Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!elroy.jpl.nasa.gov!decwrl!sgi!rpw3@rigden.wpd.sgi.com From: rpw3@rigden.wpd.sgi.com (Rob Warnock) Newsgroups: comp.compression Subject: Re: IP gnitaluclaC rof margorP (Was Re: Program for Calculating PI) Message-ID: <97034@sgi.sgi.com> Date: 12 Apr 91 04:28:43 GMT References: <28916@dime.cs.umass.edu> <24380001@hpdmd48.boi.hp.com> Sender: guest@sgi.sgi.com Reply-To: rpw3@sgi.com (Rob Warnock) Organization: Silicon Graphics, Inc., Mountain View, CA Lines: 35 In article jpc@fct.unl.pt (Jose Pina Coelho) writes: +--------------- | Certainly. I would like to compute PI to one billion decimal places on my | Apple IIe, and then store it on a single floppy, so that I can sneakernet | it to my friend's Cray MP to compare results. In order to do so, I need | to have really excellent data compression. | | Won't work, PI is uncompressible, one billion digits won't fit in a | diskette. On the other hand, if you run the program on the cray you'll | have PI in the cray in a lot less time (after all, diskettes are slow). +--------------- But the point was that he wanted to *compare results*. That is, he's not sure if his IIe is computing the same value for Pi as the Cray. For purposes of *comparison*, you can indeed achieve a very *high* degree of "compression", by transporting only "signatures", a.k.a. checksums. That is, on each machine compute a billion digits of Pi, and checksum it with several (hopefully) independent checksums. CRC-32, Snuffle, and MD4 come to mind. [Try sci.crypt for more examples.] Then transport the checksums and compare those. If all are equal, then with some fairly high degree of probability, the billion digits of Pi were equal. Note that since the result is not certain, only highly confident, I guess this is a form of "lossy compression", no? ;-} ;-} -Rob ----- Rob Warnock, MS-1L/515 rpw3@sgi.com rpw3@pei.com Silicon Graphics, Inc. (415)335-1673 Protocol Engines, Inc. 2011 N. Shoreline Blvd. Mountain View, CA 94039-7311