matrix addition. + @param[in] pSrcA points to first input matrix structure + @param[in] pSrcB points to second input matrix structure + @param[out] pDst points to output matrix structure + @return execution status + - \ref ARM_MATH_SUCCESS : Operation successful + - \ref ARM_MATH_SIZE_MISMATCH : Matrix size check failed + */ +#if defined(ARM_MATH_NEON) +/* + +Neon version is assuming the matrix is small enough. +So no blocking is used for taking into account cache effects. +For big matrix, there exist better libraries for Neon. + +*/ +arm_status arm_mat_add_f32( + const arm_matrix_instance_f32 * pSrcA, + const arm_matrix_instance_f32 * pSrcB, + arm_matrix_instance_f32 * pDst) +{ + float32_t *pIn1 = pSrcA->pData; /* input data matrix pointer A */ + float32_t *pIn2 = pSrcB->pData; /* input data matrix pointer B */ + float32_t *pOut = pDst->pData; /* output data matrix pointer */ + + float32_t inA1, inA2, inB1, inB2, out1, out2; /* temporary variables */ + + uint32_t numSamples; /* total number of elements in the matrix */ + uint32_t blkCnt; /* loop counters */ + arm_status status; /* status of matrix addition */ + +#ifdef ARM_MATH_MATRIX_CHECK + /* Check for matrix mismatch condition */ + if ((pSrcA->numRows != pSrcB->numRows) || + (pSrcA->numCols != pSrcB->numCols) || + (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols)) + { + /* Set status as ARM_MATH_SIZE_MISMATCH */ + status = ARM_MATH_SIZE_MISMATCH; + } + else +#endif + { + float32x4_t vec1; + float32x4_t vec2; + float32x4_t res; + + /* Total number of samples in the input matrix */ + numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; + + blkCnt = numSamples >> 2U; + + /* Compute 4 outputs at a time. + ** a second loop below computes the remaining 1 to 3 samples. */ + while (blkCnt > 0U) + { + /* C(m,n) = A(m,n) + B(m,n) */ + /* Add and then store the results in the destination buffer. */ + vec1 = vld1q_f32(pIn1); + vec2 = vld1q_f32(pIn2); + res = vaddq_f32(vec1, vec2); + vst1q_f32(pOut, res); + + /* update pointers to process next samples */ + pIn1 += 4U; + pIn2 += 4U; + pOut += 4U; + /* Decrement the loop counter */ + blkCnt--; + } + + /* If the numSamples is not a multiple of 4, compute any remaining output samples here. + ** No loop unrolling is used. */ + blkCnt = numSamples % 0x4U; + + while (blkCnt > 0U) + { + /* C(m,n) = A(m,n) + B(m,n) */ + /* Add and then store the results in the destination buffer. */ + *pOut++ = (*pIn1++) + (*pIn2++); + + /* Decrement the loop counter */ + blkCnt--; + } + + /* set status as ARM_MATH_SUCCESS */ + status = ARM_MATH_SUCCESS; + } + + /* Return to application */ + return (status); +} +#else +arm_status arm_mat_add_f32( + const arm_matrix_instance_f32 * pSrcA, + const arm_matrix_instance_f32 * pSrcB, + arm_matrix_instance_f32 * pDst) +{ + float32_t *pInA = pSrcA->pData; /* input data matrix pointer A */ + float32_t *pInB = pSrcB->pData; /* input data matrix pointer B */ + float32_t *pOut = pDst->pData; /* output data matrix pointer */ + + uint32_t numSamples; /* total number of elements in the matrix */ + uint32_t blkCnt; /* loop counters */ + arm_status status; /* status of matrix addition */ + +#ifdef ARM_MATH_MATRIX_CHECK + + /* Check for matrix mismatch condition */ + if ((pSrcA->numRows != pSrcB->numRows) || + (pSrcA->numCols != pSrcB->numCols) || + (pSrcA->numRows != pDst->numRows) || + (pSrcA->numCols != pDst->numCols) ) + { + /* Set status as ARM_MATH_SIZE_MISMATCH */ + status = ARM_MATH_SIZE_MISMATCH; + } + else + +#endif /* #ifdef ARM_MATH_MATRIX_CHECK */ + + { + /* Total number of samples in input matrix */ + numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; + +#if defined (ARM_MATH_LOOPUNROLL) + + /* Loop unrolling: Compute 4 outputs at a time */ + blkCnt = numSamples >> 2U; + + while (blkCnt > 0U) + { + /* C(m,n) = A(m,n) + B(m,n) */ + + /* Add and store result in destination buffer. */ + *pOut++ = *pInA++ + *pInB++; + + *pOut++ = *pInA++ + *pInB++; + + *pOut++ = *pInA++ + *pInB++; + + *pOut++ = *pInA++ + *pInB++; + + /* Decrement loop counter */ + blkCnt--; + } + + /* Loop unrolling: Compute remaining outputs */ + blkCnt = numSamples % 0x4U; + +#else + + /* Initialize blkCnt with number of samples */ + blkCnt = numSamples; + +#endif /* #if defined (ARM_MATH_LOOPUNROLL) */ + + while (blkCnt > 0U) + { + /* C(m,n) = A(m,n) + B(m,n) */ + + /* Add and store result in destination buffer. */ + *pOut++ = *pInA++ + *pInB++; + + /* Decrement loop counter */ + blkCnt--; + } + + /* Set status as ARM_MATH_SUCCESS */ + status = ARM_MATH_SUCCESS; + } + + /* Return to application */ + return (status); +} +#endif /* #if defined(ARM_MATH_NEON) */ + +/** + @} end of MatrixAdd group + */ -- cgit v1.2.3