home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.parallel
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!gatech!hubcap!fpst
- From: adriaan@dcs.qmw.ac.uk (Adriaan Joubert)
- Subject: BLAS-1 and 2 for iPSC
- Message-ID: <1992Sep5.130137.8014@dcs.qmw.ac.uk>
- Sender: usenet@dcs.qmw.ac.uk (Usenet News System)
- Nntp-Posting-Host: flint.dcs.qmw.ac.uk
- Organization: Computer Science Dept, QMW, University of London
- Date: Sat, 5 Sep 1992 13:01:37 GMT
- Approved: parallel@hubcap.clemson.edu
- Lines: 28
-
- Hi,
-
- I have written a few linear system solvers for the iPSC/2 in which
- I use a lot of BLAS routines. The BLAS-1 routines are all available
- in the vector library, but I haven't been able to find BLAS-2 and 3
- routines optimised for the sxvx (floating point co-processor & Weitek
- vector unit) nodes. Do such routines exist?
-
- Alternatively, are there versions of BLAS routines which would run more
- efficiently with Weitek vector processors than the standard Fortran
- versions from Netlib? I sent these routines through Vast, but that
- still does not seem to make them very fast...
-
- Thanks,
-
- Adriaan
-
- --
-
- ***************************************************************************
- * Adriaan Joubert * Janet : adriaan@uk.ac.qmw.dcs *
- * Computer Science Dept * Internet : adriaan@dcs.qmw.ac.uk *
- * QMW University of London * Fax : +44-81-980 6533 *
- * Mile End Road * Telephone : +44-71-975 5555 ext.3232 *
- * London E1 4NS * *
- * United Kingdom * *
- ***************************************************************************
-
-