BOINC?

KEP · 2006-11-07, 20:47

Hi I've done a bit of thinking, and come up with this disturbing conclusion, that even though we maintain the userinterface of about 10 (currently registered users at www.twinprimesearch.org) users, it will take us an painstakenly 2 years to reach to 20G, and that is with no guarentee of ever finding a twin. So teherefor, also given the fact that there are so many various prime projects, maybe it was about time, that the various developers came together and developed a common prime searching project under BOINC, in which people can decide to support the search for serpinski primes, mersenne primes, 321 primes or twin primes or others. Maybe Rytis at the www.primegrid.com project, which looks for any prime from 2+ up to about just as much as they can store, can help the developers get started or maybe he can workout a solution from which we are allowed to work on any kind of primefinding project or just one we choose for ourself, which choose to work with such a BOINC based project. I'm sure that at least the twinprimes wan't have trouble attracting users, since they are small WU (2 minutes - 5 minutes max), and the chance of finding something is actually rather strong. Well it was just a thought, and at least it would be a chance to decrease the manual work, on the various projects, and also a chance to increase the userinterface.

Regards!

KEP

MooMoo2 · 2006-11-08, 01:29

(slightly off topic, considering the title)

Quote:

it will take us an painstakenly 2 years to reach to 20G, and that is with no guarentee of ever finding a twin

I was thinking about this the other day, and it turns out that there are some DC projects out there that have much longer completion times. For example, the RC5 - 72 project (see: http://www.distributedcomputing.info...tributeddotnet) has been going on for 4 years, and it hasn't even finished searching 0.5% of its searchspace.

There's no guarantee that you'll find a twin, but the chance that a specific k will be twin is actually over 5 times greater than the chance of one lottery ticket winning the jackpot (1 in ~8 million, compared with 1 in ~41 million).

Back on topic, if anyone has more suggestions on BOINC, feel free to post.

jasong · 2006-11-08, 22:47

Quote:

Originally Posted by MooooMoo

Back on topic, if anyone has more suggestions on BOINC, feel free to post.

Riesel Sieve, another prime-finding project, is going to start work on an LLR BOINC client sometime in the future. That could be very helpful to a project like this.

pacionet · 2006-11-09, 08:26

But LLR BOINC client will be fast as our current LLR client ?

KEP · 2006-11-09, 17:28

Hi folks,

I've just recieved a reply from Rytis, that he wouldn't mind setting up the project such as it can work in BOINCwith a perlBOINC solution for 300€. The only thing then needed, is a server strong enough to handle the BOINC demands, and hopefully increased load when going BOINC. So now the discussion might very well be, not if it is possible, but if the twinprimesearch project can organise a fundraising, to collect the money needed to get Rytis to create our beautifull project in BOINC, so can we and will we do that?

I personally have no posibility of transferring any money to the project, since it would cost me more than I transfer to support this project, since I don't owe a mastercard nor a visa card :(

Regards!

victor · 2006-11-09, 17:41

Couldn't we do it ourselves?
It looks really easy, using that : http://boinc.berkeley.edu/wrapper.php

[edit]
I'll give it a try next week-end :)

axn · 2006-11-09, 18:41

Quote:

Originally Posted by victor

Couldn't we do it ourselves?
It looks really easy, using that : http://boinc.berkeley.edu/wrapper.php

[edit]
I'll give it a try next week-end :)

Isn't that just the client bit? AFAIK, perlBOINC is the server side app. You need both to complete the system.

KEP · 2006-11-09, 19:16

Quote:

Originally Posted by victor

Couldn't we do it ourselves?
It looks really easy, using that : http://boinc.berkeley.edu/wrapper.php

[edit]
I'll give it a try next week-end :)

Honestly I don't know if we can do it ourself, but if we can, of course there is no need to get help from outsiders. But one thing is for sure, if we should ever have some hope of finding a megatwinprime, (1,000,000 digits +), we will have to go BOINC, so the faster we can go BOINC the merryer it will be for out lovely project with a way to small userinterface. I still think that we have the advantage of other BOINC projects, that many pentiums, will only need 6 hours or less to complete one K test, when searching for megatwinprimes, and only few moments searching for any other size of primes, being only a little bigger than our search.

And, yeah I know I'm not a very small thinker, I think big, and fast, and see BOINC as the only opprtunity to find thousands of normal primes, and find hundreds of twins, maybe it even gives us the opportunity to create a "prime ladder", in which we find the first twin prime, for any k*2^n(-/+)1, and increments the bit with 1 when we find a twin. Doing that will ensure that we keep up maintaining the number one place at the twinprime top20 list.

Remember to think big and get far, rather than thinking small and get nowhere :)

Regards!

KEP

pacionet · 2006-11-09, 20:06

Of course turning our project in a BOINC project would be beautiful but it is not cheap. There are some problems:

1) Collect 300$: the current (simple and poor

but quite working ) site costs me $8 for a year. I bought the domain and I administer it from remote.

2) I cannot upload an application on the server , but only PHP / HTML pages (maybe also JSP pages but I am not sure) which can access a database. So it would be necessary a different (and not cheap!) kind of server. BOINC requires (I presume) a server-side application and not only the web server.

