[RC5] Re: rc5-digest V1 #261

Kai Rode kai_rode at bigfoot.com
Sun Jan 24 17:10:12 EST 1999


On Sat, 23 Jan 1999 15:32:58 -0500, you wrote:

>The other one I think that could gain publicity is the search for
>Pi..ie..PiHex that Colin Percivel(sp?) wrote.  Just about everyone knows
>what Pi is and they also know that it (so far) is an endless number.

Is PiHex suited for distributed computing?

The best known algorithm for calculating digits of pi is as far as I
know the binsplit algorithm. Calculating pi to 1 billion places (that's
one trillion in American English) using distributed computing would be
feasible, but:

- The results sent back to the server have a size of several hundred
megabytes per "block".

- After all blocks have been computed, several computers with 16 GB each
(and much more hard disc space) are needed to put the blocks together
and do some additional calculations on them.

- When this is done, a supercomputer is needed to complete the final
step which has one Terabyte (!) of memory to again do some calculations
on the intermediate results and output pi to one billion digits.

Definitely not something distributed.net can do without sponsors. Let's
rather try some OGRs, they even have some real world applications while
only pi-fanatics care about the billionth digit of pi.

	Kai
-- 
PGP public key available from servers. 0x71D4AD11/0x5CCA01F5
Join http://www.distributed.net/ or GIMPS at www.entropia.com

--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list