What is the fastest way to divide data by the covariance matrix?

I am doing the following calculation:

X_adj = X * C^(-1/2);

where X is my data, and C is a covariance matrix. I need to do this repeatedly. Is there a faster way? Typical data size varies between 10x200 and 100x1000.

Answer 1

[Star Strider]

I would see if using the sqrtm function and doing a regular matrix right division works:

X_adj = X/sqrtm(C);

[precuneus]

This is only 5% faster on my machine. I tried also the following (inlining the code):

[E,D] = eig(C);
X_adj = X*E*diag(read(diag(D).^-.5))*E';

which turned out to be 20% faster. Would be happy about other ideas.