mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > No Prime Left Behind

Reply
 
Thread Tools
Old 2010-04-14, 05:54   #45
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

1012510 Posts
Default

Results have been processed for n=680K-747K. Results and primes have been confirmed against the usual places. No problems found.

Last fiddled with by gd_barnes on 2010-04-16 at 08:27
gd_barnes is offline   Reply With Quote
Old 2010-09-10, 07:39   #46
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

34·53 Posts
Default

Primes and results for this drive for n=747K-757K have been verified against the usual places. No problems found.
gd_barnes is offline   Reply With Quote
Old 2010-09-20, 20:34   #47
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

1012510 Posts
Default

Reserving n=760K-770K for port 5000.
gd_barnes is offline   Reply With Quote
Old 2010-11-06, 02:55   #48
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3×2,083 Posts
Default

As mentioned yesterday in the "Plans and goals for 2011" thread, I have moved n=765K-770K from LLRnet port 5000 to LLRnet port 3000. This still leaves a few (320 as of this writing) pairs in port 5000 below n=765K, but will ensure that port 3000 does not abruptly run out of work when its current 5th Drive workload runs out.

Note that primes found on this drive for n<770K should still be reported with your old "[your name], Srsieve, NPLB, LLR" proof codes. I'm currently working on extracting the range for this drive from the latest sieve file from the NPLB/PrimeGrid sieving drive; however, srfile takes a LONG time to convert a ~600MB ABCD format file to NewPGen format (over 1.2GB), and at this point the situation is getting sufficiently critical on port 3000 that Gary and I decided to just keep using the p=26T file for now. We will start using the new sieve file, sieved to p=~90T, for n>770K on this drive, so all primes found above that level should be reported using a proof code like that described above but with "Psieve" added as well. (Some of you already have such proof codes from the 5th Drive, but if not, you'll need to create a new one.)

I will be loading n>770K work from this drive into port 9000 from the new sieve file tomorrow as it has another ~32 hours of work left in it.

Last fiddled with by gd_barnes on 2010-11-06 at 05:26
mdettweiler is offline   Reply With Quote
Old 2010-11-06, 03:53   #49
MyDogBuster
 
MyDogBuster's Avatar
 
May 2008
Wilmington, DE

B1416 Posts
Default

Quote:
from LLRnet port 5000 to LLRnet port 5000.
Short move?????
MyDogBuster is offline   Reply With Quote
Old 2010-11-06, 05:26   #50
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

34·53 Posts
Default

Quote:
Originally Posted by MyDogBuster View Post
Short move?????
corrected
gd_barnes is offline   Reply With Quote
Old 2010-11-06, 08:40   #51
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

53548 Posts
Default

Quote:
Originally Posted by mdettweiler View Post
I have moved n=765K-770K from LLRnet port 5000 to LLRnet port 3000.
Pro:
- The heavy hitters don't have to change their configuration of the LLRnet ini-file.
-> Less work for them.

Contra:
- The result-file mixed up with results from two drives.
-> More work for me!

Also:
Please begin to create a table with the sieve-limits of the old and new sieve.
Using the new sieve-file for the other drives there'll be much confusion at what range I use which proof-code!

So far:
5th Drive (k=400-600):
n-range: 0 |<---old proof code--->| 945k |<---new proof code--->| oo

6th Drive (k=600-800):
n-range: 0 |<---old proof code--->| 770k |<---new proof code--->| oo
kar_bon is offline   Reply With Quote
Old 2010-11-06, 12:51   #52
vaughan
 
vaughan's Avatar
 
Jan 2005
Sydney, Australia

1010011112 Posts
Default

Good observation kar_bon. I was wondering about the proof code boundaries too. I still get confused by the terminology used here with Drive # and port = x000
vaughan is offline   Reply With Quote
Old 2010-11-06, 19:51   #53
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3×2,083 Posts
Default

@Karsten: yes, the proof code table you wrote is correct. The 5th Drive uses the new sieve for n>945K and the 6th Drive uses the new sieve file for n>770K. Good idea on making a master reference table for this--I'll do that as soon as I get the chance.

As for loading the 6th Drive work into port 3000: we considered the extra work it would entail, but realized it was unavoidable since we don't want all those resources currently on port 3000 to go idle. If it helps, you may want to try using the DB-driven results query page that I coded up a while back (I sent you the link in an email; let me know if you need it again); you can punch in any range of k and n and get the results (either in LLR or LLRnet format) irrespective of what port they were done in. That may simplify some of your processing tasks quite a bit.
mdettweiler is offline   Reply With Quote
Old 2010-11-07, 02:27   #54
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

141518 Posts
Default

We currently have about 6 hours worth of work left in port 3000 at current processing, so reserving 770K-780K for port 3000. Gary mentioned to me via PM that Vaughan is planning to stick around at least until Sunday so that should just about cover it, though of course I will continue to watch the server to load in more if necessary.

Once PCZ and Vaughan have dropped off port 3000 we'll let it dry out and start loading stuff into port 5000 instead. (Not to rush you guys or anything...we can of course use whatever help you can provide. )

Also, reserving 780K-782K for port 9000.

Just as a reminder: all ranges for n>770K are being loaded with the new sieve file. This means you need to use "Psieve" proof codes for primes found above this point. Please pardon my redundancy on this matter; better safe than sorry, since it can be a pain in the butt to have primes moved to a new proof code if they're reported incorrectly the first time around.

Last fiddled with by mdettweiler on 2010-11-07 at 02:30
mdettweiler is offline   Reply With Quote
Old 2010-11-08, 07:51   #55
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3·2,083 Posts
Default

Reserving 782K-785K for port 9000 and 785K-795K for port 3000.

Apologies for the downtime of ~20 minutes on port 9000. I should have loaded work earlier rather than let it get this low...you never know when Lennart will drop by.

In other news, Gary and I got to thinking that it would be easier to just keep port 3000 as our main 6th Drive server from here on out. Port 5000 dried out earlier today (thanks Karsten!) and at this point, everybody is on port 3000; with the 5th Drive done, we were going to shut down port 3000, but we can just as easily do that for 5000. So, to summarize: port 3000 will now be our 6th Drive LLRnet port. Karsten, hopefully this won't be too much trouble for you--we're hoping it will be easier than having port 3000 be only a temporary 6th Drive server as originally planned.
mdettweiler is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Team drive #7 k=800-1001 n=600K-1M gd_barnes No Prime Left Behind 127 2011-07-15 14:25
Team drive #5: k=400-600 n=600K-1M gd_barnes No Prime Left Behind 135 2010-11-23 14:27
Team drive #1: k=400-1001 n=333.2K-600K gd_barnes No Prime Left Behind 675 2009-02-24 16:37
Team drive #3: k=300-400 n=260K-600K gd_barnes No Prime Left Behind 255 2008-11-12 10:43
Team drive #4, 15 k's < 300 for n=600K-1M gd_barnes No Prime Left Behind 38 2008-10-22 16:20

All times are UTC. The time now is 07:10.

Wed May 27 07:10:27 UTC 2020 up 63 days, 4:43, 0 users, load averages: 1.12, 1.28, 1.34

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.