Ian Davis has written this blog entry, which investigates the performance of several matrix multiplication implementations, including an implementation parallelized with the Task Parallel Library. It’s a very interesting and informative read; be sure to check it out when you get a chance!

James Rapp – Parallel Computing Platform