mersenneforum.org  

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

Reply
 
Thread Tools
Old 2009-07-28, 16:08   #1
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

624110 Posts
Cool PRPnet servers for NPLB

The following is a list of all of NPLB's active PRPnet servers. The admins will keep this thread updated as often as possible.

A daily general report for all PRPnet servers is at www.noprimeleftbehind.net/prpnet.
An hourly/daily progress report for all NPLB servers is at www.noprimeleftbehind.net/stats/index.php?content=port.
Currently processing ranges last updated 2020-09-28 00:30 GMT.

NPLB PRPnet server #1:
Short identification: G1468
server = www.noprimeleftbehind.net
port = 1468
k's: 317/329/335/341/363/391/393/399 (8k mini-drive)
n-range: 1.575M-2M
currently processing at n= Complete
k's: 347/351/355/357/359/361/365/367/369/371/375/377/379/381/383/385/387 (17k maxi-drive)
n-range: 1.61M-2M
currently processing at n= Complete
k-range: 300-350 (15th Drive)
n-range: 2M-3M
currently processing at n= ~2.73M
server summary*: www.noprimeleftbehind.net:1468/all.html

NPLB PRPnet server #2:
Short identification: G9000
server: www.noprimeleftbehind.net
port: 9000
k-range: 400-600 (13th Drive)
n-range: 1M-1.02M, 1.025M-1.035M, 1.05M-1.7M
currently processing at n= ~1.59M
k-range: 800-1001 (7th Drive)
n-range: 790K-793K, 800K-810K, 820K-830K, 840K-870K, 880K-920K, 940K-950K, 960K-970K, 980K-995K, 998K-1M
currently processing at n= (complete)
server summary*: www.noprimeleftbehind.net:9000/all.html

NPLB PRPnet server #3:
Short identification: G2000
server: www.noprimeleftbehind.net
port: 2000
k-range: 600-1001 (14th Drive)
n-range: 1M-1.6M
currently processing at n= ~1.56M
server summary*: www.noprimeleftbehind.net:2000/all.html


*The user and team scores given by the server's pages are calculated directly by the server, and may not match up with our official stats at www.noprimeleftbehind.net/. This score is primarily useful for comparing different users on a specific server, and use a completely different formula for calculating score than what is used in the NPLB database. Therefore the scores are not equivalent to the database scores for primes and results.

Last fiddled with by gd_barnes on 2020-09-28 at 00:46 Reason: status update
mdettweiler is offline   Reply With Quote
Old 2009-08-03, 09:07   #2
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

72·11·19 Posts
Default

Max,

You have used the following link suffix in 3 places:

server_stats.html

Instead of this one:

server_status.html


Both of them were incorrect in the 1st post here and it was also incorrect over at CRUS. This is just a heads up. I've now corrected all 3 places.

I've also corrected the k-range in port 5000 from 3200-3400 to 3000-3400 per our email discussion.

I've now checked all of the related web pages. Everything was correct except for one thing: http://nplb-gb1.no-ip.org:2000/ is missing k=343, 359, 361, and 375. Can you please check into why those 4 k's are missing? If they have been omitted from the file that is loaded into the server, please add those k's to the server so that they aren't missed. If not, then a correction is needed with the interface from the server to that web page.


Thanks,
Gary

Last fiddled with by gd_barnes on 2009-08-03 at 09:21
gd_barnes is offline   Reply With Quote
Old 2009-08-03, 09:23   #3
vaughan
 
vaughan's Avatar
 
Jan 2005
Sydney, Australia

5×67 Posts
Default

I get "Problem loading page" messages for all the port:5000 links in your post Max.
When is this machine coming on-line?

Last fiddled with by vaughan on 2009-08-03 at 09:24
vaughan is offline   Reply With Quote
Old 2009-08-03, 09:29   #4
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

72·11·19 Posts
Default

Quote:
Originally Posted by vaughan View Post
I get "Problem loading page" messages for all the port:5000 links in your post Max.
When is this machine coming on-line?
See this statement above:

NPLB PRPnet server #1 (updated 2009-08-03 09:00 GMT):
Note: Server is not online yet. Links below will not work until server is brought online.

maintained by mdettweiler on gd_barnes machine


We are having difficulting with testing over at CRUS. Although the PRPnet servers themselves seem to be working mostly correctly, interfaces to the web pages seem to be having some problems both there and here.

Max, if you can see what is causing the problem on port G2000 here with the 4 k's not being shown on the one web page, fix it, and then start up G2000 again to where the web page shows correctly, then I'm good with us starting on G5000. Please remember to load the file by k-value. n=50K-75K to start with should be a good stress test. I'll throw at least 3 quads on it to give it a good stress. It will be interesting to see if my servers can handle such a load for extremely fast tests.


Gary
gd_barnes is offline   Reply With Quote
Old 2009-08-03, 09:48   #5
vaughan
 
vaughan's Avatar
 
Jan 2005
Sydney, Australia

5178 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
See this statement above:
Is it days or weeks away?

