/* Copyright (C) 2016 - IIT Bombay - FOSSEE This file must be used under the terms of the CeCILL. This source file is licensed as described in the file COPYING, which you should have received as part of this distribution. The terms are also available at http://www.cecill.info/licences/Licence_CeCILL_V2-en.txt Author: Mushir Organization: FOSSEE, IIT Bombay Email: toolbox@scilab.in */ /* ** \brief Compute a multiplication for int8 matrixes. ** \param in1 : input matrix. ** \param lines1 : lines of in1 matrix. ** \param columns1 : columns of in1 matrix. ** \param in2 : input arry. ** \param lines2 : lines of in2 matrix. ** \param columns2 : columns of in2 matrix. ** \param out : Matrix that contains the multiplication in1 * in2. */ /* dgemm function of lapack library does not support uint8,uint16,int8 and int16 datatype so removed*/ #include "matrixMultiplication.h" void i8mulma(int8 *in1,int lines1,int columns1, int8 *in2,int lines2,int columns2, int8 *out) { /* Do Not Use Any BLAS Function*/ int i = 0; int k = 0; int8 accu = 0; /* How to convert 2 index matrixes to one. ** #define in1(a,b) in1[a+b*lines1] ** #define in2(c,d) in2[c+d*lines2] */ for(i=0;i < lines1 * columns2; ++i) { accu = 0; for(k=0;k < columns1; ++k) { accu += in1[i % lines1 + k * lines1] * in2[k + (i / lines1) * lines2]; } out[i] = accu; } }