mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Open Projects (https://www.mersenneforum.org/forumdisplay.php?f=80)
-   -   Small distributed projects? (https://www.mersenneforum.org/showthread.php?t=5893)

Citrix 2006-05-19 23:33

Small distributed projects?
 
There are several intresting puzzles/open problems on the internet that require some one to write some code and others to run it for a 1-2 months and the puzzle/open problem can be solved.

I would like to start/suggest such a forum where people can post such problems. If the problem is intresting, people can write some code and then others can run it. Most projects here will be small, basically finished in a few weeks.

Here are a few examples of the kind of problems that can be attempted with very little computing power.

[URL]http://www.mersenneforum.org/showthread.php?t=5866[/URL]
[URL]http://www.mersenneforum.org/showthread.php?t=5630[/URL]
[URL]http://www.mersenneforum.org/showthread.php?t=3474[/URL]

Any thoughts:question:


:bow:

jasong 2006-05-26 04:37

The third thread led me [url=http://www.research.att.com/~njas/sequences/?q=keyword:more]here.[/url]

I would very much enjoy running code for any of these projects.

bearnol 2006-05-26 13:27

UPINT D9
 
Citrix,
wrt your third thread quoted...
It's easy to prove a^x-b^y = n [_any_ n] with slight modification of my FLT/Catalan reasoning.
J

jasong 2006-05-28 19:29

It obviously isn't a "small" project, but I just read in "Scientific America"(I think) that they would like to use computers to prove that there is no Sudoku puzzle with 16 numbers given that has a unique answer.

So far they've found one that has two unique answers, and they've proven it only takes 17 numbers to make a puzzle unique. But they've never found a unique puzzle with 16 numbers given.

The article said it would take about 10,000 computer years to solve with present technology.

Citrix 2006-05-28 22:15

I don't think there are any easy solutions to this that can be found in 2-3 weeks on a few computers.

xilman 2006-05-29 10:07

[QUOTE=jasong]So far they've found one that has two unique answers ...[/QUOTE]
How can something have two [b]unique[/b] answers?

Surely, if the answer is unique there must be precisely one of them.

One of us is missing something. :unsure:

Paul

Citrix 2006-05-29 10:35

Aren't all solutions unique in themselves?

Wacky 2006-05-29 11:17

[QUOTE=xilman]How can something have two [b]unique[/b] answers?[/QUOTE]

Paul,
I think that the difficulty is a typographic spelling error. Rather than [b]unique[/b], I think that it should be spelled [b]distinct[/b], or something like that.

jasong 2007-08-09 21:46

[QUOTE=Wacky;81330]Paul,
I think that the difficulty is a typographic spelling error. Rather than [b]unique[/b], I think that it should be spelled [b]distinct[/b], or something like that.[/QUOTE]
Since this thread has been resurrected by a link from another thread, I'd like to say, yes, it was a typographic error.

Zeta-Flux 2007-08-10 15:41

I'm surprised that there isn't a distributed project to solve the 7-piece endgame problem (or the 6-piece endgame problem with castling). [For chess.]

Citrix 2007-08-11 03:02

Is there a reference to this problem? Any code available?

Citrix 2007-08-11 03:41

This might make for some good short projects
[url]http://www.recmath.org/contest/[/url]
We can donate the money to charity or distribute amongst users... whatever the users decide on....

Citrix 2007-12-17 12:34

Rules
 
Threads that are inactive for more than 2 weeks will be locked. If you want a thread opened PM me. This will help reduce the clutter. :smile:
Try to keep your project limited to 1-2 threads.

gd_barnes 2007-12-17 20:29

[quote=Citrix;120924]Threads that are inactive for more than 2 weeks will be locked. If you want a thread opened PM me. This will help reduce the clutter. :smile:
Try to keep your project limited to 1-2 threads.[/quote]

Thanks for making "Sticky's" out of the "Conjectures 'R Us" effort, Citrix. I was going to ask for that today and you beat me to it! :smile:

I know we have 4-5 threads for the effort but it's potentially so large that I think we'll need them. Also Xyzzy indicated that we may get our own project forum at some point.


Thanks,
Gary

gd_barnes 2007-12-18 21:01

[quote=Citrix;120924]Threads that are inactive for more than 2 weeks will be locked. If you want a thread opened PM me. This will help reduce the clutter. :smile:
Try to keep your project limited to 1-2 threads.[/quote]

Citrix,

You can now lock the Sierpinski bases 6 thru 18, base 10, base 22, and base 23 threads whenever you think it is appropriate.

