From: Spike Jones (spike66@ibm.net)
Date: Mon Aug 09 1999 - 21:46:19 MDT
At Extro4, Dr. Gregory Stock gave a pitch on the effect of
cash prizes on research. [Extropians has a great collection of
Gregs: Stock, Bear, Burch, and jones, just to name 4]. We
have an outstanding datum for this topic, found at the URL
http://entropia.com/ips/stats.html
This is the GIMPS project, or Great Internet Mersenne
Prime Search, which uses your spare CPU cycles (behind screen savers)
to search for Mersenne primes. I have had my computers working
on for a little over a year. The URL shows the amount of
computing time in the effort as a function of time, since the
inception. The project's CPU time fits a parabolic curve.
Nowthen, you will notice a dip in the curve starting at about
1 April 1999. There had been a $1500 token award for the next
Mersenne to be found [they were found at a rate of approximately
1 every year]. The project was moving along smoothly, picking
up participants and faster computers all along, until 1 April, 1999
when two things were announced. On that day, the project's software
author announced that he had found a bug in the code which
invalidated much the results of the previous several months. Secondly,
on that day the EFF [Electronic Frontier Foundation] announced
a 50,000 dollar prize for the first million digit prime, which would
most likely be the next prime discovered by GIMPS.
I and most others assumed both were April Fools jokes, but, as
it turns out, both were true. It was a pure Stephenson-esque
coincidence that they both happened on April 1. The bug required
downloading a new version, and all results turned in under the old
version after 1 April were awarded zero credit on the chart.
This explains the dip in the growth history curve. It is in a way most
unfortunate that the award and the bug coincided, for this prevents
us from studying their effects on group participation separately.
If any of you follow IdeasFutures, you may have seen a meme in
which some yahoo proposed that a two million digit prime would
be discovered before 1-1-2000. This meme was recently retired in the
affirmative, as a GIMPS participant (happily, it was a young newlywed)
discovered the 38th known Mersenne prime on June 1. He and
his young bride are 50k richer. Turns out, I was the yahoo that proposed
the 2E6Pri meme on IdeasFutures. (If you are interested I will share
the mathematics of how I calculated that a two million digit prime
would probably show up before 1-1-2000.)
In conclusion, I make this offer: if Extropians wish to form a
GIMPS team, I will donate all my currently held CPU time, which
is about 3.4 CPU years. The EFF prize for the first ten million digit
prime is $100,000, which if we should win, I propose we donate
to extropians.org.
Note, I do not wish to compete for spare CPU
cycles with the extropian SETI@home team, but I know several
extropians have expressed disdain for SETI@home. Perhaps if
some of these individuals do not have a similar disdain for
mathematics, they would wish to take me up on my offer of
3.4 CPU years as a seed for an extropians GIMPS team.
Mathematically yours, spike
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:04:42 MST