Mersenne BOINC coming?
 

[URL="http://mersenneathome.net/index.php"]Interesting...[/URL]

Maybe GIMPS has decided to join BOINC? George?

[QUOTE=NBtarheel_33;242390]Maybe GIMPS has decided to join BOINC? George?[/QUOTE]

That is an independent project.

Wow, I wonder if they are aware of GIMPS. This reminds me of the fellow who is re-running every LL test on his home server.

Will be interesting to see how their code stacks up against Prime95.

[QUOTE=Prime95;242396]That is an independent project.[/QUOTE]

It would be a real shame (IMHO) if they started from scratch and re-did all the work GIMPS has been doing the last 15 or so years.

Other than a smidgen of benefit from re-double checking GIMPS results.

[QUOTE=frmky;242385][URL="http://mersenneathome.net/index.php"]Interesting...[/URL][/QUOTE]My security software says that's a suspicious site and advises me to leave immediately.

[QUOTE=cheesehead;242418]My security software says that's a suspicious site and advises me to leave immediately.[/QUOTE]Seems to be time to remove your useless security software then. There is not even any JS on that site (a rare thing these days).

Perhaps this project intends to serve as a "wrapper" platform for GIMPS? Sort of like how yoyo@home and dnetc@home serve as BOINC wrappers for various distributed.net projects; they use a modified version of the distributed.net client to crunch work reserved and reported [I]en masse[/I] under a collective username at the mother project. This Mersenne BOINC project may be planning to do a similar thing with Prime95.

[QUOTE=mdettweiler;242421]Perhaps this project intends to serve as a "wrapper" platform for GIMPS? Sort of like how yoyo@home and dnetc@home serve as BOINC wrappers for various distributed.net projects; they use a modified version of the distributed.net client to crunch work reserved and reported [I]en masse[/I] under a collective username at the mother project. This Mersenne BOINC project may be planning to do a similar thing with Prime95.[/QUOTE]

That would certainly increase throughput around here!

[QUOTE=cheesehead;242418]My security software says that's a suspicious site and advises me to leave immediately.[/QUOTE]

Norton 360 didn't give me a peep. What software are you using?

[QUOTE=mdettweiler;242421]Perhaps this project intends to serve as a "wrapper" platform for GIMPS? Sort of like how yoyo@home and dnetc@home serve as BOINC wrappers for various distributed.net projects; they use a modified version of the distributed.net client to crunch work reserved and reported [I]en masse[/I] under a collective username at the mother project. This Mersenne BOINC project may be planning to do a similar thing with Prime95.[/QUOTE]

You know, I think that would be pretty cool. If done correctly, it could greatly benefit GIMPS because BOINC now has GPU support. However, we would still need a way to integrate it with PrimeNet. Perhaps George would be willing to share checksum-generating code/API keys with trusted third-party developers?

There is also a legal issue: if someone uses "Mersenne@Home" to find a prize-winning prime, would it still be considered a GIMPS discovery? If the discoverer used PrimeNet to get assignments and report results, and the wrapper was written completely independent of Prime95, a dispute could result.

[QUOTE=mdettweiler;242421]Perhaps this project intends to serve as a "wrapper" platform for GIMPS? Sort of like how yoyo@home and dnetc@home serve as BOINC wrappers for various distributed.net projects; they use a modified version of the distributed.net client to crunch work reserved and reported [I]en masse[/I] under a collective username at the mother project. This Mersenne BOINC project may be planning to do a similar thing with Prime95.[/QUOTE]

BTW: yoyo@home doesn't use a modified version of distributed.net client. It is the original one. There is just a wrapper arround it to handle the Boinc api. Don't know about dnetc@home, but I think it is the same there.
yoyo

[url]http://mersenneathome.net/apps.php[/url]

There are only Linux@Intel and Linux@AMD versions of the application - perfect match for me. I'd like to hear more from Sebastian on the implementation and availability of the source code.

I attached to the project and received a few tasks, some of which [URL="http://www.mersenne.org/report_exponent/?exp_lo=21005219&exp_hi=1&B1=Get+status"]left[/URL] [URL="http://www.mersenne.org/report_exponent/?exp_lo=21008011&exp_hi=&B1=Get+status"]me[/URL] [URL="http://www.mersenne.org/report_exponent/?exp_lo=21005767&exp_hi=&B1=Get+status"]head-scratching[/URL]!

...Why would I want to waste electricity on LL tests on Mp's that apparently weren't even trial factored there?! :max: they have trivial factors!

(the silver lining is ...well, they are only trying prime p's. That's a relief.)

[QUOTE=NBtarheel_33;242426]Norton 360 didn't give me a peep. What software are you using?[/QUOTE]ZoneAlarm

[QUOTE=retina;242420]Seems to be time to remove your useless security software then. There is not even any JS on that site (a rare thing these days).[/QUOTE]I've noticed that sites it's jumpy about are usually new. I should have said that previously.

