Automated primality testing with LLRnet for CRUS

[em99010pepe]

I need more testers....the servers have small candidates to test. On a P4 3.0 GHz kn pair is tested in less than 500 seconds from Sierpinski side and less than 300 seconds on Riesel side.


Hi all,

We now have an official Conjectures 'R Us LLRnet server! (Actually we have two of them. More on that later...)

With LLRnet, you can perform primality tests (LLR/Proth/PRP) automatically. No need to mess with reserving/reporting ranges and all that stuff. No need to worry about whether your machine will run out of work in the middle of the night or while you're on vacation. Just set it and forget it!

LLRnet is very simple to set up--in less than 5 minutes you can be all set and crunching away! It's perfect for users who just want to donate their spare CPU cycles to a worthy cause without adding a degree of complexity to their computer use. (That's not to say that it isn't great for advanced users, too. ) It's also great for work computers, since you can set it up on a computer and not have to worry about having it run out of work while it's not accessible. (We've set up one of our LLRnet servers on port 443 for just this reason. If this whole "port" thing is just going in one ear and out the other, though, here's essentially what it means: you're still able to use LLRnet at work even if your company has a restrictive firewall. Sorry, there's currently no way to run LLRnet through an HTTP or SOCKS proxy server, however.)

Here's all you need to do:

1. Download the LLRnet client software for your operating system. LLRnet is available for Windows, Linux, and FreeBSD, in two versions: the "stock" version that only works with work for bases that are powers of 2 (but has a nice GUI), and the SR5 version, which works for all bases (but doesn't have the GUI). If you'd rather not have to worry about using different LLRnet installs for different bases, get the SR5 version. If the GUI is a must-have, though, you'll want to use the stock version for bases that are powers of 2.

Click one of the following links to download LLRnet:

Windows (SR5 version) (http://bugmesticky.googlepages.com/LLRnet-win32-sr5-CRUS.zip) <<--------- Recommended for most users
Linux (SR5 version) (http://bugmesticky.googlepages.com/LLRnet-linux-sr5-CRUS.tar.bz2)<<------------ Recommended for most users

Windows (http://bugmesticky.googlepages.com/LLRnet-win32-CRUS.zip) (stock version, power-of-2 bases only)
Linux (http://bugmesticky.googlepages.com/LLRnet-linux-CRUS.tar.bz2) (stock version, power-of-2 bases only)
FreeBSD (http://bugmesticky.googlepages.com/LLRnet-freebsd-CRUS.tar.gz) (stock version, power-of-2 bases only)

2. Extract the files from the archive you just downloaded. C:\llrnet is a good choice (or /home/username/llrnet for Linux users).

3. Navigate to the folder where you extracted LLRnet. Open up the llr-clientconfig.txt file in your favorite text editor. You'll notice a little ways down from the top a line that says the following:

port = 443

This is the port on the LLRnet server that the client will connect to. Currently, the LLRnet server is set up to answer on two different ports, each with a different type of numbers. Here's what they currently are:

443 - Sierpinski Base 16 (Team Drive #1) (default)
444 - Riesel Base 16 (Team Drive #2)

Simply change the port value to work on different numbers. Please note that if you're behind a corporate firewall, you might only be able to use port 443. (Corporate firewalls usually allow access on port 443.)

A little farther down you'll notice another line that says the following:

username = "nobody"

Change "nobody" to your mersenneforum.org username. For example, I would set this to "Anonymous". This ensures that numbers you work on are credited properly to you in the stats that will be published periodically. (If you'd rather crunch anonymously, i.e. not have your crunching associated with your mersenneforum.org username, simply set this to something different. You can actually punch in anything you want here, though most users will probably want to punch in your mersenneforum username.)

Save the file, and close out your text editor.

