Matlab function slow first time, much faster subsequently. Why?

Question

Kobye 2013-10-21

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/90968-matlab-function-slow-first-time-much-faster-subsequently-why

评论： dpb 2013-10-21

I have a large MATLAB function file. It first creates a zero matrix, then updates about 70% of the cells by evaluating a number of corresponding (long) algebraic expressions that are hard coded in the function. Once this is complete, a numeric matrix is returned.

The .m file is about 4MB large (I have 100 of these m. files, but that is not directly relevant). When I evaluate the function the first time, it takes about 9s to evaluate. Subsequent runs, however, only take about 0.1s, which is more what I was expecting.

Why does the first evaluation take 9s? Anytime I close and reopen matlab, I each time have this slow first evaluation, with subsequent runs being much faster. Why is this?

The m. file can be found at the below public link (you can copy the text from the browser): https://dl.dropboxusercontent.com/u/157153767/K_a_12_102x.m

The command window input you should use is: [test]=K_a_12_102x(414000000,1.1095e+09,1.2500e-04,0.0840,0.0840,0.0240,0.0240,0.0020,0.0020,0,0,0,0,3.0397e+08,8.9930e+07,0,3.0397e+08,0,1.0702e+08,0,0,0,0,0,0,497.7389,80.7355,-15.9811,391.1985,-15.9811,103.5248,20440000,0,20440000,0.06)

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

dpb 2013-10-21

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/90968-matlab-function-slow-first-time-much-faster-subsequently-why#answer_100407

Matlab is interpreted so when the extremely large code file is read the first time the JIT compiler has to first load it into memory, parse it and compile it. Once that is done, the function is cached in memory until it is released by memory (say by a clear). Obviously when the Matlab session is closed anything in memory is lost so the whole process has to be redone next time.

It's an inevitable result of having such a large function...whether you might gain by factoring into more smaller functions I don't know--I've certainly never tried writing such large m-files.

2 个评论
显示无隐藏无

Kobye 2013-10-21

Thanks very much for your reply which has some good advice. Apparently I may be able to overcome this problem by transferring the number crunching part of the code to C, C# or something similar and including it as dll. This negates the need for JIT compiling everytime I start MATLAB.

As an aside, these 4MB files are already the "smaller" functions, hence why I have 100 of them!