LLRnet/PRPnet rally August 12th-19th

mdettweiler · 2010-07-29, 18:07

Hi everyone,

Based on comments provided after our last rally, we've decided to try something new this time: a week-long rally!

We figured that if people are going to go to all the trouble of moving their cores over to the rally servers, why do it all over again to take them back off in only 48 hours? PrimeGrid has done week-long rallies in the past with success, so it's not an entirely new idea; we're hoping it will go over well here too.

The rally will begin Thursday, August 12 at 7:00 PM GMT and finish 168 hours later on Thursday, August 19 at 7:00 PM GMT. For those of you who can't convert GMT to local time

, that's 3:00 PM EDT and 2:00 PM CDT. Last time we had to do crazy stuff with PRPnet port 9000 since LLRnet reports local time and PRPnet GMT, but since then PRPnet has been changed to report local time as well, so the official start/end times should be pretty straightforward.

As last time, this rally will be focusing on the 5th Drive, our priority for the rest of the year (or until we finish it to n=1M, which will hopefully come first). That means LLRnet port 3000 and PRPnet port 9000, both of which should be in the vicinity of n=850K-900K during the rally.

LLRnet config info.:
server = "www.noprimeleftbehind.net"
port = 3000

PRPnet config info.:
server = "www.noprimeleftbehind.net"
port = 9000
Or, in terms of a prpclient.ini config line:
server=G9000:100:1:noprimeleftbehind.net:9000

Be sure and change your userID to whatever ID you prefer. Most people use their ID here at mersenneforum. It is also best to set the cache at from around 2 to 10 pairs.

Info. for downloading the newest LLRnet client can be found at LLRnet supports LLR V3.8! (LLRnet2010 V0.73L).

Info. for downloading the newest PRPnet client can be found at PRPnet.

As a point of interest for those who might be new to NPLB and for others who haven't done any testing with us for a while, we have some milestones that we would like to shoot for in the near future. Please refer to our primes graph and drive progress as well as to our "Countdown meter". NPLB's goal all along has been to fill in holes in the prime search ranges for Riesel base 2. We'd like to get the dots in that k=300-1001 range fully extended to n=1M by year end 2011 so that they are closer to where k<300 (worked by RPS) is at. In targeting k=400-600 for the 3rd consecutive time in this rally, we have set a relatively aggressive goal to have it tested completely to n=1M by year end 2010.

Last time we had a very close race between ROLP and PrimeSearchTeam, the latter ending up winning by a mere 194 k/n pairs.

We also had participation from both AMD Users and Team-Goobee.org, coming in 3rd and 4th place respectively. And despite dutifully posting announcements at Free-DC, they went AWOL for the second rally in a row. Come on guys, what's up with our former repeated rally champs?

Let's make this rally even more exciting than the last one!

Max

mdettweiler · 2010-08-11, 03:17

Just a reminder for everyone: I will be heading out of town early tomorrow morning (as detailed earlier in the News thread). This will quite likely be my last check of the forum before leaving. I have confirmed that I will almost definitely NOT be able to use the internet during my trip, so until I return Thursday (8/12) evening, I will not be available to deal with questions/problems/etc. I realize this means I will not be available during the crucial time of the days before and of the start of the rally. If any issues crop up, post here in this thread. Gary can answer policy questions regarding the rally and the project in general, and questions about the LLRnet servers; Karsten can answer questions about the LLRnet client; and Dave can add people to teams in the database. Most of our PRPnet participants should be already familiar with the client from either here or PrimeGrid, so I don't anticipate any problems with that. If all else fails, we have a few PRPnet experts here (Lennart, and Mark himself if he's monitoring this thread) who can probably give you a hand.

There is enough loaded in PRPnet port 9000 to last at least until I get back (probably plenty longer). Gary is in charge of loading and keeping track of work in LLRnet port 3000.

As for my own clients, my dualcore's cores are working on various manual jobs, and will join the rally automatically as those complete. (One in particular may last until a ways into the rally, but that shouldn't make any noticeable impact in the rally throughput.) There's a good chance that my quad will be turned off while I'm away; either way, it will be doing personal CRUS work until I get back at which point I will switch it to PRPnet port 9000. I should only miss the first few hours of the rally.