4. Doubleclick on the file named "llrnet.exe" to start LLRnet and get crunching! (On Linux and FreeBSD this will be just "llrnet".) You'll see an LLRnet icon appear in your system tray if you're running Windows, and, if you're using the "stock" version of LLRnet, you can right-click on the icon and click a button to pop up a GUI from which you can monitor your work. If you're using the SR5 version (the recommended version since it works for all bases), you can click another button on the right-click menu to open up a console and monitor your work from there. (Note: When you want to close out the console window, don't just close out the window. Instead, right-click on the LLRnet icon, and click "detach console". Otherwise, you'll close out the LLRnet program, which you probably don't want to do.) You can also use the console with the stock version, just like with the SR5 version. Linux users won't see a system tray icon, but they will get a GUI that pops up if they're using the stock version. (If you want to get the GUI back up after closing it out, navigate to your LLRnet folder and run the remote-gui.sh file.) To see the console, you'll want to start LLRnet from a command window--navigate to the LLRnet folder, and type the command "./llrnet" (without quotes). (With the console open, simply press Ctrl-C to stop LLRnet.)

5. If you want to have LLRnet run automatically when you start your computer, simply right-click the LLRnet system tray icon and click "Add LLRnet service". Or, if you'd rather only run LLRnet when your username is logged on, click the Start button, and navigate to All Programs, then double-click on the item named "Startup". A folder window will open. In a separate window, navigate to your LLRnet folder. Drag the llrnet.exe icon, with the right mouse button, to the Startup folder, and let go of the mouse button. Choose "Create Shortcut" from the menu that appears.

For Linux, how you set a program to run at startup depends on your distribution. Just post a message here and someone will try to help.

That's all!

Please post any questions/comments here, or send a private message to me (Anonymous).

Special thanks to Carlos (em99010pepe) for graciously hosting the LLRnet server for us!

Let's get crunching!


Q: Can I run LLRnet on a multi-core system and still utilize all the cores?
Yes, though LLRnet by default can only utilize one core. Thus, you'll have to run multiple copies of LLRnet. Simply, when installing, extract the archive you downloaded into as many folders as you have cores (such as c:\llrnet, c:\llrnet2, etc.). Then, configure the llr-clientconfig.txt for each core (you have to configure each installation separately--and if you're using the stock version, make sure to change the GUIPort value to something different on everything except the first installation. You'll also have to change this value in the llr-guiconfig.txt for the installations that have been set to a different GUIPort.), and run the llrnet.exe (or "llrnet" on Linux/FreeBSD) in each installation. You'll see as many LLRnet icons in your system tray as you have cores--you can control each core's LLRnet separately. (You can also configure one core's GUI, if you're using the stock LLRnet, to control all the other cores, too, though that's a little bit more of an advanced topic--you can either try to figure it out on your own, or just post here begging for help. )

[em99010pepe]

NPLB (No Prime Lef Behind Project) server is up and running on crus.dynip.telepac.pt port 300.

NPLB is searching for all 400 < k <= 1001 for n=260K-320K and n=320K-333.2K at the same time. LLRNet with 321.6K-322.6K range, less than two minutes per candidate. We really need to clean these small ranges. PLEASE HELP.

[em99010pepe]

BTW, for future stats please use your username as:

username = "TeamName_yournickname"

My case:

username="Free-DC_em99010pepe"

[IronBits]

Ok, I'll lend a hand for a bit.

[em99010pepe]

Ok, I'll lend a hand for a bit.

Please use port 300 and not 443, faster wu's. Thanks.

[IronBits]

ok,
The server refused your new result :
either someone else computed it already,
either the server is now configured to
work on other numbers.
The server refused your new result :
either someone else computed it already,
either the server is now configured to
work on other numbers.

[em99010pepe]

ok,
The server refused your new result :
either someone else computed it already,
either the server is now configured to
work on other numbers.
The server refused your new result :
either someone else computed it already,
either the server is now configured to
work on other numbers.

That's normal because you changed ports...don't worry, I'll clean the mess from my side.

[IronBits]

Ah, not a problem, just did it once.
How long do you suppose it will take to clear these out?
I assume you have knowledge of my wu completions.
Are there any Stats or we just cleaning up?

