CRUS LLRnet server discussion

[mdettweiler]

Originally Posted by em99010pepe

It will take some time until I figure it out how to do it but meanwhile you can manual generate the stats. As a backup I'll send you by email the current results files I have for both servers. Please check your email.

Okay, thanks.

[mdettweiler]

Well, I posted a message over at Riesel Sieve asking what the rejected.txt file from the LLRnet server contains (since they developed LLRnet), and here's the answer I got:

Originally Posted by b2 (on Riesel Sieve forums)

I believe....and Bryan can correct me on this....it has been a LONG TIME... but rejected.txt holds the information returned from a client where the following takes place.

Say you download a wu of k*2^n+/-1 and your time limit is 7 days....you are doing other stuff and it takes you 9 days to actually crunch it. Now after 7 days the LLRNet server times out your wu and reissues it to Joe ServerFarm and he completes the wu in 1 day....so Joe beat you with a returned result by one day and when you turn yours in...bam..it is rejected because you no longer are the user on record. However...these can be saved and used as double checks and match residuals.

I think this is how we designed it...I can't remember....I'm getting old http://www.rieselsieve.com/forum/ima.../icon_wink.gif

Lee

I guess that explains the mystery of the rejected.txt file.

[kar_bon]

only for stats:

my client says: 17798*2^389576-1 is prime!

so 17798*16^97394-1 is prime.

PS: my doublecheck: 8899*2^389577-1 is prime!

[em99010pepe]

user=kar_bon
[01/31/08 21:37:16]
17798*2^389576-1 is prime! Time : 438.0 sec.

My double check using LLR 3.7.1c

17798*2^389576-1 is prime! Time : 455.254 sec.

[gd_barnes]

SUHWEEET!!

Karsten is smokin' hot both at NPLB and CRUS!!

[em99010pepe]

Riesel Server will be down for a few minutes, just found a bug on the knpairs.txt.....

[gd_barnes]

I'll leave it up to you guys as to whether you remove k=17798 from your server files. Since we're in top-5K range, who knows...we might find another for k=17798.

I'll remove the k from all posted files for n>=110K in a while.

[em99010pepe]

Riesel Server is up. When removing k=17798 from knpairs.txt srfile detected an error on it. The file had repeated k/n numbers to test. Very sorry. Now everything is ok.

Carlos

EDIT: Anon, please check the riesel-base16-knpairs-100K_110K.txt file you sent me.

[mdettweiler]

Originally Posted by em99010pepe

Riesel Server is up. When removing k=17798 from knpairs.txt srfile detected an error on it. The file had repeated k/n numbers to test. Very sorry. Now everything is ok.

Carlos

EDIT: Anon, please check the riesel-base16-knpairs-100K_110K.txt file you sent me.

Aha! Apparently I messed up. I accidentally inserted one of the individual files twice while combining files from the drive into one big file. Whoops! Sorry about that.

Thanks for fixing it!

Besides that, though, we can all celebrate: our first LLRnet prime! Now we just need to find one for the Sierp. side.

[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...

[mdettweiler]

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...

Hmm...probably the loss of activity on the Riesel side is due to the fact that I was the main person working on the Riesel side, and I took my machine off of CRUS LLRnet for a while while working on some manual-LLR files. I'll be back before too long, though.

As for the Sierp. server--okay, I'll go get some more work for it. Wow, it really eats those k/n pairs quickly!