Some time ago, there was a discussion on this group of an IBM program product that provided a library of floating point matrix subroutines that were optimized for the Rs.6000. Can anyone post/email me the name of that library? Thanks.