[em99010pepe]

Ah, not a problem, just did it once.
How long do you suppose it will take to clear these out?
I assume you have knowledge of my wu completions.
Are there any Stats or we just cleaning up?

Right now the server has 11490 k/n pairs (565 tested as I post) on it but I still need to add something like 122000 pairs. The latter number will be reduced because people are reserving ranges manually. I'm almost finishing my current ranges and I'll jump to this effort not with LLRNet but with the LLR manual client.
This phase, n=320.0k to n=333.2k, can be done in less than a month. Yes, I can see your timings but they don't reflect the real time but the sum of the time necessary to test the number plus the time the client needs to connect to the server. If your wu cache is set to 10 the timings will be 10 times higher.

At this page (http://www.rieselprime.org/nplb/NPLB_TeamDrive2.htm) you can see the progress. Stats are manually, still not automatic, I'll post here in a few days.

We are not just cleaning it up the range, we are filling the gaps for those k's. It's an important task for prime world.

The automatic stats, like the ones you have with PSP PRP and RS5 PRP, are in discussion with Bok and ltd...

Carlos

PS( I'm tired of running a DC project just for the stats but I know it's necessary to bring more people to this effort)

EDIT: I get 90 seconds per candidate on a overclocked quad-core to 2.9GHz. You are getting 250 seconds but they are not the real timings.

[IronBits]

Thanks for the explanations.
140-160 sec is what I'm seeing, but I'm doing more than just this project.
@ 2.5GHz quad core.

[em99010pepe]

This NPLB LLRnet server has tested 896 numbers so far.
Those 896 numbers were divided up among various users as follows:

Free-DC_IronBits: 745 numbers tested, 83% of total
kar_bon: 128 numbers tested, 14% of total
Free-DC_em99010pepe: 21 numbers tested, 2% of total
MiniGeek: 2 numbers tested, 0% of total

--------------------
Team Stats

Free-DC: 766 numbers tested, 85% of total
[Users w/o a team]: 130 numbers tested, 15% of total

Reservation board is as follows:

http://www.rieselprime.org/nplb/NPLB_TeamDrive2.htm

As seen there I am using the manual client.

Carlos Eduardo

[IronBits]

ok, let me know when we are done. :)

[em99010pepe]

Yes I will. Meanwhile I added one spare core to LLRNet, I don't want you to go far away in the stats, and another one for the manual LLRing...

[em99010pepe]

This NPLB LLRnet server has tested 2906 numbers so far.
Those 2906 numbers were divided up among various users as follows:

Free-DC_IronBits: 2351 numbers tested, 80% of total
Free-DC_em99010pepe: 223 numbers tested, 7% of total
kar_bon: 213 numbers tested, 7% of total
Anonymous: 117 numbers tested, 4% of total
MiniGeek: 2 numbers tested, 0% of total

----------------------
Team Stats

Free-DC: 2574 numbers tested, 89% of total
[No Team]: 332 numbers tested, 11% of total

[IronBits]

At this page (http://www.rieselprime.org/nplb/NPLB_TeamDrive2.htm) you can see the progress. Stats are manually, still not automatic, I'll post here in a few days.
I'm not seeing my name in lights on that page.
Am I crunching under your name and using your 'proxy' port to crunch work for you automatically so I don't have to reserve anything? :thumbs:

[em99010pepe]

I'm not seeing my name in lights on that page.
Am I crunching under your name and using your 'proxy' port to crunch work for you automatically so I don't have to reserve anything? :thumbs:

You found two primes: 693*2^321959-1 and 795*2^322014-1.

Your are using LLRNet to crunch so you don't have to reserve anything and not crunching under my name.

Stats by number of primes found in here (http://www.rieselprime.org/nplb/NPLB_Scoring.htm).