Have fun and try not to wreck the PRPnet server while I'm gone!

(Feel free to mess with the LLRnet server--Gary can handle that.

)

AMDave · 2010-08-12, 13:46

Kick off is in just under 6 hours.
I will be asleep then (I hope).

Don't forget to post your team sign-up requests in the "Teams: Join a team" thread in this forum:
http://www.mersenneforum.org/showthr...=10826&page=10

I'll check the stats in about 7 hours from now and then head to the office.
I have no access from there, so any outstanding team sign-ups will be done in about 16 hours time.

If you haven't fired up a client yet; "What are you waiting for? Permission?"

gd_barnes · 2010-08-12, 22:45

The rally has been running for nearly 4 hours now. Good luck everyone!

mdettweiler · 2010-08-13, 03:45

Well, I'm back--I missed the rally entirely thusfar due to a strange spontaneous-rebooting problem on my dualcore picking 8:00 PM yesterday, of all times, to begin manifesting itself (again--it's happened before but not for a while). I'm not sure of the exact cause but I do know from last time this happened that it does not cause bad results. I shut down the computer and kept it off for about 30 minutes, which in the past has done the trick; hopefully it's gone now.

My quad has been turned back on and I will, if I can, switch it over to port 9000 tonight as well. That will make for a total of 5 cores of mine participating in the rally (to be upped to 6 as soon as my CRUS S288 work finishes on the dualcore).

I see nobody's found any primes yet in the rally, which I suppose is not surprising since it's only been going for ~9 hours. Let's get some nice juicy medium-large ~850K-870K primes rolling!

Sloth · 2010-08-13, 17:32

The PRPnet server is saying too many connections

Mini-Geek · 2010-08-13, 17:40

Quote:

Originally Posted by Sloth

The PRPnet server is saying too many connections

I'm getting this too. First got it at 12:08 (~30 minutes ago). Edit: It's working again, as of 12:43 (5 minutes ago from this edit)

gd_barnes · 2010-08-13, 17:47

I just now stopped and restarted the server. It now seems to be handing out and accepting pairs again,

I'll keep an eye on it. Max, if you can help with that as you have time, that would help also. Do you know why there would be a maximum # of connections?

Edit: I just did one more thing to prevent a problem that we had with the last rally: I renamed the prpserver.log file to prpserver-0813.log. It was already at 755 MB. I'll keep doing that at least once/day. Something about its size caused the server to have problems last time.

mdettweiler · 2010-08-13, 18:10

Quote:

Originally Posted by gd_barnes

I just now stopped and restarted the server. It now seems to be handing out and accepting pairs again,

I'll keep an eye on it. Max, if you can help with that as you have time, that would help also. Do you know why there would be a maximum # of connections?

Edit: I just did one more thing to prevent a problem that we had with the last rally: I renamed the prpserver.log file to prpserver-0813.log. It was already at 755 MB. I'll keep doing that at least once/day. Something about its size caused the server to have problems last time.

Thanks--that happens every once in a while and neither Mark or I are really sure why. What happens is the server keeps opening sockets with increasing frequency but never closing them--this used to just go on until infinity, hence why Mark implemented a maximum number of communications. I think ours is 1000 or something like that, so we should never reach that number in normal operation.

At this time I'm not so sure that the log file is the cause of the issue--it may be, it may not be. Renaming the log file on a daily basis surely won't hurt. (I usually try to do that periodically anyway, but once daily during the rally should be a good precaution.)

Meanwhile, I just tried to VNC into jeepford now and everything seemed frozen at first. Before I got to the chance to do any investigating, the VNC server crashed. The PRPnet server still responds to queries, so it looks like everything is OK--I just can't get on. I'm going to try a few more tricks and see if I can get it fixed. Restarting the computer would surely do the trick, but that's a pain and not exactly something we'd want to do during a rally.

kar_bon · 2010-08-13, 18:10

LLRnet server offline, too? Getting no new pair!

OK, working again!

mdettweiler · 2010-08-13, 18:29

Quote:

Originally Posted by kar_bon

LLRnet server offline, too? Getting no new pair!

