Matlab与C/C++混合编程（mex文件）

Matlab是一种矩阵语言，是为vector和matrix操作设计的，这两种类型的运算速度非常快，但是当涉及到大量的循环处理时，Matlab的速度就有点慢了。因此，Matlab提供了MEX脚本来使用C/C++函数以提高循环的运算速度。

现在我们来介绍一下如何编写相应的C/C++函数以生成对应的mexw32或mexw64文件。

MEX文件的编写和编译需要两个基本条件：一是必须按照Matlab应用程序接口组件设计；二是要有相应的C/C++编译器。这些我们先不谈，先来看看如何编写Source File。

我们以一个例子来讲——编写一个Mex文件来调用一个C函数：arrayProduct【这个函数实现一个常量与数组相乘】

文件信息，非必需，可选：

  1: 
  2: /*==========================================================
  3:  * arrayProduct.c - example in MATLAB External Interfaces
  4:  *
  5:  * Multiplies an input scalar (multiplier) 
  6:  * times a 1xN matrix (inMatrix)
  7:  * and outputs a 1xN matrix (outMatrix)
  8:  *
  9:  * The calling syntax is:
 10:  *
 11:  *    outMatrix = arrayProduct(multiplier, inMatrix)
 12:  *
 13:  * This is a MEX-file for MATLAB.
 14:  * Copyright 2007-2012 The MathWorks, Inc.
 15:  *
 16:  *========================================================*/
 17: /* $Revision: 1.1.10.4 $ */
 18: 

添加头文件mex.h，该文件包含相应的MATLAB API函数声明。

  1: #include "mex.h"

核心计算程序。

  1: /* The computational routine */
  2: void arrayProduct(double x, double *y, double *z, mwSize n)
  3: {
  4:     mwSize i;
  5:     /* multiply each element y by x */
  6:     for (i=0; i<n; i++) {
  7:         z[i] = x * y[i];
  8:     }
  9: }
 10: 

MEX文件的入口函数——mexFunction

参数	描述
nlhs	# of output arguments, or the size of the plhs array
plhs	Array of output arguments
nrhs	# of input arguments, or the size of the prhs array
prhs	Array of input arguments.

  1: 
  2: /* The gateway function */
  3: void mexFunction( int nlhs, mxArray *plhs[],
  4:                   int nrhs, const mxArray *prhs[])
  5: {
  6:     double multiplier;              /* input scalar */
  7:     double *inMatrix;               /* 1xN input matrix */
  8:     size_t ncols;                   /* size of matrix */
  9:     double *outMatrix;              /* output matrix */
 10: 
 11:     /* check for proper number of arguments */
 12:     if(nrhs!=2) {
 13:         mexErrMsgIdAndTxt("MyToolbox:arrayProduct:nrhs","Two inputs required.");
 14:     }
 15:     if(nlhs!=1) {
 16:         mexErrMsgIdAndTxt("MyToolbox:arrayProduct:nlhs","One output required.");
 17:     }
 18:     /* make sure the first input argument is scalar */
 19:     if( !mxIsDouble(prhs[0]) || 
 20:          mxIsComplex(prhs[0]) ||
 21:          mxGetNumberOfElements(prhs[0])!=1 ) {
 22:         mexErrMsgIdAndTxt("MyToolbox:arrayProduct:notScalar","Input multiplier must be a scalar.");
 23:     }
 24:     
 25:     /* make sure the second input argument is type double */
 26:     if( !mxIsDouble(prhs[1]) || 
 27:          mxIsComplex(prhs[1])) {
 28:         mexErrMsgIdAndTxt("MyToolbox:arrayProduct:notDouble","Input matrix must be type double.");
 29:     }
 30:     
 31:     /* check that number of rows in second input argument is 1 */
 32:     if(mxGetM(prhs[1])!=1) {
 33:         mexErrMsgIdAndTxt("MyToolbox:arrayProduct:notRowVector","Input must be a row vector.");
 34:     }
 35:     
 36:     /* get the value of the scalar input  */
 37:     multiplier = mxGetScalar(prhs[0]);
 38: 
 39:     /* create a pointer to the real data in the input matrix  */
 40:     inMatrix = mxGetPr(prhs[1]);
 41: 
 42:     /* get dimensions of the input matrix */
 43:     ncols = mxGetN(prhs[1]);
 44: 
 45:     /* create the output matrix */
 46:     plhs[0] = mxCreateDoubleMatrix(1,(mwSize)ncols,mxREAL);
 47: 
 48:     /* get a pointer to the real data in the output matrix */
 49:     outMatrix = mxGetPr(plhs[0]);
 50: 
 51:     /* call the computational routine */
 52:     arrayProduct(multiplier,inMatrix,outMatrix,(mwSize)ncols);
 53: }