[em99010pepe]

833*2^322530-1 is prime! (IronBits, LLRNet)

[em99010pepe]

A 24 hour rally will be held this Friday or Saturday for k/n pairs of port 300. Soon I'll post more details...we really need to clean the lower ranges.

[IronBits]

I'll be there with two Q6600s, one 32bit, the other 64bit. :)
Hot diggity dog, found me another Prime :D

[Paratima]

Ah'm in with one 3GHz AMD core. :cheers:

[Paratima]

Using Port 300.

Seems to be averaging about 165 seconds to find a non-prime. :beep:

[em99010pepe]

I already moved one Q6600 and three P4 3.0GHz...
Just a note, the primes found won't get into Top 5000, we are cleaning the lower ranges. Next rally will be in Top 5000 range...

Carlos

[em99010pepe]

Thanks Paratima for your help. Meanwhile I added another core to LLRNet and added more k/n pairs to the server.
Hey Helix what about you? Fast wu's on the server...

[em99010pepe]

LLRNet stats update (http://www.rieselprime.org/nplb/NPLB_Scoring.htm).

A 24 hour rally will take place next Saturday starting at 7 PM GMT.

[IronBits]

This Saturday the 9th ?

[Paratima]

What, if anything, do we do special? :confused:

[em99010pepe]

This Saturday the 9th ?

Yes.

[em99010pepe]

user=Free-DC_IronBits
[02/08/08 14:12:33]
473*2^325256-1 is prime!

[IronBits]

:rock:

[em99010pepe]

user=Free-DC_Petey
[02/08/08 16:45:35]
789*2^325375-1 is prime! Time : 266.0 sec.

[Petey]

:woot:

My first ever prime.

[em99010pepe]

:woot:

My first ever prime.

How many candidates did you test? Check GUI window.

[Petey]

Hi Carlos,

I'm not running the Gui but a quick word count on the lresults file says I have done 86 tests so far.

Petey

[em99010pepe]

Hi Carlos,

I'm not running the Gui but a quick word count on the lresults file says I have done 86 tests so far.

Petey

Lucky indeed...
All scores can be checked here (http://www.rieselprime.org/nplb/NPLB_Scoring.htm).

[Paratima]

Still yesterday's results.

How often does she update?

If you're going to invite stats whores to come in & play, you gotta provide stats!

[em99010pepe]

Still yesterday's results.

How often does she update?

If you're going to invite stats whores to come in & play, you gotta provide stats!

Please read my posts...You can always move to another project. I'm here for the goal and not for the stats.

Carlos

[Paratima]

All of that selfless dedication is certainly a fine thing.

Another thing entirely is wanting to know how you're progressing on a project and how well your pharm is performing.

When can we expect the next stats update?

[em99010pepe]

All of that selfless dedication is certainly a fine thing.

Another thing entirely is wanting to know how you're progressing on a project and how well your pharm is performing.

When can we expect the next stats update?

In an hour. I just sent the results to the stats man...

[em99010pepe]

user=Free-DC_IronBits
[02/08/08 14:33:23]
705*2^325274-1 is prime!

[em99010pepe]

Stats were updated. Stats are updated once a day, like dnet.

[em99010pepe]

user=Free-DC_Petey
[02/08/08 22:47:09]
741*2^325661-1 is prime! (confirmed prime)

[Petey]

Now I'm just being greedy :D

[em99010pepe]

Karsten post:


for information:
between 1:30 AM GMT and 6:30 AM GMT (08 Feb 08) there were about 3800 k/n-pairs handed out from the server.

current rate is about 50 k/n-pairs per 5 minutes.

current k/n-pair at n=327270 (and about n -> n+4 every 5 minutes).

so about 68000 pairs left on server -> 113 hours of work (~50 pairs per 5 min)

[em99010pepe]

The rally will start within 8 hours.

[Paratima]

Well, Carlos, you ignored my question of whether there's anything special required of us and I won't be around when it starts at noon EST. :bonk:
I do most of what I laughingly call "work" on weekends. Actually, almost entirely on Saturdays.