Should I move my machines from IB:9000 and co at noprimeleftbehind.net to PRPnet?

BOINC Simap is out of work now so once my cached tasks are crunched and reported I'll be back to full strength. I noticed that my I7-920 (Win 7 64bit) is pretty quick at BOINC ABCatHome.
vaughan is offline   Reply With Quote
Old 2009-08-03, 13:39   #6
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

72×11×19 Posts
Default

Quote:
Originally Posted by vaughan View Post
Is it days or weeks away?

Should I move my machines from IB:9000 and co at noprimeleftbehind.net to PRPnet?

BOINC Simap is out of work now so once my cached tasks are crunched and reported I'll be back to full strength. I noticed that my I7-920 (Win 7 64bit) is pretty quick at BOINC ABCatHome.
It depends on what you want to do. IB9000 has some of the biggest top-5000 tests that we have. Only G8000 is higher for servers. I think you said you like those the best.

Unfortunately at this time, these are the only two PRPnet servers and have/will be running the smallest tests that we have and are not top-5000. We're still in a testing and transition phase yet. G2000 that currently has work is double check. G5000 will be extremely small (n=50K-75K vs. 800K-900K in port IB9000) new tests and will take only a few seconds each to test. So you'll get to complete a lot of pairs and find a lot of small primes but their stats score here will be very small.

As for when G5000 will start up, I see it as 1-3 days away yet. Max is the one coordinating the running of the PRPnet servers so he could be more specific. I'm the checker and verifier of every little issue that I can find. I wanted to do the non-top-5000 and double-check work on the PRPnet servers until we defenitively have all of the smallest issues worked out. Then we'll begin transitioning perhaps 1-2 of our LLRnet servers at a time that are doing top-5000 tests to PRPnet servers as we get to nice round n-ranges.

Once G5000 starts up, we'll announce it here. In order to do a good stress test on it, we'll want lots of cores running it. I'll probably put several quads on it and if others can do the same, that would be a great stress test. The idea is to see how much load my servers and PRPnet can handle at the moment. With very fast tests, this will be a very big load as huge #'s of pairs are handed out and completed in short order.


Gary

Last fiddled with by gd_barnes on 2009-08-03 at 13:42
gd_barnes is offline   Reply With Quote
Old 2009-08-03, 13:45   #7
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

792 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Max,

You have used the following link suffix in 3 places:

server_stats.html

Instead of this one:

server_status.html


Both of them were incorrect in the 1st post here and it was also incorrect over at CRUS. This is just a heads up. I've now corrected all 3 places.
Whoops, thanks.

Quote:
I've also corrected the k-range in port 5000 from 3200-3400 to 3000-3400 per our email discussion.
Duh, should have remembered that one. Thanks.

Quote:
I've now checked all of the related web pages. Everything was correct except for one thing: http://nplb-gb1.no-ip.org:2000/ is missing k=343, 359, 361, and 375. Can you please check into why those 4 k's are missing? If they have been omitted from the file that is loaded into the server, please add those k's to the server so that they aren't missed. If not, then a correction is needed with the interface from the server to that web page.
Hmm...did we originally leave those k's out of the first-pass 3rd Drive? Because I just used the original 3rd Drive sieve file for this.
mdettweiler is offline   Reply With Quote
Old 2009-08-03, 13:57   #8
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

72·11·19 Posts
Default

Quote:
Originally Posted by mdettweiler View Post
Hmm...did we originally leave those k's out of the first-pass 3rd Drive? Because I just used the original 3rd Drive sieve file for this.
Hum. That's very possible but we do need to double-check all of them. I'll have to investigate and get back with you late this afternoon.

In the mean time, can you check the file that is in G2000 to make sure none of those k's are there? If none are there, the web page interfaces appear to be completely correct for this particular testing and we can start G5000 at any time.

Once G5000 starts handing out pairs, can you do a quick spot check to make sure there are a full 195 k's listed on the appropriate web page interface? The file that I sent you should have all k's in there for k=3010-3400.

BTW, I goofed in two ways on the range for G5000. We're only testing k=3010-3400, not 3000-3400 since Lennart has done 3000-3010. Also, we're only doing n=50K-400K not 50K-425K. We're stopping at 400K because that is where the mini-drive started. When people reserve individual k's on the k=3200-3400 range, they'll just need to do a little non-top-5000 work, which would also be the case if we loaded n=50K-425K so we may as well be consistent with k=3010-3200. I've corrected that (once again) in the 1st post. Sheesh. lol

Just a friendly reminder to be sure and only initially load n=50K-75K and in k-value sequence. Thanks.


Gary

Last fiddled with by gd_barnes on 2009-08-03 at 14:03
gd_barnes is offline   Reply With Quote
Old 2009-08-03, 14:28   #9
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

792 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Hum. That's very possible but we do need to double-check all of them. I'll have to investigate and get back with you late this afternoon.

In the mean time, can you check the file that is in G2000 to make sure none of those k's are there? If none are there, the web page interfaces appear to be completely correct for this particular testing and we can start G5000 at any time.
I've checked and it looks like all four of those k's were omitted from the file for n<500K. (Since the range we're doing right now in the server is 300K-350K, the k's don't have any incidence at all in the server.)

