I see I guess I did not fully understand em99010pepe post
I see I guess I did not fully understand em99010pepe post
Yeah, as PCZ said.The plan is to not enable the additional manual data in the stats until at least sometime in June, after the PoTM push is long complete and you guys have had time to savor your victory.
BTW, Free-DC has a lot of manual work under their belt, too: Carlos and Beyond both put in quite a lot in the early days of the project. So, ROLP will by no means be the only team to benefit from inclusion of manual results in the stats.
What is this "manual work" ?
Reserve ranges manually by asking in the Forum.
Run the work offline. Send the results to Gary when finished.
http://www.mersenneforum.org/showthread.php?t=9890
Originally Posted by IronBits
Can somebody please explain all that in English?
I thought NPLB was running the llrnet software?
Can't anyone in the project explain anything clearly about how this works?
That article is just
Never mind. I don't know why I even bother to ask.
Why don't you ask your questions on the project's forum?
Most of us get by with doing the online llrnet server stuff that you are already doing.
If you want to try doing it the manual way, then you should go ask the pros.
There is no simple answer, and will require reading and studying on your part to learn something new.
It's not a Boinc project that's for sure.
I have spent ALL EVENING reading on that forum, and I'm more confused than ever, and know less about what the whole thing is about.Why don't you ask your questions on the project's forum?
Most of us get by with doing the online llrnet server stuff that you are already doing.
If you want to try doing it the manual way, then you should go ask the pros.
There is no simple answer, and will require reading and studying on your part to learn something new.
It's not a Boinc project that's for sure.
How can they have so many sub-projects that all seem to be doing different things? I mean - aren't they just looking for prime numbers like 1,3,5,7,11,13 but really big ones?
Yes, as IB said. It's mainly useful for users on dial-up or with machines that don't have internet access, though before we got LLRnet it was the primary mode of contribution for the project so it adds up to a pretty sizeable total score.
Edit: Ah, I just now saw your post requesting further clarification on this. (Man, those page cutoffs are rather annoying.) Here's how it works: we have ranges of k and n (the respective variables in the k*2^n-1 mathematical form for our numbers that we test for primality). The project, at this time, covers the entire range of k=300-2000, and n increases as we do more and more work.
Before the advent of LLRnet, the primary way people distributed tasks like this would be to have a forum post listing available ranges, with files containing the numbers to test (having been pre-screened by a process called "sieving" for candidates that had small factors and could be quickly removed) linked to. Each file would take anywhere from a few days to a couple weeks (depending on how big a range the file contained, and the speed of the user's machine). A user would post in the thread stating that they would like to reserve a certain file, download that file, and run it through a client application (the most commonly used one being LLR). The admins would edit the reservations table to mark that range as reserved by user x.
When the user was done with their range, they would take the results file produced by LLR, email it to the admins, and post in the forum saying that their range was complete. Rinse, lather, and repeat.
Once we set up an LLRnet server at NPLB, the manual system became rather outmoded. Still, it is often useful for offline computers and other such situations, so we continue to have manual work available. We integrate this with an LLRnet setup by "reserving" large ranges for the LLRnet servers, and then reporting back in the manual threads when a given range is complete.
Regarding why we have different "team drives" (subprojects, so to speak) for manual work: each of those team drives is for a different range of k and n. For example, Drive #5 covers k=400-600, n=600K-1M (currently testing at about n=670K). Last year we completed Drive #1, which did k=400-1001, n=333K-600K. Et cetera, et cetera. We generally have different LLRnet servers keyed to particular drives: for example, IB8000 (shorthand for nplb.ironbits.net port 8000) does the 10th Drive (and earlier on covered the now-complete 8th Drive, which the 10th Drive is an extension of). IB5000 does the 6th drive, IB4000 does the 5th, and...well, you get the picture.
Hope this explains things a bit!
Max
Last edited by Max Dettweiler; 05-05-2009 at 02:26 PM.
Excellent explanation .
Because we are still migrating to the new domain for your project, everywhere you see nplb.ironbits.net, can be changed to your new domain of noprimeleftbehind.net.
The old address will still work so nobody has to run around and change all their client configurations.
Eventually, nplb.ironbits.net will go away, say in about a few months, so if you are starting up new, use the new domain name.
If you are stopping and starting clients, and you only have a handle full, you could use the new domain name instead.
We did it
Opportunities
Rank -Team -----------------Score -----Average -Daily Gain -Days to Overtake
1 -Raiders of the Lost Primes 5,562,306 -25,865 ---60,122---- 23.88 Days
>>>>>>>>Go Team Go<<<<<<<<
Opportunities
Rank --Team ----------------Score----- Average-- Daily Gain -Days to Overtake
1-- Raiders of the Lost Primes -5,495,100 --28,087 -50,070---->> 31.88days
Outstanding Output Free-DC
Machines at home going down now.
I should be back up this evening.
Posting Permissions
Reply With Quote