So I'm starting a couple of clients now and I'll look in again when I get home tonight. Good luck!

[IronBits]

I'm already on board and cranking. Not waiting for the party to start :rotfl:
Extract the contents someplace or install it to \NPLB

Make a dir NPLB - edit the config file
Make a dir NPLB2 - copy contents of NPLB to it
Make a dir NPLB3 - copy contents of NPLB to it
Make a dir NPLB4 - copy contents of NPLB to it
@echo off
cd \NPLB
start llrnet
cd \NPLB2
start llrnet
cd \NPLB3
start llrnet
cd \NPLB4
start llrnet

:D

[em99010pepe]

Well, Carlos, you ignored my question of whether there's anything special required of us and I won't be around when it starts at noon EST. :bonk:


I only asked for your help, nothing special, I think it's well exposed on my previous posts.

Well IronBits, I started yesterday and I'll keep until I dry the server out, check the stats!

Thanks alpha, Paratima, Petey and IronBits.

[IronBits]

Nice output :thumbs:

[em99010pepe]

15 mins. A stats update will follow shortly...

[em99010pepe]

15 mins. A stats update will follow shortly...

Only now updated. Next one tomorrow at 9 pm GMT.

[IronBits]

Primes all over the place!!!
How the heck did they miss so many?

Stats goodness: http://www.rieselprime.org/nplb/NPLB_Scoring.htm

[em99010pepe]

Primes all over the place!!!
How the heck did they miss so many?

Stats goodness: http://www.rieselprime.org/nplb/NPLB_Scoring.htm

Sometimes primes are miss, this case a few k's are being double-check, that's why you have the confirmed primes, other's k's have never been tested. Sometimes people just start searching for big primes and leave the lower ranges to process, they start at a higher n, it's like starting one thing from the middle and not from the beginning...so many variables.


Detailed Search Information:
At first there are three different ranges for 400 < k < 1001 for searching with the following conditions:

* 333200 < n < 600000 to find many Top-5000 primes as possible (Forum Team Drive #1)
* 320000 < n < 333200 to fill the n = 20k (320k to 340k) chunk for reporting primes to PrimeSearch (Forum Team Drive #2)
* 260000 < n < 320000 to fill the lowest untested range at PrimeSearch (Forum Team Drive #2)

The whole range from n = 260000 to 600000 was presieved up to p = 5 T).


After cleaning 2nd goal, I'll put Top 5000 candidates into the server but I'll put my machines manually on 3rd goal.

[em99010pepe]

user=Free-DC_IronBits
[02/10/08 08:47:05]
833*2^329490-1 is prime!

[em99010pepe]

user=Free-DC_IronBits
[02/10/08 10:34:19]
987*2^329652-1 is prime!

[em99010pepe]

Thanks Beyond for joining us.

[Paratima]

Make a dir NPLB - edit the config file
Make a dir NPLB2 - copy contents of NPLB to it
Make a dir NPLB3 - copy contents of NPLB to it
Make a dir NPLB4 - copy contents of NPLB to it
@echo off
cd \NPLB
start llrnet
cd \NPLB2
start llrnet
cd \NPLB3
start llrnet
cd \NPLB4
start llrnet
:D
Hey, IB
Did all that before your post. Won't run multiple copies on my 6600+ X2 Winders XP boxen. I get:
"Could not run the GUI server, make sure another LLRnet client is not running"
So I've got a couple running, but have to do it on different machines. :looney: :confused:

[IronBits]

-- minimal configuration
server = "CRUS.dynip.telepac.pt"
port = 300

username = "Free-DC_IronBits" <--- put yours here, or not ;)

-- change this value to a unique name for each client you want to run as service on the same computer (multiple cores/processors)
serviceName = "LLRnet-CRUS"

-- if several clients are running on the same computer, give each of them a different GUI port to listen to (multiple cores/processors)
GUIport = 6999

GUImask = "255.255.255.255"
GUIallow = "127.0.0.1"
GUIpassword = "llrnet" <--- change it if you like (I don't use it)
workFile = "workfile.txt"
tosendFile = "tosend.txt"
sendRetries = 0

WUCacheSize = 5

-- end of config file

When you are done, be sure to change WUCacheSize to 1
WUCacheSize = 1

Then when it is down to just that one WU, stop the client.

