MATLAB: How to make the chi2rnd(size) Matlab function faster for big matrices

optimizationrandom number generator

I'm running a program a couple thousand times, and the bottleneck of it is just creating the chi-squared noise matrix. It has size 2048 by 100,000 and it takes more than 80% of each iteration (I'm doing 'chi2rnd(2048, 1000,000)) Is there anyway to make this function faster? Or to code my own?

Best Answer

storing the chi2rnd(2048, 1000000) results require about 16Gb of memory, perhaps the extra time is due to memory-swapping rather than actually computing these values? (I suggest to evaluate whether your code actually needs the entire matrix being computed in a single step)

Related Solutions

MATLAB: About memory manipulation/speed up program

There are some situations in which removing unused large variables could speed up calculations a lot, but if the variables are just sitting in memory without being touched then probably it is not significantly affecting the speed of your calculations.

What hammers speed the most in MATLAB tends to be failing to pre-allocate arrays to reasonable sizes, and then to increase the size of the array on (most or every) iteration.

MATLAB: U = chi2rnd(nu, [N,1]);

Hi, The degrees of freedom has nothing to do with N = 10000

The degrees of freedom is the parameter that characterizes the chi-square pdf, just like the mean and the variance are the parameters that characterize the Gaussian. For a chi-square density, if you know the degrees of freedom, you know everything about the distribution.

N = 10000 is just the number of samples you want.

u = chi2rnd(5,1000,1); % draw 1000 samples chi-square density  with
% 5 degrees of freedom
u = chi2rnd(5,1e4,1); % draw 10000 samples

Best Answer

Related Solutions

MATLAB: About memory manipulation/speed up program

MATLAB: U = chi2rnd(nu, [N,1]);

Related Question