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_scale_f32.c | 508 +++++++++------------ 1 file changed, 221 insertions(+), 287 deletions(-) (limited to 'Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_scale_f32.c') diff --git a/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_scale_f32.c b/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_scale_f32.c index daebbb3..91c56b1 100644 --- a/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_scale_f32.c +++ b/Drivers/CMSIS/DSP/Source/MatrixFunctions/arm_mat_scale_f32.c @@ -1,287 +1,221 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_mat_scale_f32.c - * Description: Multiplies a floating-point matrix by a scalar - * - * $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 MatrixScale Matrix Scale - - Multiplies a matrix by a scalar. This is accomplished by multiplying each element in the - matrix by the scalar. For example: - \image html MatrixScale.gif "Matrix Scaling of a 3 x 3 matrix" - - The function checks to make sure that the input and output matrices are of the same size. - - In the fixed-point Q15 and Q31 functions, scale is represented by - a fractional multiplication scaleFract and an arithmetic shift shift. - The shift allows the gain of the scaling operation to exceed 1.0. - The overall scale factor applied to the fixed-point data is - 
-      scale = scaleFract * 2^shift.
-
- */ - -/** - @addtogroup MatrixScale - @{ - */ - -/** - @brief Floating-point matrix scaling. - @param[in] pSrc points to input matrix - @param[in] scale scale factor to be applied - @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_scale_f32( - const arm_matrix_instance_f32 * pSrc, - float32_t scale, - arm_matrix_instance_f32 * pDst) -{ - arm_status status; /* status of matrix scaling */ - #ifdef ARM_MATH_MATRIX_CHECK - /* Check for matrix mismatch condition */ - if ((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) - { - /* Set status as ARM_MATH_SIZE_MISMATCH */ - status = ARM_MATH_SIZE_MISMATCH; - } - else -#endif /* #ifdef ARM_MATH_MATRIX_CHECK */ - { - float32_t *pIn = pSrc->pData; /* input data matrix pointer */ - float32_t *pOut = pDst->pData; /* output data matrix pointer */ - uint32_t numSamples; /* total number of elements in the matrix */ - uint32_t blkCnt; /* loop counters */ - f32x4_t vecIn, vecOut; - float32_t const *pInVec; - - pInVec = (float32_t const *) pIn; - /* - * Total number of samples in the input matrix - */ - numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; - blkCnt = numSamples >> 2; - while (blkCnt > 0U) - { - /* - * C(m,n) = A(m,n) * scale - * Scaling and results are stored in the destination buffer. - */ - vecIn = vld1q(pInVec); - pInVec += 4; - - vecOut = vecIn * scale; - - vst1q(pOut, vecOut); - pOut += 4; - /* - * Decrement the blockSize loop counter - */ - blkCnt--; - } - /* - * tail - */ - blkCnt = numSamples & 3; - if (blkCnt > 0U) - { - mve_pred16_t p0 = vctp32q(blkCnt); - vecIn = vld1q(pInVec); - vecOut = vecIn * scale; - - vstrwq_p(pOut, vecOut, p0); - } - /* Set status as ARM_MATH_SUCCESS */ - status = ARM_MATH_SUCCESS; - } - - /* Return to application */ - return (status); - -} -#else -#if defined(ARM_MATH_NEON_EXPERIMENTAL) -arm_status arm_mat_scale_f32( - const arm_matrix_instance_f32 * pSrc, - float32_t scale, - arm_matrix_instance_f32 * pDst) -{ - float32_t *pIn = pSrc->pData; /* input data matrix pointer */ - 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 scaling */ - - -#ifdef ARM_MATH_MATRIX_CHECK - /* Check for matrix mismatch condition */ - if ((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) - { - /* Set status as ARM_MATH_SIZE_MISMATCH */ - status = ARM_MATH_SIZE_MISMATCH; - } - else -#endif /* #ifdef ARM_MATH_MATRIX_CHECK */ - { - float32x4_t vec1; - float32x4_t res; - - /* Total number of samples in the input matrix */ - numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; - - blkCnt = numSamples >> 2; - - /* 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) * scale */ - /* Scaling and results are stored in the destination buffer. */ - vec1 = vld1q_f32(pIn); - res = vmulq_f32(vec1, vdupq_n_f32(scale)); - vst1q_f32(pOut, res); - - /* update pointers to process next sampels */ - pIn += 4U; - pOut += 4U; - - /* Decrement the numSamples 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) * scale */ - /* The results are stored in the destination buffer. */ - *pOut++ = (*pIn++) * scale; - - /* 