Add this line to the config file and save it
once = 1

Start the client one last time and it will finish the last WU, upload the result and terminate. :)

[em99010pepe]

user=Free-DC_IronBits
[02/10/08 16:55:15]
973*2^330363-1 is prime!

[Paratima]

Thanks, Mr. Bits!

You're all right. I take back half the bad things I said about you. :D

[em99010pepe]

user=Free-DC_IronBits
[02/10/08 17:53:02]
705*2^330472-1 is prime!

[Beyond]

Thanks Beyond for joining us.

Anything math related is good, your welcome.

1 question, what is the limit for cache size? set at 10 ATM, but runs thru them pretty quickly.

[IronBits]

I noticed it uploads as soon as it finishes a wu, so I really don't think a cache is helpful.
Unless you want to make sure you get more WUs to work on than anyone else has in their cache,
to ensure you complete more work than anyone else, after the server runs out of WUs to hand out. ;)

[Beyond]

I noticed it uploads as soon as it finishes a wu, so I really don't think a cache is helpful, unless you want to make sure you get more WUs than anyone else to work on after the server runs out . ;)


Right!

In a word "Comcast". (read as being down 30% to 40% of the time) So lost processing time on a regular basis if no cache. :swear:

[IronBits]

Try WUCacheSize = 100 and see what happens ;)

[Beyond]

:thumbs:

[em99010pepe]

Try WUCacheSize = 100 and see what happens ;)

Only if you are going to test them....

The rally will be over in 20 min.

[IronBits]

Damn the rally! ;)
We can still work on the whole thing right?
You're not pulling the Server offline are you?

[em99010pepe]

Damn the rally! ;)
We can still work on the whole thing right?
You're not pulling the Server offline are you?

Yes you can still work on this thing, server up and running. I'm going to finish my reserved ranges....

[em99010pepe]

Is it possible to create a dedicated sub forum for NPLB (No Prime Left Behind Project) project?

I already added a few Top 5000 candidates and soon I'll need to post a few threads explaining how to submit primes, how the score system works, how to set up the LLRNet client, etc...

[em99010pepe]

i tried to be fast.....
so here is the first glance of our 24h-race on LLRnet:

www.rieselprime.org/nplb/NPLB_Race20080902_24h.htm

this link in not yet inserted in the NPLB-homepage (will be done) and the scorings for LLRnet from the race are not yet included in the 'Overall'-LLRnet stats page (next to do too).

please have a look and perhaps more suggestions to implement.
i will do:
- listing primes found in race

:|party|:

[em99010pepe]

user=Free-DC_Beyond
[02/11/08 05:58:21]
939*2^331229-1 is prime!

[em99010pepe]

We are almost in the Top 5000 territory (n>349.2K) within LLRnet server. In this case all primes must be submitted to top-5000 prime:


Instructions for submitting a top-5000 prime:

Go to http://primes.utm.edu/primes/home.php.

Step 1: For people who have not submitted top-5000 primes previously, create a prover account:
1. Select 'Submit' on the lower left side under 'Join in'.
2. Select the 'click here to create a prover-account' link in the middle of the page.
3. Fill out the form and submit it. You will be assigned a prover account.

Step 2: Create a proof code:
1. Go to the home page in the link above and select 'Index' on the left side under 'Provers'.
2. Next to 'search prover-accounts', type yours and press enter.
3. Click on your prover account (there may be one or several) and press enter.
4. Towards the bottom, click on 'Create a New Proof-Code' and press enter.
5. If necessary in the little pop-up box, type in your user name (prover account) and password and press enter.
6. You should now have a list of proof programs. Select LLR.
7. In the space below all the programs, type in 'PrimeSearch' for the project and 'Srsieve' for the sieving software. Separate each selection by a comma. (Obviously if you did an individual-k search on your own, then the software may vary.)
8. You should now have a new proof code and can submit the prime. Example...L442.