About the "even k's and the Sierprinski and Riesel conjectures" threads. We'd want to keep those open for now. At the current time, I am not including any multiples of the bases in the conjectures searches. But within the next month, I plan to create a separate page just for k's that are multiples of the base but have a different lowest first primes for n>=1, i.e. 45*10^n-1 has a prime at n=1, but 450*10^n-1 doesn't have a prime until n=11958; so that is one that would be shown. k=4500 would not be shown because it has the same lowest prime as k=450. The same thing for 22*22^n+1, which has no prime at all, although 1*22^n+1 is prime for n=1.

The even k's and the Sierpinski and Riesel conjectures threads will fit right in with the 'multiples of the base' effort. If those aren't 'official prime search projects', I'll check with the folks in those threads and see if they are OK with including them in this effort.


Thanks,
Gary

jasong 2008-01-13 06:20

The "I Hate n=333,333" Project
 
Would anyone like to help me start an "I hate n=333,333" Project? Basically, the goal of this project is to find and prove primes bigger than about 100,365 digits, and submit them to Prime Pages. There is something very annoying to me about how boring the bottom of the top-5000 list looks.

Basically, we'd pick an n-value slightly above n=333,333, and try to knock all the n=333,333 primes off the list as quickly as possible.

Anybody interested?

gd_barnes 2008-01-13 08:31

[quote=jasong;122750]Would anyone like to help me start an "I hate n=333,333" Project? Basically, the goal of this project is to find and prove primes bigger than about 100,365 digits, and submit them to Prime Pages. There is something very annoying to me about how boring the bottom of the top-5000 list looks.

Basically, we'd pick an n-value slightly above n=333,333, and try to knock all the n=333,333 primes off the list as quickly as possible.

Anybody interested?[/quote]

I feel similarly about the n=333333 primes and would love to knock all of them off.

I don't agree with fixed-n searches unless all k-values with primes are 'filled in' up to the n-value of the prime afterwords. Look for something big from me in less than 2 weeks that is not fixed-n but that will help with this. You'll know it when you see it.


Gary

axn 2008-01-13 15:23

[QUOTE=jasong;122750]Basically, we'd pick an n-value slightly above n=333,333, and try to knock all the n=333,333 primes off the list as quickly as possible[/QUOTE]

And then, once all the 333333 primes have been knocked off, you (or someone else) just turns around and start a "I hate n = whatever" thread? And so on and so forth?

You should think this thru so that you don't end up repeating history.

jasong 2008-01-13 16:28

[QUOTE=axn1;122763]And then, once all the 333333 primes have been knocked off, you (or someone else) just turns around and start a "I hate n = whatever" thread? And so on and so forth?

You should think this thru so that you don't end up repeating history.[/QUOTE]
Yeah, I thought about commenting on this, I guess I should have gone with my gut. Basically, I'm thinking we would pick a small range, something that amounts to a year or two of LLR on one core when it's sieved to a sufficient depth. When that's done, simply move on to another number.

mdettweiler 2008-01-18 18:40

[quote=jasong;122767]Yeah, I thought about commenting on this, I guess I should have gone with my gut. Basically, I'm thinking we would pick a small range, something that amounts to a year or two of LLR on one core when it's sieved to a sufficient depth. When that's done, simply move on to another number.[/quote]
Or, you could just crunch for the Conjecture 'R Us project--it's pretty easy to find a prime there, and it won't be long before their second team drive is in top-5000 territory. :smile: Not to mention that you can help prove some conjectures while you're at it.

gd_barnes 2008-01-21 20:03

[quote=jasong;122750]Would anyone like to help me start an "I hate n=333,333" Project? Basically, the goal of this project is to find and prove primes bigger than about 100,365 digits, and submit them to Prime Pages. There is something very annoying to me about how boring the bottom of the top-5000 list looks.

Basically, we'd pick an n-value slightly above n=333,333, and try to knock all the n=333,333 primes off the list as quickly as possible.

Anybody interested?[/quote]


Jasong,

You HAVE your wish!! Now announcing the 'No Prime Left Behind' project. Time to take out many of those n=333333 primes in rapid succession and make the bottom of the top-5000 list look good again! Come join us [URL="http://www.mersenneforum.org/showthread.php?t=9887"]here[/URL].


Gary

Joshua2 2010-01-26 09:04

I would like to write code for something that would do something significant within a year. Any ideas? I thought maybe solving mancala with several different rule sets? Only one has been done.


All times are UTC. The time now is 02:07.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.