The fine print says, "... This recommendation is based on how long the site has been around and the strength of its security. ... Site is not a known phishing or spyware distributor. Site was first registered on 11/03/10 (less than 3 month). Site is located in Poland, Eastern Europe."

[QUOTE=ixfd64;242429]There is also a legal issue: if someone uses "Mersenne@Home" to find a prize-winning prime, would it still be considered a GIMPS discovery?[/QUOTE]Yes, if it's running prime95, whose EULA begins with, "The full legal rules are at http://mersenne.org/legal/" and includes "At the time of this software release, the Electronic Frontier Foundation is offering a $150,000 award to the first person or group to discover a hundred million digit prime number. If you find such a prime with the software provided, GIMPS will claim the award and distribute the award according to the rules published at http://mersenne.org/legal/#awards"

As for Linux: mprime's EULA is around here somewhere ...

- -

No one has to share with GIMPS ... if they write their own prime-testing software.

Is it correct that the prize for the 10 million digit (Mersenne) prime
has not yet been officially awarded due to the fact that not all
smaller Mersenne Numbers have been fully tested?

If it has not yet been awarded, then it is not because the other numbers have not been fully tested. The prize was merely for finding the first 10 million digit prime, (as well as having it verified and published in a suitable journal and possibly some other conditions too).

[QUOTE=Batalov;242569]I attached to the project and received a few tasks.[/QUOTE]


I attached to the project and got this message:
...........................................................................
Requesting new tasks for CPU and GPU
Scheduler request completed: got 0 new tasks
Message from Server: Project has no tasks available
...............................................................................

What a pity!

Is there any prize in the offing for finding/proving the entire collection
of the first 50, 60, or 100 Mersenne Prime Exponents?

My Multi-Precison (Integer) Calculator does have a function for testing
Mersenne numbers in a loop, but it's too slow to reach 5-digit numbers (at this time).

[QUOTE=moebius;242662]I attached to the project and got this message:
...........................................................................
Requesting new tasks for CPU and GPU
Scheduler request completed: got 0 new tasks
Message from Server: Project has no tasks available
...............................................................................

What a pity![/QUOTE]
927 users attached at the moment (some familiar faces there!).
Most obviously have the credit of 0, because there are no Windows workers.

[QUOTE=davar55;242638]Is it correct that the prize for the 10 million digit (Mersenne) prime
has not yet been officially awarded due to the fact that not all
smaller Mersenne Numbers have been fully tested?[/QUOTE]

[url=http://www.eff.org/press/archives/2009/10/14-0]It's not correct[/url].

I forgot to ask earlier:

How does Mersenne@home choose the exponents it assigns as workunits?

Now, in its early stages, it would make sense to assign only exponents that had already been tested, as shown in the PrimeNet database -- i.e., already-DCed exponents -- so that the Mersenne@home admins can verify accuracy.

But once they have confidence in their accuracy of results, and demonstrates that to GIMPS's satisfaction, wouldn't it make sense for the Mersenne@home server to request bulk assignments from PrimeNet, deal them out to its users, then report results back to PrimeNet?

Batalov said in post #14 that they don't even trialfactor exponents first, so the site doesn't seem too trustworthy?

[QUOTE=ATH;247386]Batalov said in post #14 that they don't even trialfactor exponents first, so the site doesn't seem too trustworthy?[/QUOTE]It's not a matter of trust. It's a matter of being still in early stages.

From [URL]http://mersenneathome.net/forum_thread.php?id=8[/URL]

[quote]As you read on the homepage in the news section, the project is under implementation and the current analyzed samples are marked as "test", because we are currently testing application, server, work generator, validator, etc. So the current stage is not the actual search.[/quote]

[QUOTE=cheesehead;242418]My security software says that's a suspicious site and advises me to leave immediately.[/QUOTE]

I get the same message from both mersenne.org, mersenneforum.org and mersenneathome.net. Rarely does this mean anything but that they do not have a VeriSign certificate, SSL encryption or other security features. Most websites fail when checking for these and other security features.

[QUOTE=imwithid;247853]Rarely does this mean anything but that they do not have a VeriSign certificate, SSL encryption or other security features. Most websites fail when checking for these and other security features.[/QUOTE]As in post #15, ZoneAlarm warns because of the short time the websites had existed.

Would GIMPS accept M@H doublecheck results?
 

[EMAIL="Mersenne@Home"]Mersenne@Home[/EMAIL] seems to be starting out by doublechecking results from roughly 21 million upward. GIMPS is also doublechecking in this range. If [EMAIL="Mersenne@Home"]Mersenne@Home[/EMAIL] happens to doublecheck exponents that GIMPS has not had a chance to doublecheck, would (or should) GIMPS be willing to incorporate those results into its own database, and hence avoid running redundant doublechecks?