CRUS LLRnet server discussion

[gd_barnes]

Originally Posted by em99010pepe

Riesel server is without activity for almost two days in a row. Sierpinski is still with activity but I suppose we need to add more work...2100 k/n pairs still in knpairs.txt file but within 2 hours the server will prune it...

Are files needed for the Riesel server or is it just that no one is testing it now? If the latter, can you or someone put a core on it? It'll be a while before I have a machine that I can comfortably run LLRnet on or I would put one on it. (My main connected machine is an Athlon; other is a work laptop.)

Edit...never mind. Anon just answered my question.


Gary

[mdettweiler]

Originally Posted by gd_barnes

Are files needed for the Riesel server or is it just that no one is testing it now? If the latter, can you or someone put a core on it? It'll be a while before I have a machine that I can comfortably run LLRnet on or I would put one on it. (My main connected machine is an Athlon; other is a work laptop.)

Edit...never mind. Anon just answered my question.


Gary

Even though I'll be back on the Riesel server before too terribly long (right now I'm working on the new NPLB LLRnet server for a while), it would be really nice to have some help. Carlos, could you please put one of your machines (preferably a quad-core!) on the Riesel server? It wouldn't be too hard for you to rise up in the stats, too, because it's not getting too much attention.

[em99010pepe]

Originally Posted by Anonymous

Even though I'll be back on the Riesel server before too terribly long (right now I'm working on the new NPLB LLRnet server for a while), it would be really nice to have some help. Carlos, could you please put one of your machines (preferably a quad-core!) on the Riesel server? It wouldn't be too hard for you to rise up in the stats, too, because it's not getting too much attention.

I can shutdown the riesel server and test the numbers using the LLR manual client. I think it's 10% faster than the automatic client, any thoughts?

[mdettweiler]

Originally Posted by em99010pepe

I can shutdown the riesel server and test the numbers using the LLR manual client. I think it's 10% faster than the automatic client, any thoughts?

I don't think we should shut down the server--what happens then if someone suddenly decides to work on the Riesel server while it's down?

Also, as for the speed difference, I think it's actually more like 2-3% faster for the manual client (because LLRnet is based on LLR 3.5.1, and the speed increases came in 3.6). The reason why tnerual saw a 10% speed difference (if that's what you're thinking of) was because LLR 3.5.1 doesn't automatically convert to base 2, so he was accidentally doing PRP tests on base 16 numbers rather than LLR on base 2. When he converted the input file to base 2 (as I do with the LLRnet input files before sending them to you), he saw more realistic speeds.

BTW, I've attached the new files I grabbed for the Sierp. server.

[em99010pepe]

Let's wait more one or two days for some Riesel activity, then I'll see what I can do but I really prefer to run NPLB....

[mdettweiler]

Originally Posted by em99010pepe

Let's wait more one or two days for some Riesel activity, then I'll see what I can do but I really prefer to run NPLB....

Okay, that sounds good. Thanks!