File size: 5,777 Bytes
25ade36 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 | #include "edge-impulse-sdk/dsp/config.hpp"
#if EIDSP_LOAD_CMSIS_DSP_SOURCES
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_mat_mult_f64.c
* Description: Floating-point matrix multiplication
*
* $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 "edge-impulse-sdk/CMSIS/DSP/Include/dsp/matrix_functions.h"
/**
* @ingroup groupMatrix
*/
/**
* @addtogroup MatrixMult
* @{
*/
/**
* @brief Floating-point matrix multiplication.
* @param[in] *pSrcA points to the first input matrix structure
* @param[in] *pSrcB points to the second input matrix structure
* @param[out] *pDst points to output matrix structure
* @return The function returns either
* <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
*/
arm_status arm_mat_mult_f64(
const arm_matrix_instance_f64 * pSrcA,
const arm_matrix_instance_f64 * pSrcB,
arm_matrix_instance_f64 * pDst)
{
float64_t *pIn1 = pSrcA->pData; /* Input data matrix pointer A */
float64_t *pIn2 = pSrcB->pData; /* Input data matrix pointer B */
float64_t *pInA = pSrcA->pData; /* Input data matrix pointer A */
float64_t *pInB = pSrcB->pData; /* Input data matrix pointer B */
float64_t *pOut = pDst->pData; /* Output data matrix pointer */
float64_t *px; /* Temporary output data matrix pointer */
float64_t sum; /* Accumulator */
uint16_t numRowsA = pSrcA->numRows; /* Number of rows of input matrix A */
uint16_t numColsB = pSrcB->numCols; /* Number of columns of input matrix B */
uint16_t numColsA = pSrcA->numCols; /* Number of columns of input matrix A */
uint64_t col, i = 0U, row = numRowsA, colCnt; /* Loop counters */
arm_status status; /* Status of matrix multiplication */
#ifdef ARM_MATH_MATRIX_CHECK
/* Check for matrix mismatch condition */
if ((pSrcA->numCols != pSrcB->numRows) ||
(pSrcA->numRows != pDst->numRows) ||
(pSrcB->numCols != pDst->numCols) )
{
/* Set status as ARM_MATH_SIZE_MISMATCH */
status = ARM_MATH_SIZE_MISMATCH;
}
else
#endif /* #ifdef ARM_MATH_MATRIX_CHECK */
{
/* The following loop performs the dot-product of each row in pSrcA with each column in pSrcB */
/* row loop */
do
{
/* Output pointer is set to starting address of row being processed */
px = pOut + i;
/* For every row wise process, column loop counter is to be initiated */
col = numColsB;
/* For every row wise process, pIn2 pointer is set to starting address of pSrcB data */
pIn2 = pSrcB->pData;
/* column loop */
do
{
/* Set the variable sum, that acts as accumulator, to zero */
sum = 0.0;
/* Initialize pointer pIn1 to point to starting address of column being processed */
pIn1 = pInA;
#if defined (ARM_MATH_LOOPUNROLL)
/* Loop unrolling: Compute 4 MACs at a time. */
colCnt = numColsA >> 2U;
/* matrix multiplication */
while (colCnt > 0U)
{
/* c(m,n) = a(1,1) * b(1,1) + a(1,2) * b(2,1) + .... + a(m,p) * b(p,n) */
/* Perform the multiply-accumulates */
sum += *pIn1++ * *pIn2;
pIn2 += numColsB;
sum += *pIn1++ * *pIn2;
pIn2 += numColsB;
sum += *pIn1++ * *pIn2;
pIn2 += numColsB;
sum += *pIn1++ * *pIn2;
pIn2 += numColsB;
/* Decrement loop counter */
colCnt--;
}
/* Loop unrolling: Compute remaining MACs */
colCnt = numColsA % 0x4U;
#else
/* Initialize cntCnt with number of columns */
colCnt = numColsA;
#endif /* #if defined (ARM_MATH_LOOPUNROLL) */
while (colCnt > 0U)
{
/* c(m,n) = a(1,1) * b(1,1) + a(1,2) * b(2,1) + .... + a(m,p) * b(p,n) */
/* Perform the multiply-accumulates */
sum += *pIn1++ * *pIn2;
pIn2 += numColsB;
/* Decrement loop counter */
colCnt--;
}
/* Store result in destination buffer */
*px++ = sum;
/* Decrement column loop counter */
col--;
/* Update pointer pIn2 to point to starting address of next column */
pIn2 = pInB + (numColsB - col);
} while (col > 0U);
/* Update pointer pInA to point to starting address of next row */
i = i + numColsB;
pInA = pInA + numColsA;
/* Decrement row loop counter */
row--;
} while (row > 0U);
/* Set status as ARM_MATH_SUCCESS */
status = ARM_MATH_SUCCESS;
}
/* Return to application */
return (status);
}
/**
* @} end of MatrixMult group
*/
#endif // EIDSP_LOAD_CMSIS_DSP_SOURCES
|