Step 3: Submit the prime:
1. Go to the home page in the link above and select 'Index' on the left side under 'Provers'.
2. Next to 'search proof-code', type your new code from step 2 and press enter.
3. Towards the bottom, next to 'Submit primes using this code as', click on your name (prover account) and press enter.
4. If necessary in the little pop-up box, type in your user name (prover account) and password and press enter.
5. You should see a big free-form box. Type in your prime (no spaces needed) and click 'Press here to submit these prime(s)'.
6. A verification screen will come up. If the prime is correct, click 'Press here to complete submission'.

Soon my strategy will be different, in case you miss the prime from your machines I'll send you a PM warning you to submit the prime and I won't post here the prime until Top 5000 confirms it as a prime number. Any doubts?

[em99010pepe]

user=Free-DC_Beyond
[02/12/08 00:07:54]
845*2^332076-1 is prime!

[em99010pepe]

user=Free-DC_Beyond
[02/12/08 09:10:22]
933*2^332466-1 is prime!

[alpha]

I'll have some of whatever Beyond is crunching. :idea:

[em99010pepe]

user=Free-DC_IronBits
[02/12/08 12:56:55]
595*2^332639-1 is prime!

[em99010pepe]

user=Free-DC_Beyond
[02/12/08 21:57:23]
597*2^333048-1 is prime!

Thanks for creating a dedicated forum for this project.

[em99010pepe]

We are now testing top 5000 candidates, stay alert!

Stats were updated (http://www.rieselprime.org/nplb/NPLB_Scoring.htm).

[em99010pepe]

IronBits found the first Top 5000 LLRnet prime. Please check your PM with instructions on how to submit the prime.

[Beyond]

IronBits found the first Top 5000 LLRnet prime. Please check your PM with instructions on how to submit the prime.


WTG Ironbits

:cheers: :thumbs: :banana:

[em99010pepe]

I just found the second one.

user=Free-DC_em99010pepe
[02/13/08 17:07:35]
573*2^350426-1 is prime!


IB, do you need help? Feel free to contact me via PM or email.

Carlos

[em99010pepe]

Beyond, check your private messages.

[IronBits]

IronBits found the first Top 5000 LLRnet prime. Please check your PM with instructions on how to submit the prime.
:woot:
I read your PM and now I'm :looney: trying to figure it out...

Re-read it, went there... that's :looney: ...
Way too many hoops to jump through just to put a name on a prime. :swear:

You can have it, claim it as your own if you like. :smoking:

[em99010pepe]

:woot:
I read your PM and now I'm :looney: trying to figure it out...

Re-read it, went there... that's :looney: ...
Way too many hoops to jump through just to put a name on a prime. :swear:

You can have it, claim it as your own if you like. :smoking:

Go here:

http://primes.utm.edu/bios/newprover.php

Fill everything. I'm going to sleep, please ask Beyond help.

[Beyond]

:woot:
I read your PM and now I'm :looney: trying to figure it out...

Re-read it, went there... that's :looney: ...
Way too many hoops to jump through just to put a name on a prime. :swear:

You can have it, claim it as your own if you like. :smoking:

If you need help, I can walk you through it. I done this many times.

[IronBits]

I did the first part, but I have to tell you, it's the most bizarre forms layout I have ever seen.
Someone needs to re-write that for them or something.

Well I got registered, I think, if that's what they call it.
I got the email with password in it.
I can PM it to you so you can finish it up if you like ;) :thumbs:

None of it makes any sense to me at all. Totally illogical.

[Beyond]

I setup your proof-code it is L566. If you like, I can submit the prime under your new proof-code for you, you will send the prime to me by PM.

