Are you interested in actual computation speed or are you interested in theoretical big-O speed? If you want the fastest computation for a specific, reasonable-sized problem with no particular structure (i.e., non-sparse), then using the ordinary matrix multiply algorithm that is coded the best will probably beat any theoretically-faster algorithm. You probably will find the fastest matrix-multiplication code in one of the sets of the so-called Basic Linear Algebra Subprograms (BLAS). Check out http://www.netlib.org/blas/faq.html, and especially 5) therein: http://www.netlib.org/blas/faq.html#5.
Dave On Dec 8, 6:09 am, Luciano Junior <luciano....@gmail.com> wrote: > What is best multiply matrix algorithm for: > > -multiply a n x n matrix by another n x n matrix > -multiply a m x n matrix by a n x p matrix > > I need a best performance cpu algorithm. > Note: it can use a parallel programming concept. > > Thankfully. > ---------------------------------------- > Luciano. -- You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To post to this group, send email to algoge...@googlegroups.com. To unsubscribe from this group, send email to algogeeks+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/algogeeks?hl=en.
