LLRnet servers for NPLB

[mdettweiler]

I've just sent another stats import to Adam--this includes all the work done during the rally (except for the C443 server, which has only had a relatively minuscule amount of work done on it so far--we'll wait to import that until it's gotten more progress, maybe when its range is complete).

[em99010pepe]

Anon,

I turned back on server #4 because I can only fetch candidates from work on port 443 and not on the others, is that ok with you? Soon I'll add another 4 cores to try to clean the server by myself.

Carlos

[gd_barnes]

Originally Posted by em99010pepe

Anon,

I turned back on server #4 because I can only fetch candidates from work on port 443 and not on the others, is that ok with you? Soon I'll add another 4 cores to try to clean the server by myself.

Carlos

Yes, that is great. We'll have ports 300 and 400 up past your range in C443 by next Tues. or Weds. so that works out well. I was going to move mine over to C443 when I get back next Weds. but may just leave them where they're at if you do C443.


Gary

[em99010pepe]

I'm trying to bring a friend with more 6 cores to the fun...let's see what happens.

Carlos

[mdettweiler]

Updated stats import data sent to Adam; this will bring the database up to date as of the results files for 9/2/08 (except C443, which we'll import all at once when its range is done).

[mdettweiler]

Hi all,

Carlos has been drying the C443 server lately, and it was on schedule to finish its current work by Sunday. We decided that we will be using C443 as one of the primary servers for our upcoming Drive #4 (k<300), so thus we won't be running it completely dry after all. However, since the sieved files won't be available for a little while yet, I've sent a sieved file to Carlos for k=289, n=520K-600K--this will get k=289, which is part of Drive #4 but is currently only searched to 520K, caught up to 600K at which point it will be searched with the rest of the k's in the drive.

Anyway, just figured I'd keep all everybody in touch with what's going on behind the scenes, so nobody gets confused if they connect to C443 and find k=289 work.

Anon

[em99010pepe]

I just took a look at the knpairs.txt of server C443 and we have less than 3400 candidates to test on the range 530.6k-532k (400<k<1001). That's good, by Sunday morning k=289 will be sent to clients or probably sooner if this server gets more help during the rally.

Carlos

[gd_barnes]

Originally Posted by Anonymous

Hi all,

Carlos has been drying the C443 server lately, and it was on schedule to finish its current work by Sunday. We decided that we will be using C443 as one of the primary servers for our upcoming Drive #4 (k<300), so thus we won't be running it completely dry after all. However, since the sieved files won't be available for a little while yet, I've sent a sieved file to Carlos for k=289, n=520K-600K--this will get k=289, which is part of Drive #4 but is currently only searched to 520K, caught up to 600K at which point it will be searched with the rest of the k's in the drive.

Anyway, just figured I'd keep all everybody in touch with what's going on behind the scenes, so nobody gets confused if they connect to C443 and find k=289 work.

Anon

Great proactive thinking Anon! Very good idea.

Originally Posted by em99010pepe

I just took a look at the knpairs.txt of server C443 and we have less than 3400 candidates to test on the range 530.6k-532k (400<k<1001). That's good, by Sunday morning k=289 will be sent to clients or probably sooner if this server gets more help during the rally.

Carlos

Thanks Carlos. That's great that we can get k=289 working ahead of time in preparation for the k<300 drive.

I'm still waiting for a response to a PM that I sent to Beyond about k=47 and 53. He hasn't been on since Aug. 22nd. He has those k's reserved at RPS and we are including them in the sieve. We'd like to include them in the drive but won't "pilfer" his reserved k's. I'm hoping to get the go ahead from him on those before the drive starts. Besides k=289, those k's have the lowest search limits of any k < 300 and are BY FAR the most behind their neighbors, something us NPLB admins like to fix.


Gary

[mdettweiler]

Originally Posted by gd_barnes

I'm still waiting for a response to a PM that I sent to Beyond about k=47 and 53. He hasn't been on since Aug. 22nd. He has those k's reserved at RPS and we are including them in the sieve. We'd like to include them in the drive but won't "pilfer" his reserved k's. I'm hoping to get the go ahead from him on those before the drive starts. Besides k=289, those k's have the lowest search limits of any k < 300 and are BY FAR the most behind their neighbors, something us NPLB admins like to fix.

I think Beyond's out of town--last I heard from him, he was away (business trip, if memory serves?). It must be somewhat long-term, since according to user profile he hasn't logged on to mersenneforum since August 22.

[em99010pepe]

Originally Posted by Anonymous

I think Beyond's out of town--last I heard from him, he was away (business trip, if memory serves?). It must be somewhat long-term, since according to user profile he hasn't logged on to mersenneforum since August 22.

I'll contact him on Ars forum, he posted there today...lol
Beyond's email sent by PM to Gary.

C443 is already sending k=289 because I took the last 800-900 candidates of 400<k<1001 to test manually. I gain 10% in speed with standard LLR.
It would be great to finish 289 up to 600k by the end of the rally.


Carlos

[gd_barnes]

Originally Posted by em99010pepe

I'll contact him on Ars forum, he posted there today...lol
Beyond's email sent by PM to Gary.

C443 is already sending k=289 because I took the last 800-900 candidates of 400<k<1001 to test manually. I gain 10% in speed with standard LLR.
It would be great to finish 289 up to 600k by the end of the rally.


Carlos

Thanks for Beyond's Email. I'll send him a note.

The rally started ~25 mins. ago. Good luck everyone! Time to do some serious crunching.


Gary