How can I speedup neural network training and simulation of a radial basis network using parallel computing toolbox in Matlab R2012A

I have made a matlab script for predicting various parameters using a radial basis network function newrb.Usually,It takes a huge time for completing the entire process as I train the network for 230 iterations in order to get the optimum result.Now,I need to speed up the training and simulation using parallel computing toolbox.I know the way of doing this using parfor.But,I have read somewhere that the entire training algorithm(built in matlab functions) can be executed in parallel using UseParallel switch.Is it possible in R2012A.Any help would be highly appreciated.

Answer 1

[Jill Reese]

I think that the useParallel flag was introduced to the Neural Network Toolbox in the R2012b release. Here is a link to the documentation:

http://www.mathworks.com/help/nnet/gs/parallel-computing.html