Per this thread, this is correct for the 3rd Drive. Nonetheless, yes, those k's should be doublechecked. However, we do have one problem: since we didn't originally test them the first time around, we don't have first-pass residuals for those k's. But, we really need to have two sets of residuals for each k/n pair in order to make a proper doublecheck. So here's what I'm proposing: we quickly re-do those k's to get "first-pass" residuals, then we add them to G2000 to be included in the doublecheck. We'll then end up with two residuals for each k/n pair, as needed.

Quote:
Once G5000 starts handing out pairs, can you do a quick spot check to make sure there are a full 195 k's listed on the appropriate web page interface? The file that I sent you should have all k's in there for k=3010-3400.

BTW, I goofed in two ways on the range for G5000. We're only testing k=3010-3400, not 3000-3400 since Lennart has done 3000-3010. Also, we're only doing n=50K-400K not 50K-425K. We're stopping at 400K because that is where the mini-drive started. When people reserve individual k's on the k=3200-3400 range, they'll just need to do a little non-top-5000 work, which would also be the case if we loaded n=50K-425K so we may as well be consistent with k=3010-3200. I've corrected that (once again) in the 1st post. Sheesh. lol

Just a friendly reminder to be sure and only initially load n=50K-75K and in k-value sequence. Thanks.
Okay, sounds good. I'll get the sieve file prepped and loaded as soon as possible, hopefully today.
mdettweiler is offline   Reply With Quote
Old 2009-08-04, 03:27   #10
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

72×11×19 Posts
Default

Quote:
Originally Posted by mdettweiler View Post
I've checked and it looks like all four of those k's were omitted from the file for n<500K. (Since the range we're doing right now in the server is 300K-350K, the k's don't have any incidence at all in the server.)

Per this thread, this is correct for the 3rd Drive. Nonetheless, yes, those k's should be doublechecked. However, we do have one problem: since we didn't originally test them the first time around, we don't have first-pass residuals for those k's. But, we really need to have two sets of residuals for each k/n pair in order to make a proper doublecheck. So here's what I'm proposing: we quickly re-do those k's to get "first-pass" residuals, then we add them to G2000 to be included in the doublecheck. We'll then end up with two residuals for each k/n pair, as needed.


Okay, sounds good. I'll get the sieve file prepped and loaded as soon as possible, hopefully today.

Ah, good. That's nice to hear that there's no interface problem. We're good to go on G5000.

Actually yes, we do have first pass residuals. You might remember that as the k=300-1001 range was nearing n=600K, I followed up the tail end by double checking other's efforts, finding one missing prime in the process, something that I'll also do as the 11th drive nears n=650K. I did it manually but still have the results saved off and available.

So, when you get a chance, can you load k=343, 359, 361, and 375 for n=300K-350K into port G2000? Assuming that this particular server naturally hands them out by size instead of entry order or k-value, then they should be handed out first up until we get to the current n=~330K depth. That would be a good thing.

That's nice to know that those interfacing web pages alert us to such things like that before they become bigger issues. Cool! :-)


Gary

Last fiddled with by gd_barnes on 2009-08-04 at 03:28
gd_barnes is offline   Reply With Quote
Old 2009-08-04, 03:52   #11
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

624110 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Ah, good. That's nice to hear that there's no interface problem. We're good to go on G5000.

Actually yes, we do have first pass residuals. You might remember that as the k=300-1001 range was nearing n=600K, I followed up the tail end by double checking other's efforts, finding one missing prime in the process, something that I'll also do as the 11th drive nears n=650K. I did it manually but still have the results saved off and available.

So, when you get a chance, can you load k=343, 359, 361, and 375 for n=300K-350K into port G2000? Assuming that this particular server naturally hands them out by size instead of entry order or k-value, then they should be handed out first up until we get to the current n=~330K depth. That would be a good thing.

That's nice to know that those interfacing web pages alert us to such things like that before they become bigger issues. Cool! :-)


Gary
Okay, sounds good. However, it seems that I don't have the 260K-600K sieve file for those four k's; I only have the sieve file for the 3rd Drive. Could you send me the part that I'm missing?

Thanks,
Max

Last fiddled with by mdettweiler on 2009-08-04 at 03:52
mdettweiler is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
PRPnet Servers for CRUS MyDogBuster Conjectures 'R Us 78 2020-08-05 17:17
LLRnet servers for NPLB kar_bon No Prime Left Behind 1343 2014-08-20 09:38
Public PRPNet Servers rogue Open Projects 26 2013-01-16 01:33
PRPNet servers down? opyrt Prime Sierpinski Project 13 2009-11-04 21:33
Servers for NPLB gd_barnes No Prime Left Behind 0 2009-08-10 19:21

All times are UTC. The time now is 20:19.

Wed Nov 25 20:19:38 UTC 2020 up 76 days, 17:30, 3 users, load averages: 1.73, 1.61, 1.60

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.