optimized assembler routines (math/scientific)

[Dan Severance]

   Hi,
     I have some applications for which a lot of matrix operations, etc.
 are performed.  Does anyone out there (sgi?) have some libraries or
 source which could be used which optimize cache performance, etc.
 My main interests are for single CPU (4D-25, 4D-35) performance enhancements.

     Dan Severance
     dan at rani.chem.yale.edu