home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!harrier.ukc.ac.uk!eagle.ukc.ac.uk!wgd
- From: wgd@ukc.ac.uk (W.G.Day)
- Newsgroups: comp.sys.transputer
- Subject: Re: Meiko
- Message-ID: <1609@eagle.ukc.ac.uk>
- Date: 12 Sep 92 15:30:17 GMT
- References: <6068@m1.cs.man.ac.uk>
- Reply-To: wgd@ukc.ac.uk (Warren Day)
- Organization: Computing Lab, The University, Canterbury, Kent. CT2 7NF, UK
- Lines: 22
-
- In article <6068@m1.cs.man.ac.uk> bane@vtx.ma.man.ac.uk writes:
- >Could somebody please confirm, or otherwise, that the Meiko machines have
- >switch chips which do not give reproducible results?
-
- Do MEiKO machines have switch chips /\ do these prevent reproducible results?
-
- Yes, MEiKO machines do have switch chips.
-
- Yes, there probably are situations when they give different results for several
- different runs of the same program with all other things fixed.
-
- However, I did some quite extensive testing on our 70 transputer multi-user
- MEiKO system and even did some of the tests (which involved 16 or sometimes
- 32 transputers and were both compute and communication bound) a number of
- times and got invariably identical results. Even with other people booting,
- using and releasing transputers presumably at different times in different
- ways. My results were timed by looking at the transputers on-chip clock at
- low priority.
-
- >Michael Bane
-
- Warren
-