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_scale_f32( - const arm_matrix_instance_f32 * pSrc, - float32_t scale, - arm_matrix_instance_f32 * pDst) -{ - float32_t *pIn = pSrc->pData; /* Input data matrix pointer */ - 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 scaling */ - -#ifdef ARM_MATH_MATRIX_CHECK - - /* Check for matrix mismatch condition */ - if ((pSrc->numRows != pDst->numRows) || - (pSrc->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) pSrc->numRows * pSrc->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) * scale */ - - /* Scale and store result in destination buffer. */ - *pOut++ = (*pIn++) * scale; - *pOut++ = (*pIn++) * scale; - *pOut++ = (*pIn++) * scale; - *pOut++ = (*pIn++) * scale; - - /* 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) * scale */ - - /* Scale and store result in destination buffer. */ - *pOut++ = (*pIn++) * scale; - - /* 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 MatrixScale group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_mat_scale_f32.c + * Description: Multiplies a floating-point matrix by a scalar + * + * $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 MatrixScale Matrix Scale + + Multiplies a matrix by a scalar. This is accomplished by multiplying each element in the + matrix by the scalar. For example: + \image html MatrixScale.gif "Matrix Scaling of a 3 x 3 matrix" + + The function checks to make sure that the input and output matrices are of the same size. + + In the fixed-point Q15 and Q31 functions, scale is represented by + a fractional multiplication scaleFract and an arithmetic shift shift. + The shift allows the gain of the scaling operation to exceed 1.0. + The overall scale factor applied to the fixed-point data is + 
+      scale = scaleFract * 2^shift.
+
+ */ + +/** + @addtogroup MatrixScale + @{ + */ + +/** + @brief Floating-point matrix scaling. + @param[in] pSrc points to input matrix + @param[in] scale scale factor to be applied + @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_EXPERIMENTAL) +arm_status arm_mat_scale_f32( + const arm_matrix_instance_f32 * pSrc, + float32_t scale, + arm_matrix_instance_f32 * pDst) +{ + float32_t *pIn = pSrc->pData; /* input data matrix pointer */ + 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 scaling */ + + + float32_t in1, in2, in3, in4; /* temporary variables */ + float32_t out1, out2, out3, out4; /* temporary variables */ + + +#ifdef ARM_MATH_MATRIX_CHECK + /* Check for matrix mismatch condition */ + if ((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols)) + { + /* Set status as ARM_MATH_SIZE_MISMATCH */ + status = ARM_MATH_SIZE_MISMATCH; + } + else +#endif /* #ifdef ARM_MATH_MATRIX_CHECK */ + { + float32x4_t vec1; + float32x4_t res; + + /* Total number of samples in the input matrix */ + numSamples = (uint32_t) pSrc->numRows * pSrc->numCols; + + blkCnt = numSamples >> 2; + + /* 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) * scale */ + /* Scaling and results are stored in the destination buffer. */ + vec1 = vld1q_f32(pIn); + res = vmulq_f32(vec1, vdupq_n_f32(scale)); + vst1q_f32(pOut, res); + + /* update pointers to process next sampels */ + pIn += 4U; + pOut += 4U; + + /* Decrement the numSamples 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) * scale */ + /* The results are stored in the destination buffer. */ + *pOut++ = (*pIn++) * scale; + + /* 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_scale_f32( + const arm_matrix_instance_f32 * pSrc, + float32_t scale, + arm_matrix_instance_f32 * pDst) +{ + float32_t *pIn = pSrc->pData; /* Input data matrix pointer */ + 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 scaling */ + +#ifdef ARM_MATH_MATRIX_CHECK + + /* Check for matrix mismatch condition */ + if ((pSrc->numRows != pDst->numRows) || + (pSrc->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) pSrc->numRows * pSrc->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) * scale */ + + /* Scale and store result in destination buffer. */ + *pOut++ = (*pIn++) * scale; + *pOut++ = (*pIn++) * scale; + *pOut++ = (*pIn++) * scale; + *pOut++ = (*pIn++) * scale; + + /* 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) * scale */ + + /* Scale and store result in destination buffer. */ + *pOut++ = (*pIn++) * scale; + + /* 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 MatrixScale group + */ -- cgit v1.2.3