Software/instructions/questions

[Mini-Geek]

Originally Posted by mdettweiler

(Actually, a batch file might work better here, since Windows will ask "terminate batch job? y/n" after LLRnet exits, thus allowing you to see at a glance whether there are any more k/n pairs to cancel without needing *really* sharp eyes... )

I don't find it hard at all to see, maybe it's different on Linux. On Windows it looks about like this when it's unreserving:

C:\Files\Prime\LLRNet>llrnet -c

C:\Files\Prime\LLRNet>Running 'init.lua' ...
Loading library 'basic' ...
Loading library 'client' ...
Loading library 'client_server' ...
LLR network client 0.9b7 started
Based on LLR version 3.5.0
Loading library 'win32' ...
LLRMain started
winloop thread
Cancelling : 1743/585603 (10000000000000:M:1:2:258)
StartServiceCtrlDispatcher returns 0
Successfully connected to server!
LOGGED ON
[04/29/09 08:27:34]
Result 1743/585603 succesfully sent to the server.
Result 1743/585603 succesfully sent to the server.

LOGGING OUT
Win32Exit

and like this when it's done:

C:\Files\Prime\LLRNet>llrnet -c

C:\Files\Prime\LLRNet>Running 'init.lua' ...
Loading library 'basic' ...
Loading library 'client' ...
Loading library 'client_server' ...
LLR network client 0.9b7 started
Based on LLR version 3.5.0
Loading library 'win32' ...
LLRMain started
StartServiceCtrlDispatcher returns 0
winloop thread
No more job to cancel !
Win32Exit

One note though: after running llrnet -c it doesn't make the C:\path\to\llrnet\> prompt reappear, but it does finish and return you to the command line as you'd expect, so you can retype llrnet -c or press up (to make it recall the last command you typed) and press enter and do that until you see "No more job to cancel !" instead of something like "Result 1743/585603 succesfully sent to the server."
Still, a batch might be easier for you, so do whatever you want to do.

[mdettweiler]

Yeah, that works too. I was mainly suggesting a batch file as an alternative to a GUI shortcut, which would open a new command window, run "llrnet -c", and when LLRnet dumps back to the command prompt it would exit.

Though, of course, users who are familiar with batch files are probably quite comfortable with using the command line directly, anyway, so this point is rather moot.

[Mini-Geek]

Oh right. Yeah, I agree that it'd be rather difficult to see whether there was something there or not in a direct shortcut, and in that case a batch would be better. Even to someone comfortable with the command line, a batch could make the process easier, especially with multiple cores and several reserved WUs (work units, i.e. numbers to test).

[Zachariassen]

Thanks guys - it worked
I delivered my first WU on port 5000, only to stop for the next 48 hours...but when I restart - I know how to change ports

Tore

[MyDogBuster]

Are there any instructions on how to get LLR 3.7.1c running on Linux machines? I have the executable but no help file.

[Lennart]

Originally Posted by MyDogBuster

Are there any instructions on how to get LLR 3.7.1c running on Linux machines? I have the executable but no help file.

-h

-m

/Lennart

[gd_barnes]

To be more specific: To start it, go to the command prompt and type:

./LLR -m

It will allow you to enter the same info. that you would from Window's LLR.

To run sr2sieve in Linux, all the commands are identical except that you'd put the ./ before everything on the command line, i.e.:

./sr2sieve -P xxex -i xxxx.txt (etc.)

[mdettweiler]

Originally Posted by gd_barnes

To be more specific: To start it, go to the command prompt and type:

./LLR -m

It will allow you to enter the same info. that you would from Window's LLR.

In addition, there are two other ways to start the client after you have initially loaded a file into it:

./llr will skip the menu and directly run the client in "quiet mode", i.e. with no screen output;
and
./llr -d will skip the menu and directly run the client, this time with screen output.

[henryzz]

Originally Posted by pb386

I'm not saying this is a good way to do it but:

On some machines I give them a particular k value (starting from the bottom of the range I am testing) & then I can leave them to get on with it with little intervention.

Other machines I take a k range (starting from the top of the range) and sort them by size & give the machines a couple of weeks work to do. This is a bit more labour intensive so I'm now wondering why I do this.

The ones I am waiting for are in the first category where a slow machine or large file is allocated to a machine & is just going to take a while to finish. I wondered about taking them off & finishing them more quickly but I can't really be bothered as they will finish soon enough anyway. I guess that's the point of the second method so you don't get 1 machine stuck with weeks of work when the others have finished.

i dont know how your machines are set up but would a personal prpnet server help you?

[pb386]

Originally Posted by henryzz

i dont know how your machines are set up but would a personal prpnet server help you?

Yes, I'm sure it would. Can some tell me how to go about setting it up?

Thanks.

[Mini-Geek]

Originally Posted by pb386

Yes, I'm sure it would. Can some tell me how to go about setting it up?

Thanks.

http://home.roadrunner.com/~mrodenkirch/prpnet_2.1.5.zip (this download includes the client, server, and source and is the most current release; I've heard a 2.2.0 is on its way, but is not available yet)
It's fairly self-explanatory. To use LLR/NewPGen files, replace the NewPGen header with "ABC $a*2^$b-1" and name it prpserver.candidates. If you need specific help with anything, ask and hopefully someone will know what to do.