MEX Function c++ vs c interface running speed

Question

Xuechao Qin 2022-4-19

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1699745-mex-function-c-vs-c-interface-running-speed

编辑： Xuechao Qin 2022-4-21

Hi Team,

Recently I am working on boosting our matlab code performance. The old c mex API works good and I noticed that matlab 2018 introduced a new c++ mex function to avoid data copy. I believe it suppose to be faster but in my code the performance is pretty bad for C++ mex function.

I tested in my local env c++ mex code need 70s for my test data about 50 rounds iteration, but c mex only took 0.089s total. In my test I know in C++ mex the iteration is not optimized but in my scenario I am unable to use sample code iterator (auto& elem : inMatrix) .

Can someone help explain? Thanks in advance

Sample code:

ret(di,:) = basedata(di,:) ./ (basedata(di-1,:) * flag(di:, );

So I converted this function for both c mex and c++ mex function:

c++ mex:

        size_t numRows = inputs[0].getDimensions()[0];
        size_t numColumns = inputs[0].getDimensions()[1];
        TypedArray<double> ret = std::move(inputs[0]);
        TypedArray<double> data = inputs[1];
        TypedArray<double> tradable = inputs[2];
        size_t i, j, pos;
        for (i = 1; i < numRows; i++) {
            for (j = 0; j < numColumns; j++) {
                if (data[i - 1][j] > 1 && data[i][j] > 1)
                    ret[i][j] = (data[i][j] / data[i - 1][j] - 1.0) * flag[i][j];
            }
        }
        outputs[0] = ret;

c mex:

    double *ret = mxGetPr(prhs[0]);
    double *data = mxGetPr(prhs[1]);
    double *tradable = mxGetPr(prhs[2]);
    size_t ROWS = mxGetM(prhs[0]);
    size_t N_INSTRUMENTS = mxGetN(prhs[0]);
    for (j = 0; j < N_INSTRUMENTS; j++) {
        for (i = 1; i < ROWS; i++) {
            pos = j * ROWS + i;
            if (data[pos - 1] > 1 && data[pos] > 1)
                ret[pos] = (data[pos] / data[pos - 1] - 1.0) * flag[pos];
        }
    }

3 个评论
显示 1更早的评论隐藏 1更早的评论

Xuechao Qin 2022-4-19

在 MATLAB Online 中打开

Yes, i think the matrix[i][j] overload has some issues. If I only do iteration like below, the speed is almost identical as c mex function. But there is no guidance about how to index 2d array faster in C++ mex.

        for (auto &elem: ret) {
            elem = 1;
        }

Bruno Luong 2022-4-19

That why I never like C++; all the inner details are hiden to user.

请先登录，再进行评论。

请先登录，再回答此问题。

MEX Function c++ vs c interface running speed

3 个评论
显示 1更早的评论隐藏 1更早的评论

回答（0 个）

另请参阅

类别

标签

产品

Community Treasure Hunt

MEX Function c++ vs c interface running speed

3 个评论 显示 1更早的评论隐藏 1更早的评论

回答（0 个）

另请参阅

类别

标签

产品

Community Treasure Hunt

3 个评论
显示 1更早的评论隐藏 1更早的评论