Software/instructions/questions

[Mini-Geek]

I've started an LLR article on the English Wikipedia, using what I got auto-translated from Google as a base.
http://en.wikipedia.org/wiki/Lucas-Lehmer-Riesel
Feel free to help improve it. (and to help me understand what starting numbers to use)

[Mini-Geek]

Is there any way to "queue up" additional files in LLR? If a file is going to run out of things to do in the middle of the night, the computer would be idle for hours.
You could combine the remaining inputs of one file and the new inputs of the next file into one, but would this produce any bad effects as far as the results it generates goes?

[kar_bon]

you can put new k/n-pairs at the end of the LLR-input file (better when stopped) but they must have the same type of testing:
a file like this

309848651713:M:1:6:258
9577 25015
600000000000:M:1:16:258
17798 25004
160802390587:M:1:2:258
6435 126832
600000000000:P:1:16:257
19687 30002

won't work because there are different bases (6,16,2,16) and Riesel and Sierpinski mode mixed (:258 and :257).
LLR would take all k/n-pairs as Riesel-base 6 (first line!).
the only thing you have to do is to divide the results-file in their original portions (for example if you send results here for special ranges).
helps this?

[Mini-Geek]

Well, they'd all be from NPLB, so that wouldn't be a problem. I just wanted to check that it doesn't mess with the results file it makes. Is the results file it makes even necessary to submit anywhere (for NPLB, at least), or is it just useful as a list of the primes I've found?

[kar_bon]

please send all results to Gary, so if a doublecheck is necessary the residues can crosschecked with your results!
the primes you found are listed in an own file you filled in when started LLR under Test -> Input Data -> Output file: this file contains the first line of your testfiles (something like this: 5000000000000:M:1:2:258) and primes found, every line one prime in the form: k <space> n

[gd_barnes]

Originally Posted by Mini-Geek

Well, they'd all be from NPLB, so that wouldn't be a problem. I just wanted to check that it doesn't mess with the results file it makes. Is the results file it makes even necessary to submit anywhere (for NPLB, at least), or is it just useful as a list of the primes I've found?

I have had the same dilemma about files finishing in the middle of the night and I despise idle machines. It is perfectly fine if you want to combine input sieve files together.

Karsten, that would be an interesting request for Jean. Since LLR just ignores subsequent 'header' lines, perhaps he could make it read any header that it comes to and 'switch forms' as it goes. Sometimes I have wanted to test both Riesel and Proth primes in the same file.

About results files...I have debated long enough about whether to collect them with this project. I think Karsten and Anon agree that we should collect them. You can do this in one of 2 ways; it doesn't matter to me. You can attach them in a status post here or Email them to me at:
gbarnes017 at gmail . com. I'll post something more visible about it.

When posting or sending results, if you have combined more than one file in a single results file, just make a quick comment about it. This is less important if the ranges are contiguous but if they are not, one of the ranges can become 'misplaced' in my mass of files if I'm unaware of it ahead of time.


Thanks,
Gary

[gd_barnes]

Originally Posted by kar_bon

you can put new k/n-pairs at the end of the LLR-input file (better when stopped) but they must have the same type of testing:
a file like this

309848651713:M:1:6:258
9577 25015
600000000000:M:1:16:258
17798 25004
160802390587:M:1:2:258
6435 126832
600000000000:P:1:16:257
19687 30002

won't work because there are different bases (6,16,2,16) and Riesel and Sierpinski mode mixed (:258 and :257).
LLR would take all k/n-pairs as Riesel-base 6 (first line!).
the only thing you have to do is to divide the results-file in their original portions (for example if you send results here for special ranges).
helps this?

Mini-Geek and Karsten,

A light just came on about this. It IS possible to test different forms in a single file that is fed into LLR. If you check the README file, you'll see that if you use ABC format, you can do more than one form:

Example:

ABC $a*2^$b$c
599 250001 -1
599 250001 +1

which is
599*2^250001-1
599*2^250001+1

These are the ABC forms accepted by LLR (from the readme file):

Fixed k forms for k*b^n+/-1 : %d*$a^$b+%d, %d*$a^$b-%d, %d*$a^$b$c
Fixed b forms for k*b^n+/-1 : $a*%d^$b+%d, $a*%d^$b-%d, $a*%d^$b$c
Fixed n forms for k*b^n+/-1 : $a*$b^%d+%d, $a*$b^%d-%d, $a*$b^%d$c
Any form of k*b^n+/-1 : $a*$b^$c+%d, $a*$b^$c-%d, $a*$b^$c$d

You can even change bases:

ABC $a*$b^$c$d
599 2 250001 -1
599 4 250001 +1

which is

599*2^250001-1
599*4^250001+1

I remember reading about this at some point many months ago but completely forgot about it. I don't know that it applies to this project but it might help at Conjectures 'R Us and is good to know for future reference.


Gary

[Mini-Geek]

EDIT: Disregard, I noticed the GUIremote option. That's what was causing the problem, and I'm leaving the rest of the post here in case anyone else has that problem.
Edit2: I just checked, and the download package has that line uncommented. Did I stupidly change that line and then forget about it, or did it come messed up originally, then get fixed later?