OK, working again!

Yeah, I just talked with Gary and it looks like jeepford just spontaneously rebooted. There's some strange stuff going on here...probably more to do with my attempts to get back on than with the PRPnet "too many connections" issue.

Anyway, everything seems OK now--Gary is restarting the servers as I speak. The rally servers were the first ones to come back online.

2010-07-29, 18:07	#1
mdettweiler A Sunny Moo Aug 2007 USA (GMT-5) 3×2,083 Posts	LLRnet/PRPnet rally August 12th-19th Hi everyone, Based on comments provided after our last rally, we've decided to try something new this time: a week-long rally! We figured that if people are going to go to all the trouble of moving their cores over to the rally servers, why do it all over again to take them back off in only 48 hours? PrimeGrid has done week-long rallies in the past with success, so it's not an entirely new idea; we're hoping it will go over well here too. The rally will begin Thursday, August 12 at 7:00 PM GMT and finish 168 hours later on Thursday, August 19 at 7:00 PM GMT. For those of you who can't convert GMT to local time , that's 3:00 PM EDT and 2:00 PM CDT. Last time we had to do crazy stuff with PRPnet port 9000 since LLRnet reports local time and PRPnet GMT, but since then PRPnet has been changed to report local time as well, so the official start/end times should be pretty straightforward. As last time, this rally will be focusing on the 5th Drive, our priority for the rest of the year (or until we finish it to n=1M, which will hopefully come first). That means LLRnet port 3000 and PRPnet port 9000, both of which should be in the vicinity of n=850K-900K during the rally. LLRnet config info.: server = "www.noprimeleftbehind.net" port = 3000 PRPnet config info.: server = "www.noprimeleftbehind.net" port = 9000 Or, in terms of a prpclient.ini config line: server=G9000:100:1:noprimeleftbehind.net:9000 Be sure and change your userID to whatever ID you prefer. Most people use their ID here at mersenneforum. It is also best to set the cache at from around 2 to 10 pairs. Info. for downloading the newest LLRnet client can be found at LLRnet supports LLR V3.8! (LLRnet2010 V0.73L). Info. for downloading the newest PRPnet client can be found at PRPnet. As a point of interest for those who might be new to NPLB and for others who haven't done any testing with us for a while, we have some milestones that we would like to shoot for in the near future. Please refer to our primes graph and drive progress as well as to our "Countdown meter". NPLB's goal all along has been to fill in holes in the prime search ranges for Riesel base 2. We'd like to get the dots in that k=300-1001 range fully extended to n=1M by year end 2011 so that they are closer to where k<300 (worked by RPS) is at. In targeting k=400-600 for the 3rd consecutive time in this rally, we have set a relatively aggressive goal to have it tested completely to n=1M by year end 2010. Last time we had a very close race between ROLP and PrimeSearchTeam, the latter ending up winning by a mere 194 k/n pairs. We also had participation from both AMD Users and Team-Goobee.org, coming in 3rd and 4th place respectively. And despite dutifully posting announcements at Free-DC, they went AWOL for the second rally in a row. Come on guys, what's up with our former repeated rally champs? Let's make this rally even more exciting than the last one! Max

2010-08-11, 03:17	#2
mdettweiler A Sunny Moo Aug 2007 USA (GMT-5) 14151₈ Posts	Heading out of town early tomorrow Just a reminder for everyone: I will be heading out of town early tomorrow morning (as detailed earlier in the News thread). This will quite likely be my last check of the forum before leaving. I have confirmed that I will almost definitely NOT be able to use the internet during my trip, so until I return Thursday (8/12) evening, I will not be available to deal with questions/problems/etc. I realize this means I will not be available during the crucial time of the days before and of the start of the rally. If any issues crop up, post here in this thread. Gary can answer policy questions regarding the rally and the project in general, and questions about the LLRnet servers; Karsten can answer questions about the LLRnet client; and Dave can add people to teams in the database. Most of our PRPnet participants should be already familiar with the client from either here or PrimeGrid, so I don't anticipate any problems with that. If all else fails, we have a few PRPnet experts here (Lennart, and Mark himself if he's monitoring this thread) who can probably give you a hand. There is enough loaded in PRPnet port 9000 to last at least until I get back (probably plenty longer). Gary is in charge of loading and keeping track of work in LLRnet port 3000. As for my own clients, my dualcore's cores are working on various manual jobs, and will join the rally automatically as those complete. (One in particular may last until a ways into the rally, but that shouldn't make any noticeable impact in the rally throughput.) There's a good chance that my quad will be turned off while I'm away; either way, it will be doing personal CRUS work until I get back at which point I will switch it to PRPnet port 9000. I should only miss the first few hours of the rally. Have fun and try not to wreck the PRPnet server while I'm gone! (Feel free to mess with the LLRnet server--Gary can handle that. ) Last fiddled with by mdettweiler on 2010-08-11 at 03:19

