Xref: utzoo comp.sys.sgi:3399 comp.sys.mips:563 sci.math.num-analysis:650
Path: utzoo!attcan!uunet!cs.utexas.edu!wuarchive!udel!udccvax1!mccalpin
From: mccalpin@vax1.acs.udel.EDU (John D Mccalpin)
Newsgroups: comp.sys.sgi,comp.sys.mips,sci.math.num-analysis
Subject: LAPACK anyone?
Keywords: linear algebra, math software
Message-ID: <5904@udccvax1.acs.udel.EDU>
Date: 21 Mar 90 14:26:37 GMT
Followup-To: comp.sys.sgi
Organization: College of Marine Studies, Univ. of Delaware
Lines: 9

Has anyone done a reasonably optimized port of LAPACK to a generic
scalar machine (specifically the mips or sgi machines)?

The ability to do block-mode level-3 blas should help the performance
of cache-busting linear algebra jobs by a factor of 2-3 on these machines.
-- 
John D. McCalpin                               mccalpin@vax1.acs.udel.edu
Assistant Professor                            mccalpin@delocn.udel.edu
College of Marine Studies, U. Del.             mccalpin@scri1.scri.fsu.edu