Скачать или смотреть Why is My Matlab Code so Slow?

Why is my Matlab code so slow?

Matlab is inherently slow when you have a "for" loop.

That has gotten better since they started using Just in Time compilation routines.

Matlab is an interpreted language. That means it is slower than code that runs through a compiler.

Queue the joke about coders being able to excuse any laziness by saying I'm not wasting my time, I'm waiting for the code to compile.

You can speed up the code by vectorizing it. Or you can compile the Matlab files.

The whole point of using Matlab instead of Fortran or C is to avoid compiling.

Sometimes Matlab itself is the source of the slowdown. Matlab uses a lot of dynamic memory allocation and takes time to re-allocate.

I knew that. But I try to keep open apps to a minimum.

Dynamic memory demands go up with matrix and array size. So you could try to shrink them down.

I don't have that option.

So pre-allocate arrays. Then the code doesn't waste time and resources re-allocating the array when it is resized.

That else can I do to speed up Matlab code?

Use functions directly instead of running scripts. Use load and save within Matlab instead of file input and output functions.

I need results saved to an external file.

If you have a large data set, use a distributed array.

That's like reducing the size of an array.

Yes, and the code no longer has to waste as much time allocating memory for a big array. And then there are the solutions that you can learn from programming.

If I liked programming, I'd be using Python to solve this stuff instead of Matlab.

For example, assign things to variables so Matlab knows what to do instead of dithering. Make sure you use the right logical operator to reduce errors.

Hard to believe math software could run into problems it couldn't solve.

It especially goes into loops with loops.