2010-08-13, 03:45	#5
mdettweiler A Sunny Moo Aug 2007 USA (GMT-5) 3·2,083 Posts	Well, I'm back--I missed the rally entirely thusfar due to a strange spontaneous-rebooting problem on my dualcore picking 8:00 PM yesterday, of all times, to begin manifesting itself (again--it's happened before but not for a while). I'm not sure of the exact cause but I do know from last time this happened that it does not cause bad results. I shut down the computer and kept it off for about 30 minutes, which in the past has done the trick; hopefully it's gone now. My quad has been turned back on and I will, if I can, switch it over to port 9000 tonight as well. That will make for a total of 5 cores of mine participating in the rally (to be upped to 6 as soon as my CRUS S288 work finishes on the dualcore). I see nobody's found any primes yet in the rally, which I suppose is not surprising since it's only been going for ~9 hours. Let's get some nice juicy medium-large ~850K-870K primes rolling! Last fiddled with by mdettweiler on 2010-08-13 at 03:46

2010-08-13, 17:47	#8
gd_barnes May 2007 Kansas; USA 3³·5·7·11 Posts	I just now stopped and restarted the server. It now seems to be handing out and accepting pairs again, I'll keep an eye on it. Max, if you can help with that as you have time, that would help also. Do you know why there would be a maximum # of connections? Edit: I just did one more thing to prevent a problem that we had with the last rally: I renamed the prpserver.log file to prpserver-0813.log. It was already at 755 MB. I'll keep doing that at least once/day. Something about its size caused the server to have problems last time. Last fiddled with by gd_barnes on 2010-08-13 at 17:57

2010-08-13, 18:10	#10
kar_bon Mar 2006 Germany 2³·3·11² Posts	LLRnet server offline, too? Getting no new pair! OK, working again! Last fiddled with by kar_bon on 2010-08-13 at 18:16

2010-08-12, 13:46	#3
AMDave Jan 2006 deep in a while-loop 2·7·47 Posts	Kick off is in just under 6 hours. I will be asleep then (I hope). Don't forget to post your team sign-up requests in the "Teams: Join a team" thread in this forum: http://www.mersenneforum.org/showthr...=10826&page=10 I'll check the stats in about 7 hours from now and then head to the office. I have no access from there, so any outstanding team sign-ups will be done in about 16 hours time. If you haven't fired up a client yet; "What are you waiting for? Permission?"

2010-08-12, 22:45	#4
gd_barnes May 2007 Kansas; USA 3³×5×7×11 Posts	The rally has been running for nearly 4 hours now. Good luck everyone!

2010-08-13, 17:32	#6
Sloth Mar 2006 1011110₂ Posts	The PRPnet server is saying too many connections

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
LLRnet/PRPnet rally April 4th-11th	mdettweiler	No Prime Left Behind	55	2011-04-25 09:35
LLRnet/PRPnet rally January 3rd-10th	mdettweiler	No Prime Left Behind	48	2011-01-12 10:14
LLRnet/PRPnet rally Oct. 27th-Nov. 3rd	mdettweiler	No Prime Left Behind	33	2010-12-24 19:16
LLRnet/PRPnet rally June 4th-6th	gd_barnes	No Prime Left Behind	61	2010-07-30 17:28
LLRnet server rally 400<k<1001 August 8-10	mdettweiler	No Prime Left Behind	66	2008-08-11 03:00