mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > PrimeNet

Reply
 
Thread Tools
Old 2009-01-03, 21:50   #1
stars10250
 
stars10250's Avatar
 
Jul 2008
San Francisco, CA

3×67 Posts
Default On Primenet credit and philosophy

When I first joined GIMPS it was because I wanted to find a Mersenne prime. I plotted the prior discoveries and noted that GIMPS was overdue. That made me decide to join, and two weeks later M45 and M46 were found. Had I started two weeks earlier I would have had a significantly better chance :(

After the two discoveries and my first "xxx is not prime," my attention switched to improving my ranking. The low probability of hitting the jackpot was (and is) a harsh reality. As I watched my stats grow, I converted more and more of my old computers to number crunchers. Recognizing that they weren't all that powerful, I set out to build a dedicated machine. Then came an interest in overclocking. Today I have 5 computers crunching and a 6th is on the way. Fair enough, I'm hooked.

As more and more people join, I watch my stats slip daily. Every now and then I jump up in rank when a computer completes a task, and then I slowly slide down again. While the "top producers" list makes sense, I feel at odds competing with the giants near the top. I'm not a math department, or a server farm, or whatever it takes for those people to produce at those incredible rates. I'm an individual, driving my family nuts and paying the power bills at the same time. Over time, as more and more people convince their university to run GIMPS, I'll become a blip on a chart. I can't compete with that, and if I ever slow down or stop I'll fade into oblivion at the rate of Moore’s law. Don’t get me wrong, I’m not mad at them for contributing so much. I find it all fascinating. But I’d rather be ranked with people a little more like myself: an individual contributor.

My goal has now flipped back to finding that next Mersenne prime. That’s the only way I see to make a lasting, notable, contribution. It won’t matter much over time that my computers proved xxx number of exponents were not prime. And my wonderful computing power of today, and even tomorrow, will eventually become too small to bother reporting on the producers list.

I think it would be fun to create a list of GIMPS emotions for new users. My experience has been something like:

Day 1) install prime95 and get excited about finding a Mersenne prime
Day 30) realize I have little chance of finding a Mersenne prime
Day 31) begin converting all available computers to GIMPS to improve odds
Day 32) realize those old computers--combined--add up to a pile of crap
Day 60) convince gf that she needs a new computer with a little green box in the lower right corner
Day 95) convince gf that the kids and me all need new computers with little green boxes
Day 125) finally understand enough about overclocking to go for it
Day 150) decide that two overclocked computers are better than one
Day 155) recognize that the top producers are accelerating away from me
Day 156) shift strategy back to finding a Mersenne prime

Comments on what my future holds?
stars10250 is offline   Reply With Quote
Old 2009-01-04, 00:05   #2
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

17×251 Posts
Default

For me:
Year 3.5) realize that I have practically no chance at finding a new mersenne prime and would rather find smaller (but still top-5000 primes) primes with the new NPLB project

Last fiddled with by Mini-Geek on 2009-01-04 at 00:14
Mini-Geek is offline   Reply With Quote
Old 2009-01-04, 04:35   #3
jinydu
 
jinydu's Avatar
 
Dec 2003
Hopefully Near M48

2×3×293 Posts
Default

I'm not so sure that you'll slide further and further down the list over time. While the giants like Curtis Cooper will accelerate away, new giants don't appear all that often. If you keep up with the technology, I would think your ranking should stay relatively stable.

As for me, I joined GIMPS in November 2003. I've managed to stay motivated by watching my stats rise (slowly approaching 1000 P90 years), my rankings rise (although the v5 transition has distorted that figure) and even trying to set a new record (see the M100,000,039 thread).
jinydu is offline   Reply With Quote
Old 2009-01-04, 05:23   #4
stars10250
 
stars10250's Avatar
 
Jul 2008
San Francisco, CA

3×67 Posts
Default

I agree that if I keep up with technology then my rank will be decent. If I ever stop then I will slide into oblivion. I guess I'm partly thinking about my recognized contribution to GIMPS. If I find a Mersenne prime, then I get remembered forever. If I don't, then a short time after I stop (for whatever reason) I'm forgotten. I'm just saying that my motivation has changed. While I too love to watch my stats, to get long-lasting (read infinity) "credit," I need to find a Mersenne prime.
stars10250 is offline   Reply With Quote
Old 2009-01-04, 11:33   #5
garo
 
garo's Avatar
 
Aug 2002
Termonfeckin, IE

23·73 Posts
Default

I think right now the stats are a bit wonky due to the v4 credit switch and as more and more people link their v4 account to v5 you see more and more of them jumping up past you. I find this slide in my ranking quite disheartening as well. I seem to jump 5 ranks one day only to slide down two a day for the next 5 days. At some point the new accounts will stabilize and after Oct 20 2009, you will magically see your rank rise as all old credit disappears from the Top Producers report as it is set by default to only show production in the past 65 days.
garo is offline   Reply With Quote
Old 2009-01-05, 09:36   #6
bayanne
 
bayanne's Avatar
 
"Tony Gott"
Aug 2002
Yell, Shetland, UK

2×151 Posts
Default

Quote:
Originally Posted by garo View Post
I think right now the stats are a bit wonky due to the v4 credit switch and as more and more people link their v4 account to v5 you see more and more of them jumping up past you. I find this slide in my ranking quite disheartening as well. I seem to jump 5 ranks one day only to slide down two a day for the next 5 days. At some point the new accounts will stabilize and after Oct 20 2009, you will magically see your rank rise as all old credit disappears from the Top Producers report as it is set by default to only show production in the past 65 days.
What's the method for linking the v4 account details to v5 .....

