home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:8450 alt.folklore.computers:11493 comp.benchmarks:1219
- Newsgroups: comp.arch,alt.folklore.computers,comp.benchmarks
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uunet.ca!geac!itcyyz!yrloc!rbe
- From: rbe@yrloc.ipsa.reuter.COM (Robert Bernecky)
- Subject: Re: dinosaur horsepower
- Message-ID: <1992Jul30.053240.20199@yrloc.ipsa.reuter.COM>
- Reply-To: rbe@yrloc.ipsa.reuter.COM (Robert Bernecky)
- Organization: Snake Island Research Inc, Toronto
- References: <Bs3oGt.6vp@helios.physics.utoronto.ca>
- Date: Thu, 30 Jul 92 05:32:40 GMT
- Lines: 47
-
- In article <Bs3oGt.6vp@helios.physics.utoronto.ca> mark@cita.toronto.edu writes:
- >Larry Stewart's proposal regarding the computer history project prompts
- >me to offer a suggestion for something I'd like to see: A compendium of
- >old machines, and just how fast (well, slow) they were compared with the
- >systems we use today, and how much they cost when they were new.
- >
-
- I'm just finishing a course in Philosophy of Math. One topic we covered
- was computer proofs, with the 4-colo(u)r theorem as the prime example.
-
- The prof (Jim Brown -- a superb lecturer, and excellent teacher, for
- those of you interested in the topic!) noted that the proof took
- months on a "supercomputer".
-
- Being the fussbudget I am, and remembering the date of the Scientific
- Armenian article which described it, I decided a bit of research was
- called for: Another article noted that the "supercomputer" was
- an IBM S/370/168. Ahah! False advertising from Big Blew strikes again!
-
- I then went and looked up rough timings on said machine. Hard, because
- the closest thing I have on hand is a System S/370 system summary
- (GA22-7001-2)which gives the following information:
-
- Mainstore 1meg-4 meg.
- 8k-16k Cache
- 80nsec clock(aka ~~12mhz)
- Mainstore cycle time: 480 nsec
- Storage access width: 8 bytes
- Storage interleaving: 4-way
-
- 4-way interleave gets effective mainstore cycle time down to 120nsec,
- for housebroken programs.
-
- This means that the "supercomputer" which solved the 4-color problem
- was about on a par with high-end PC's of today, IF YOU IGNORE THE
- I/O BANDWIDTH of the system.
-
- Given that the 4-color theorem was not I/O intensive, this means that
- today's PCs could solve the same problem in a similar timeframe, a
- mere 20 years or so later.
-
-
- Robert Bernecky rbe@yrloc.ipsa.reuter.com bernecky@itrchq.itrc.on.ca
- Snake Island Research Inc (416) 368-6944 FAX: (416) 360-4694
- 18 Fifth Street, Ward's Island
- Toronto, Ontario M5J 2B9
- Canada
-