Subversion Repositories canSerial

/* ----------------------------------------------------------------------

 * Project:      CMSIS DSP Library

 * Title:        arm_mat_mult_f32.c

 * Description:  Floating-point matrix multiplication

 *

 * $Date:        27. January 2017

 * $Revision:    V.1.5.1

 *

 * Target Processor: Cortex-M cores

 * -------------------------------------------------------------------- */

/*

 * Copyright (C) 2010-2017 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 MatrixMult Matrix Multiplication

 *

 * Multiplies two matrices.

 *

 * \image html MatrixMultiplication.gif "Multiplication of two 3 x 3 matrices"

 * Matrix multiplication is only defined if the number of columns of the

 * first matrix equals the number of rows of the second matrix.

 * Multiplying an <code>M x N</code> matrix with an <code>N x P</code> matrix results

 * in an <code>M x P</code> matrix.

 * When matrix size checking is enabled, the functions check: (1) that the inner dimensions of

 * <code>pSrcA</code> and <code>pSrcB</code> are equal; and (2) that the size of the output

 * matrix equals the outer dimensions of <code>pSrcA</code> and <code>pSrcB</code>.

 */

/**

 * @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_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 *pInA = pSrcA->pData;                /* input data matrix pointer A  */

  float32_t *pOut = pDst->pData;                 /* output data matrix pointer */

  float32_t *px;                                 /* Temporary output data matrix pointer */

  float32_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 */

#if defined (ARM_MATH_DSP)

  /* Run the below code for Cortex-M4 and Cortex-M3 */

  float32_t in1, in2, in3, in4;

  uint16_t col, i = 0U, j, 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 the row being processed */

      px = pOut + i;

      /* For every row wise process, the column loop counter is to be initiated */

      col = numColsB;

      /* For every row wise process, the pIn2 pointer is set

       ** to the starting address of the pSrcB data */

      pIn2 = pSrcB->pData;

      j = 0U;

      /* column loop */

      do

      {

        /* Set the variable sum, that acts as accumulator, to zero */

        sum = 0.0f;

        /* Initiate the pointer pIn1 to point to the starting address of the column being processed */

        pIn1 = pInA;

        /* Apply loop unrolling and compute 4 MACs simultaneously. */

        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) */

          in3 = *pIn2;

          pIn2 += numColsB;

          in1 = pIn1[0];

          in2 = pIn1[1];

          sum += in1 * in3;

          in4 = *pIn2;

          pIn2 += numColsB;

          sum += in2 * in4;

          in3 = *pIn2;

          pIn2 += numColsB;

          in1 = pIn1[2];

          in2 = pIn1[3];

          sum += in1 * in3;

          in4 = *pIn2;

          pIn2 += numColsB;

          sum += in2 * in4;

          pIn1 += 4U;

          /* Decrement the loop count */

          colCnt--;

        }

        /* If the columns of pSrcA is not a multiple of 4, compute any remaining MACs here.

         ** No loop unrolling is used. */

        colCnt = numColsA % 0x4U;

        while (colCnt > 0U)

        {

          /* c(m,n) = a(1,1)*b(1,1) + a(1,2) * b(2,1) + .... + a(m,p)*b(p,n) */

          sum += *pIn1++ * (*pIn2);

          pIn2 += numColsB;

          /* Decrement the loop counter */

          colCnt--;

        }

        /* Store the result in the destination buffer */

        *px++ = sum;

        /* Update the pointer pIn2 to point to the  starting address of the next column */

        j++;

        pIn2 = pSrcB->pData + j;

        /* Decrement the column loop counter */

        col--;

      } while (col > 0U);

#else

  /* Run the below code for Cortex-M0 */

  float32_t *pInB = pSrcB->pData;                /* input data matrix pointer B */

  uint16_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 pInA with each column in pInB */

    /* row loop */

    do

    {

      /* Output pointer is set to starting address of the row being processed */

      px = pOut + i;

      /* For every row wise process, the column loop counter is to be initiated */

      col = numColsB;

      /* For every row wise process, the pIn2 pointer is set

       ** to the starting address of the pSrcB data */

      pIn2 = pSrcB->pData;

      /* column loop */

      do

      {

        /* Set the variable sum, that acts as accumulator, to zero */

        sum = 0.0f;

        /* Initialize the pointer pIn1 to point to the starting address of the row being processed */

        pIn1 = pInA;

        /* Matrix A columns number of MAC operations are to be performed */

        colCnt = numColsA;

        while (colCnt > 0U)

        {

          /* c(m,n) = a(1,1)*b(1,1) + a(1,2) * b(2,1) + .... + a(m,p)*b(p,n) */

          sum += *pIn1++ * (*pIn2);

          pIn2 += numColsB;

          /* Decrement the loop counter */

          colCnt--;

        }

        /* Store the result in the destination buffer */

        *px++ = sum;

        /* Decrement the column loop counter */

        col--;

        /* Update the pointer pIn2 to point to the  starting address of the next column */

        pIn2 = pInB + (numColsB - col);

      } while (col > 0U);

#endif /* #if defined (ARM_MATH_DSP) */

      /* Update the pointer pInA to point to the  starting address of the next row */

      i = i + numColsB;

      pInA = pInA + numColsA;

      /* Decrement the 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

 */