[RC5] Quick positive press effort
ayatolla at lhq.com
Wed Sep 23 14:31:07 EDT 1998
On Tue, 22 Sep 1998, Mike Faunce wrote:
> Positive press for Distributed.net is that we DID something, no matter what it is. I personally don't understand why we didn't FINISH DES-II to see how long it would have taken us, even though we wouldn't have won any money, we would have completed it and could have been listed that as an accomplishment. Who cares that EFF did it faster? They spent money to do it. They had a highly organized team, we're a bunch of people who have never met, working with widely varied equipment, scattered all over. It's not how fast we can do it, it's that we CAN do it.
> So, why not do "quickie" contests or OGR or Mersiennes or whatever that we can solve in months rather than years? If we take 2.5 years to solve RC5-64, what does that prove? Nothing, as at the end of 2.5 years, the computing world will be so different that it won't matter. If we assume that we are doubling our rate every year, then in 2 years we'd be 4 times faster. If we had waited and started 2 years later, we'd finish what we've done so far in 1/4 the time (3 months instead of a year). Using the doubling rate as a given (2% Y1, 4& Y2, 8% Y3, 16% Y4, and so on), if we skip the first year, we've only lost about 7 days at the "big end". If you follow that out, if you skip the first 3 years, you only loose about 2 months.
> But, if we can knock out the "gazillionth" digit of pi next month
and OGR-22 ruler the month after that, and the next Merseinne (sp?)
prime the month after that, then we DONE something. And, what have
we lost? .05% more on RC5-64? Nothing. What have we gained?
A press release.
> Mike Faunce
> email mdfaunce at paranet.com
Certainly worth a shot. The big deal of distributed net still is working
out how to do parallel/distributed processing. The journey in this case
is more the point than just finding the solution.
On another note, would be interesting to run the numbers on our "dead" as
it were, or those who joined up for a time, cranked out blocks, then
vanished. Maybe call a "dead" individual 3 months inactive for the study
once fall kicks in(to rule out those who claim to have lives and went away
for the summer) . See what the breakdown is for the dead. If the majority
are one block wonders, or who ran a normal course, then just faded out.
Would go to make a better model if you figured attrition rates and stats.
One other point. If your CPU is as good as dead in 6-9 months, distributed
type background projects are ideal. If the CPU is cranking out at 95-100%
active time and idle time, you've truely milked the CPU for all its worth
before replacing it.
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5