/github/workspace/src/MatrixFunctionsStride/mat_add_stride/kernels/plp_mat_add_stride_i32s_xpulpv2.c

Functions

	Name
void	plp_mat_add_stride_i32s_xpulpv2(const int32_t restrict pSrcA, const int32_t restrict pSrcB, uint32_t M, uint32_t N, uint32_t strideA, uint32_t strideB, uint32_t strideY, int32_t *restrict pDst) strided matrix addition of 32-bit integer matrices kernel for XPULPV2 extension.

Functions Documentation

function plp_mat_add_stride_i32s_xpulpv2

void plp_mat_add_stride_i32s_xpulpv2(
    const int32_t *__restrict__ pSrcA,
    const int32_t *__restrict__ pSrcB,
    uint32_t M,
    uint32_t N,
    uint32_t strideA,
    uint32_t strideB,
    uint32_t strideY,
    int32_t *__restrict__ pDst
)

strided matrix addition of 32-bit integer matrices kernel for XPULPV2 extension.

Parameters:

• pSrcA Points to the first input matrix
• pSrcB Points to the second input matrix
• M Height of all matrices
• N Width of all matrices
• strideA Stride of matrix A (elements between each row)
• strideB Stride of matrid B (elements between each row)
• strideY Stride of output matrix (elements between each row)
• pDst Points to the output matrix

Return: none

matrix addition of a 32-bit integer matrices for XPULPV2 extension.

Source code

/* =====================================================================
 * Project:      PULP DSP Library
 * Title:        plp_mat_add_stride_i32s_xpulpv2.c
 * Description:  32-bit integer strided matrix addition for XPULPV2
 *
 * $Date:        1. July 2020
 * $Revision:    V0
 *
 * Target Processor: PULP cores
 * ===================================================================== */
/*
 * Copyright (C) 2020 ETH Zurich and University of Bologna.
 *
 * Author: Tibor Schneider, ETH Zurich
 *
 * 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 "plp_math.h"

void plp_mat_add_stride_i32s_xpulpv2(const int32_t *__restrict__ pSrcA,
                                     const int32_t *__restrict__ pSrcB,
                                     uint32_t M,
                                     uint32_t N,
                                     uint32_t strideA,
                                     uint32_t strideB,
                                     uint32_t strideY,
                                     int32_t *__restrict__ pDst) {

//#define BASIC_VERSION // if used don't forget to also use the undefine at end of file
#ifdef BASIC_VERSION

    uint32_t m, n; // loop counters

    for (m = 0; m < M; m++) {
        for (n = 0; n < N; n++) {
            pDst[m * strideY + n] = pSrcA[m * strideA + n] + pSrcB[m * strideB + n];
        }
    }

#else

    uint32_t m, n; // loop counters

    unsigned int n_iter = N >> 1;
    unsigned int n_rem = N & 0x1;

    unsigned int step_a = strideA - N;
    unsigned int step_b = strideB - N;
    unsigned int step_y = strideY - N;

    for (m = 0; m < M; m++) {
        for (n = 0; n < n_iter; n++) {
            int32_t a1 = *pSrcA++;
            int32_t a2 = *pSrcA++;
            int32_t b1 = *pSrcB++;
            int32_t b2 = *pSrcB++;
            *pDst++ = a1 + b1;
            *pDst++ = a2 + b2;
        }
        if (n_rem) {
            *pDst++ = *pSrcA++ + *pSrcB++;
        }
        pSrcA += step_a;
        pSrcB += step_b;
        pDst += step_y;
    }

#endif
    //#undef BASIC_VERSION
}

---

Updated on 2023-03-01 at 16:16:33 +0000