3) 300$ would be the initial cost. But it would be necessary to update the site regularly ... and other $$ (and much time!)

So I think that for the moment the site is the best I can offer !

(moreover I have not so much time to spend for the site ...)

Anyway if somebody else have time/dollars to spend ... good luck !

axn · 2006-11-09, 20:11

I think the relevant documentation is here: http://boinc.berkeley.edu/create_project.php

pacionet · 2006-11-09, 20:20

Hardware requirements:

"Hardware: use a host with good CPU capacity (dual Xeon or Opteron), at least 2 GB of RAM, and at least 40 GB of free disk space. Do whatever you can to make it highly reliable (UPS power supply, RAID disk configuration, hot-swappable spares, tempurature-controlled machine room, etc.). If you anticipate a high-traffic project, use a machine whose RAM capacity is 8 GB or more, and that uses 64-bit processors. "

2006-11-07, 20:47	#1
KEP Quasi Admin Thing May 2005 7×11×13 Posts	BOINC? Hi I've done a bit of thinking, and come up with this disturbing conclusion, that even though we maintain the userinterface of about 10 (currently registered users at www.twinprimesearch.org) users, it will take us an painstakenly 2 years to reach to 20G, and that is with no guarentee of ever finding a twin. So teherefor, also given the fact that there are so many various prime projects, maybe it was about time, that the various developers came together and developed a common prime searching project under BOINC, in which people can decide to support the search for serpinski primes, mersenne primes, 321 primes or twin primes or others. Maybe Rytis at the www.primegrid.com project, which looks for any prime from 2+ up to about just as much as they can store, can help the developers get started or maybe he can workout a solution from which we are allowed to work on any kind of primefinding project or just one we choose for ourself, which choose to work with such a BOINC based project. I'm sure that at least the twinprimes wan't have trouble attracting users, since they are small WU (2 minutes - 5 minutes max), and the chance of finding something is actually rather strong. Well it was just a thought, and at least it would be a chance to decrease the manual work, on the various projects, and also a chance to increase the userinterface. Regards! KEP

2006-11-09, 17:28	#5
KEP Quasi Admin Thing May 2005 3E9₁₆ Posts	Setup for 300€ Hi folks, I've just recieved a reply from Rytis, that he wouldn't mind setting up the project such as it can work in BOINCwith a perlBOINC solution for 300€. The only thing then needed, is a server strong enough to handle the BOINC demands, and hopefully increased load when going BOINC. So now the discussion might very well be, not if it is possible, but if the twinprimesearch project can organise a fundraising, to collect the money needed to get Rytis to create our beautifull project in BOINC, so can we and will we do that? I personally have no posibility of transferring any money to the project, since it would cost me more than I transfer to support this project, since I don't owe a mastercard nor a visa card :( Regards!

2006-11-09, 17:41	#6
victor Oct 2005 Fribourg, CH 374₈ Posts	Couldn't we do it ourselves? It looks really easy, using that : http://boinc.berkeley.edu/wrapper.php [edit] I'll give it a try next week-end :) Last fiddled with by victor on 2006-11-09 at 17:42

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
BOINC	Unregistered	Information & Answers	6	2010-09-21 03:31
BOINC.BE	BATKrikke	Teams	2	2010-03-05 18:57
Boinc	Xentar	Sierpinski/Riesel Base 5	4	2009-04-25 10:26
BOINC?	masser	Sierpinski/Riesel Base 5	1	2009-02-09 01:10
BOINC	bebarce	Software	3	2005-12-15 18:35

2006-11-09, 08:26	#4
pacionet Oct 2005 Italy 3×113 Posts	But LLR BOINC client will be fast as our current LLR client ?

2006-11-09, 20:06	#9
pacionet Oct 2005 Italy 523₈ Posts	Of course turning our project in a BOINC project would be beautiful but it is not cheap. There are some problems: 1) Collect 300$: the current (simple and poor but quite working ) site costs me $8 for a year. I bought the domain and I administer it from remote. 2) I cannot upload an application on the server , but only PHP / HTML pages (maybe also JSP pages but I am not sure) which can access a database. So it would be necessary a different (and not cheap!) kind of server. BOINC requires (I presume) a server-side application and not only the web server. 3) 300$ would be the initial cost. But it would be necessary to update the site regularly ... and other $$ (and much time!) So I think that for the moment the site is the best I can offer ! (moreover I have not so much time to spend for the site ...) Anyway if somebody else have time/dollars to spend ... good luck !

2006-11-09, 20:11	#10
axn Jun 2003 1554₁₆ Posts	I think the relevant documentation is here: http://boinc.berkeley.edu/create_project.php

2006-11-09, 20:20	#11
pacionet Oct 2005 Italy 3×113 Posts	Hardware requirements: "Hardware: use a host with good CPU capacity (dual Xeon or Opteron), at least 2 GB of RAM, and at least 40 GB of free disk space. Do whatever you can to make it highly reliable (UPS power supply, RAID disk configuration, hot-swappable spares, tempurature-controlled machine room, etc.). If you anticipate a high-traffic project, use a machine whose RAM capacity is 8 GB or more, and that uses 64-bit processors. "