"James Tursa" wrote in message <email@example.com>... > > 3) If you need help setting up a custom mex routine to do this I can probably write it for you pretty quickly. Just ask if you need this.
I just did a quick test with a custom mex routine. Calling dpptrf is taking 50% longer than calling chol. My guess is this is reflecting the necessary packing and unpacking of the data in order to call it (all that data copying takes time). So unless you can keep all the data in a packed state through all your iterations I doubt a custom mex routine using dpptrf (or spptrf as the case may be) will help your timing.