Last fiddled with by bayanne on 2009-01-05 at 09:37
bayanne is offline   Reply With Quote
Old 2009-01-05, 11:53   #7
garo
 
garo's Avatar
 
Aug 2002
Termonfeckin, IE

23·73 Posts
Default

Alas bayanne, the team credit transfer still does not work. So only those who had their own account in v4 can link it to their account in v5 using the v4 migration page.

PS: The 65 is my previous post should read 365.
garo is offline   Reply With Quote
Old 2009-01-05, 16:27   #8
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

26·67 Posts
Default Aiming for 90th Percentile....

I realize there is still some dust to settle up to Oct 20, 2009 when the V4 Anonymous is dropped and while other V4 people continue to migrate....

That being said, my current goal is to attain the 90th Percentile in:
TF: 94%
P1: 92%
LL: 81% (this one will be tough - it's the most popular)
DC: 88%
ECM: I haven't tried this yet ... maybe I will when I get the other 4.

I currently have 7 full time PC's ranging from a PIII 866 to a Quad-Q9550. I hope that is enough to make my goals.
petrw1 is offline   Reply With Quote
Old 2009-01-05, 16:39   #9
Mr. P-1
 
Mr. P-1's Avatar
 
Jun 2003

49116 Posts
Default

Quote:
Originally Posted by stars10250 View Post
When I first joined GIMPS it was because I wanted to find a Mersenne prime. I plotted the prior discoveries and noted that GIMPS was overdue. That made me decide to join, and two weeks later M45 and M46 were found. Had I started two weeks earlier I would have had a significantly better chance :(
That's the gambler's fallacy. If Mersenne primes are effectively randomly distributed, and all the evidence is that they are, then there is no sense in which one can be considered "due" or "overdue", and the chance of your discovering one is independent of the discovery of any other.

To answer your question, although rationally I know that turning off a tap in my home to conserve water has a negligible effect on the water supply, I still do it. It just feels wrong to leave it running.

An idling computer to me feels like a running tap.

When the P-1 factoring method was first implemented, I saw a way to maximize the value of my meager resources, even though it was a lot of effort setting up the worktypes manually, and the credit given was crap. On the plus side, finding really big factors - 100+ bits - is rather cool.

With V5, It's easy to specialize in this type of work, and you get good credit for it. Currently I'm at 22 in the individual producer rankings for this work, and rapidly gaining on one of those university departments now barely 10GHz days ahead. and I have only one computer doing this. Long-term, I know that I will slip as new, better resourced contributers register with V5 or decide to specialize in this kind of work. But I'm hopeful that I will reach the top 20 before this happens.

Last fiddled with by Mr. P-1 on 2009-01-05 at 16:41
Mr. P-1 is offline   Reply With Quote
Old 2009-01-05, 18:03   #10
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

10000101010112 Posts
Default

Quote:
Originally Posted by Mr. P-1 View Post
That's the gambler's fallacy. If Mersenne primes are effectively randomly distributed, and all the evidence is that they are, then there is no sense in which one can be considered "due" or "overdue", and the chance of your discovering one is independent of the discovery of any other.
Mostly this is true, but on the other hand, perhaps he was referring to the possibility of being assigned M45 or M46, not any gambler's fallacy effect.
Mini-Geek is offline   Reply With Quote
Old 2009-01-05, 18:34   #11
stars10250
 
stars10250's Avatar
 
Jul 2008
San Francisco, CA

110010012 Posts
Default

Yes, I was speaking of the possibility of being assigned M45 or M46. If it takes maybe a month or so to do a LL test, and I was two weeks into mine when the discoveries were made, had I started a few weeks earlier I may have been assigned one of the winning exponents.

I plotted the discovery dates and (while I forget the exact details) noticed that new Mersenne primes were being discovered something like every 1.5-2 years. At that time, it had already been close to 2 years so I figured GIMPS was about due. I don't claim this to be at all scientific, but it was generally true on my plot for several of the last discoveries. Of course, M46 came along right after M45, which discounts all of this! Regardless, it was my motivation for joining when I did.
stars10250 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Philosophy davieddy Lounge 12 2011-07-11 19:21
PrimeNet credit CPU efficiency calculator James Heinrich PrimeNet 22 2009-05-09 19:21
Soap Box Philosophy AES Soap Box 3 2009-03-22 05:34
How much credit for LL? hj47 PrimeNet 26 2009-01-23 22:14
v4 Credit on v5 precius1 Information & Answers 3 2008-11-03 22:23

All times are UTC. The time now is 13:41.

Wed Jun 3 13:41:00 UTC 2020 up 70 days, 11:14, 2 users, load averages: 1.41, 1.43, 1.41

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.