(Not sure if this is the right place)
Why is my LLRnet GUI not connecting? It just endlessly tries to connect, obviously getting nowhere.
LLRnet is running, and I can open a console to it just fine, but the GUI doesn't connect.
Here's my llr-client/guiconfig.txt files so you can see if a setting is wrong.

--
-- LLRNet - network part of LLR
--
-- (C) 2004-2005 Vincent Penne
--
-- Released under GNU LIBRARY GENERAL PUBLIC LICENSE
-- (See file LICENSE that must be included with this software)
--
--
-- CLIENT CONFIGURATION FILE
--
-- WARNING: comments start with two dashes 
--

-- server location
server = "crus.dynip.telepac.pt"
port =   300

-- username
username = "Mini-Geek"

-- uncomment the following line to allow remote control of your client
-- see also options below to configure who can connect
--GUIremote = 1

-- if several clients are running on the same machine, give each of them
-- a different GUI port to listen to
GUIport = 6999

-- GUI clients connection configuration
-- by default, allow only local connections
GUImask  = "255.255.255.0"
GUIallow = "192.168.1.0"
-- example allowing any connection on a local network 192.168.1.*
--GUImask  = "255.255.255.0"
--GUIallow = "192.168.1.0"

-- GUI password, it is strongly recommended to change it 
-- if you don't restrict GUI clients to local connections
GUIpassword = "hi"

-- name of the file containing current job
workFile = "workfile.txt"

-- name of the file containing results to send to the server
tosendFile = "tosend.txt"

-- number of time to retry to send results before going for next WU.
-- set it to 0 for no retries, for user with no permanent connection
-- this is the suggested setting as it will save them precious 
-- minutes of useless retries (the program sleep for one minute
-- between two retries)
sendRetries = 0

-- uncomment line below to work only once
--once = 1

-- uncomment the line below for verbose output
--llrVerbose(1)

-- to ask several WU job to the server at a time, set the variable below
-- to something greater than 1 (but no more than 100)
WUCacheSize = 2


-- Uncomment the two functions below to launch some script each
-- time you respectively find a prime, and submit a new result

--function OnPrime(t, k, n)
--   local script = "onprime.sh"
--   if WIN32 then
--      script = "onprime.bat"
--   end
--   execute(format("%s %d %d %d", script, t, k, n))
--end

--function OnSubmit(t, k, n, result, residue)
--   -- notice : result is 0 for a prime, -2 for non prime, and -1 on error
--   local script = "onsubmit.sh"
--   if WIN32 then
--      script = "onsubmit.bat"
--   end
--   execute(format("%s %d %d %d %d %s", script, t, k, n, result, residue))
--end


-- WIN32 specific settings

-- uncomment the line below to hide the tray icon on startup
--hideTrayIcon = 1

-- change this value to a unique name for each client you want to run
-- as service on the same machine
serviceName = "LLRnet"


-- MY TEST
--server = "localhost"
--port =   7000
--username = "zig"

I had tried it first with the GUImask/allow settings default, but it didn't work that way, which is why I tried to change it to be like that.

clientsDesc = {
[1]={
  ["password"]="hi",
  ["port"]=6999,
  ["name"]="Local",
  ["address"]="192.168.1.19",
  ["connected"]=308,
},
[2]={
  ["password"]="hi",
  ["port"]=6998,
  ["name"]="Local",
  ["address"]="localhost",
  ["connected"]=236,
},
}

I tried to add the GUI for the second core's setup. Neither worked.

[mdettweiler]

Mini-Geek, are you using Linux? I've never been able to get the GUI and the LLRnet client to see eye-to-eye on Linux, though it's always worked fine on Windows.

[Mini-Geek]

I'm using Windows.

I think I must have commented out that line, not realizing what I was doing.
I'm not certain because I downloaded and briefly played around with LLRnet a while ago (that's where my two in the stats came from), and I don't remember for sure what I did or didn't change, but I do remember the GUI didn't work last time, when I did the two.

Trying to remember back, yeah I think I commented out the line thinking it meant a remote connection only, not referring to a local connection which just happens to technically be done like a remote.

[mdettweiler]

Originally Posted by Mini-Geek

I'm using Windows.

I think I must have commented out that line, not realizing what I was doing.
I'm not certain because I downloaded and briefly played around with LLRnet a while ago (that's where my two in the stats came from), and I don't remember for sure what I did or didn't change, but I do remember the GUI didn't work last time, when I did the two.

Trying to remember back, yeah I think I commented out the line thinking it meant a remote connection only, not referring to a local connection which just happens to technically be done like a remote.

Okay--now that I read your post here, as well as read through your original post again, I think I see what might be the problem keeping you from controlling both cores with one GUI. In llr-guiconfig.txt, both of your listings are named "Local". I would suggest you try renaming one of them to "Local2" or something else like that.

Not that I have ever tried it myself, just a hunch.