LLRnet and PRPnet servers for automated LLR

[Oddball]

Originally Posted by mdettweiler

Uh...you sure about those figures? By my reckoning (based on the ~3.5 days it took to cover 487.5K-489K), we have another 2.3 days left with the currently loaded work.

Yup, I'm sure. Remember, n=489K-490K had already been done before.

If you didn't load the n=487101-487500 work, we would have been out of k/n pairs now. The "count untested" column now says that 12,667 candidates haven't been tested, and all but one has n's from 487101-487500.

[mdettweiler]

Originally Posted by Oddball

Yup, I'm sure. Remember, n=489K-490K had already been done before.

If you didn't load the n=487101-487500 work, we would have been out of k/n pairs now. The "count untested" column now says that 12,667 candidates haven't been tested, and all but one has n's from 487101-487500.

Ah, of course--I forgot that 489K-490K was already done. That makes all the difference.

[Oddball]

Max, could you load k=100K-200K, n=484K-490K into the PRPnet server and k<100K, n=490K-491K into the LLRnet server? Both servers have dried out, and the files should be posted by gribozavr in this thread soon:
http://www.mersenneforum.org/showthread.php?t=12260

[mdettweiler]

Originally Posted by Oddball

Max, could you load k=100K-200K, n=484K-490K into the PRPnet server and k<100K, n=490K-491K into the LLRnet server? Both servers have dried out, and the files should be posted by gribozavr in this thread soon:
http://www.mersenneforum.org/showthread.php?t=12260

I don't see the files yet...am I missing something, or have they not actually been posted yet?

[Oddball]

Originally Posted by mdettweiler

I don't see the files yet...am I missing something, or have they not actually been posted yet?

They haven't been posted yet

[Oddball]

Originally Posted by mdettweiler

I don't see the files yet...am I missing something, or have they not actually been posted yet?

It's posted now. See here:
http://www.mersenneforum.org/showpos...00&postcount=5

[mdettweiler]

Originally Posted by Oddball

It's posted now. See here:
http://www.mersenneforum.org/showpos...00&postcount=5

All righty--the files are loading right now. LLRnet is all ready to go, and 484K-485K has been loaded into PRPnet; 485K-490K is still loading (the file is rather big and takes a while to upload), but in the meantime the server should have work available from 484K-485K.

[Oddball]

I checked the http://noprimeleftbehind.net/tps/ page and found something weird. It said,

Results submitted today by:
sm5ymt : 17774
odicin : 31
: 1
Total results submitted today: 17806

Does anyone know what happened with that ": 1" result?

[mdettweiler]

Originally Posted by Oddball

I checked the http://noprimeleftbehind.net/tps/ page and found something weird. It said,

Results submitted today by:
sm5ymt : 17774
odicin : 31
: 1
Total results submitted today: 17806

Does anyone know what happened with that ": 1" result?

The script that converts the results from PRPnet to LLRnet format sometimes barfs on slightly-malformed results lines. Since the (separate) script that does the daily stats reads the LLRnet format results, it is affected by this as well. The problem is here:

[2011-07-05 06:10:03 CDT] 161631*2^484489-1 received by Email: ****@****  User: sm5ymt  Client: 161 Program: llr  27A376A2E3F90409  
[2011-07-05 06:10:05 CDT] 128367*2^484490-1 received by Email: ****@****  User: sm5ymt  Client: 60 Program: llr.exe  4417C4BA0584AB32  
[2011-07-05 06:10:08 CDT] 151881*2^484489-1 received by Email: ****@****  User: sm5ymt  Client: 161 Program: llr  5843637031083387  
[2011-07-05 06:10:08] 155895*2^484489-1 received by Email: ****@****  User: sm5ymt  Client: 161 Program: llr  D0F1A5278A5D8425  
[2011-07-05 06:10:10 CDT] 119667*2^484488-1 received by Email: ****@****  User: sm5ymt  Client: 47 Program: llr  AA5336CC0DB4C6CD  
[2011-07-05 06:10:11 CDT] 134655*2^484490-1 received by Email: ****@****  User: sm5ymt  Client: 60 Program: llr.exe  BCC14265341684CD  
[2011-07-05 06:10:25 CDT] 130917*2^484490-1 received by Email: ****@****  User: sm5ymt  Client: 160 Program: llr  AAC07F03BDB0CB77  
[2011-07-05 06:10:28 CDT] 145845*2^484489-1 received by Email: ****@****  User: sm5ymt  Client: 103 Program: llr  F6579543046AA5E6

As you may notice, one of the lines is missing the "CDT" at the end of the timestamp, which was enough to throw off the conversion script, resulting in this LLRnet-format output:

user=sm5ymt
[2011-07-05 06:10:05]
128367*2^484490-1 is not prime.  Res64: 4417C4BA0584AB32  Time : 0.0 sec.
user=sm5ymt
[2011-07-05 06:10:08]
151881*2^484489-1 is not prime.  Res64: 5843637031083387  Time : 0.0 sec.
user=
[2011-07-05 06:10:08]]
 is not prime.  Res64:   Time : 0.0 sec.
user=sm5ymt
[2011-07-05 06:10:10]
119667*2^484488-1 is not prime.  Res64: AA5336CC0DB4C6CD  Time : 0.0 sec.
user=sm5ymt
[2011-07-05 06:10:11]
134655*2^484490-1 is not prime.  Res64: BCC14265341684CD  Time : 0.0 sec.

This particular error tends to happen every once in a while; I think it's occurred only once or twice in the past here at TPS, but it's shown up a number of times at NPLB. (There, it manifests itself as a result with a blank username in the hourly stats tables.) So far I've been fixing them all manually as they come up; I'll do that shortly with this one.

I really should add some code to the conversion script to handle these errors properly...shouldn't be too hard.

[lsoule]

Are the binaries or sources available for the server part of llrnet?

thanks,
Larry

[rogue]

Originally Posted by lsoule

Are the binaries or sources available for the server part of llrnet?

Out of curiosity, why? Does llrnet provide any functionality that is not already covered by PRPNet? Is it in some way easier to set up and use than PRPNet?