home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!destroyer!gumby!wupost!bcserv!ponder
- From: ponder@bcserv.wustl.edu (Ponder (Jay))
- Newsgroups: comp.unix.ultrix
- Subject: Anonymous ftp Site for 5000/240 FPU Test
- Date: 11 Nov 92 00:11:16 GMT
- Organization: Washington University in Saint Louis, MO USA
- Lines: 35
- Message-ID: <ponder.721440676@bcserv>
- NNTP-Posting-Host: bcserv.wustl.edu
- Summary: Anonymous ftp Site for 5000/240 FPU Test Case
- Keywords: DEC 5000/240 FPU Test Case
-
- In response to the recent requests for the 5000/240
- FPU test program, I've put it up for anonymous ftp
- on dasher.wustl.edu (128.252.162.151) in the
- subdirectory /pub/test240. There are actually two
- test suites: the original "smith" test that was
- posted to the net over the summer, and the "ponder"
- test that we developed (to our horror!!) locally.
-
- Source is given for the Smith test, an executable
- for the Ponder test (since it is part of a very
- large program). Most machines with the FPU errors
- will fail both tests, but we do know of machines
- that pass Smith and fail Ponder.
-
- Maybe I'm naive, but I'm EXTREMELY PEEVED that DEC
- apparently still has machines out there with this
- problem. There is a hardware fix, an altered CPU
- card that DEC gave us several weeks ago. I'm
- AMAZED that DEC has not "quietly" replaced the
- CPU on the early 240's with the fix. The FPU error
- can, in our experience, range from very, very
- slight to obvious. It certainly can vary from
- machine to machine. This is a VERY, VERY serious
- problem (we were lucky and only lost three months
- of computations) -- please test your 240's if
- you haven't yet! And demand repair if they are
- faulty....
-
- Jay W. Ponder
- Asst. Prof, Biochem & Mol Biophys
- Washington Univ. Med School
- ponder@comet.wustl.edu
- (314) 362-4195
-
-
-