[IronBits]

I setup your proof-code it is L566. If you like, I can submit the prime under your new proof-code for you, you will send the prime to me by PM. Thanks! :thumbs:
How do I send you the prime? I thought it was
user=Free-DC_IronBits
[02/12/08 12:56:55]
595*2^332639-1 is prime!

[Beyond]

That is not it, it should have been in a PM

[em99010pepe]

That is not it, it should have been in a PM

It's there.

[IronBits]

He's got the PM now.

Thanks Beyond! :hifi:

[Beyond]

He's got the PM now.

Thanks Beyond! :hifi:

Your Welcome

It has been submitted ---> http://primes.utm.edu/primes/page.php?id=84070

Next time it will be much easier now that everything is setup.

[em99010pepe]

Thank you both.

[Beyond]

Carlos, can you bump up the k/n pairs on the LLRNet Server (if needed), I'm going to make a run on this project for a week or 2 while the exponents are still small enough for AMD machines to make a good contribution?



/edit Never mind, I've decided to do some sieving...maybe next time.

[em99010pepe]

Carlos, can you bump up the k/n pairs on the LLRNet Server (if needed), I'm going to make a run on this project for a week or 2 while the exponents are still small enough for AMD machines to make a good contribution?



/edit Never mind, I've decided to do some sieving...maybe next time.

The server has plenty of work and I still have a few ranges reserved to add, just waiting one or two days because of the size of the knpairs.txt. Next 24 hour rally will take within two weeks, this Saturday I can't be at home just for the case the server needs to be restarted.
Please tell me when you're going to free up your machines to LLRNet, I also would like to move all cores to it to boost even more the project. BTW, if you read the project forum the guys without team they are willing to race against Free-DC...I just told them to not underestimate our power...lol


Carlos

EDIT: more than 65 000 k/n pairs on the server...

[em99010pepe]

Beyond,

If you want to add your AMD's we still have some lower ranges to test. I'm there with all my machines but with the manual client. I can always setup a server just for you.
Please say something.

Carlos

[Beyond]

I have a 8 hour refresher course to take on Saturday, so it will be late that night or Sunday before I switch over to LLRNet.

I did add 2 or 3 machines to the current LLRNet server last night to see what timings I would get with various processors, the AMDs @ 2.2GHz are arveraging 290 sec, the C2D's @ 2.2GHz 162 sec.

If you want to finish off the lower ranges I can provide at least 16, maybe 20 AMD cores.

[em99010pepe]

I have a 8 hour refresher course to take on Saturday, so it will be late that night or Sunday before I switch over to LLRNet.

I did add 2 or 3 machines to the current LLRNet server last night to see what timings I would get with various processors, the AMDs @ 2.2GHz are arveraging 290 sec, the C2D's @ 2.2GHz 162 sec.

If you want to finish off the lower ranges I can provide at least 16, maybe 20 AMD cores.

Yes I want. Each one takes between 55 and 75 seconds on a 2.9GHz quad machine. My AMD is averaging that timing of 290 seconds...
Saturday 11 pm GMT I'll set up the server on the same address but port 80, ok?

[Beyond]

Sounds good. And Thanks. :cheers:

[em99010pepe]

Sounds good. And Thanks. :cheers:

I already reserved 276k-320k range just waiting for the file. I'll move all cores to help you clean this up in no time.

Carlos

[em99010pepe]

Server is up and running on port 100.

[Beyond]

Server is up and running on port 100.

I'll try to find time to start moving some of the AMD over this evening.

[em99010pepe]

Thank you. I only ask you to set up the wu cache to 50 or so....I still have 4 days to finish my manual reservations then I'll help you out.

[Beyond]

Thank you. I only ask you to set up the wu cache to 50 or so....I still have 4 days to finish my manual reservations then I'll help you out.

:thumbs:

Know how many k/n pairs?

[em99010pepe]

:thumbs:

Know how many k/n pairs?

On the server I suppose something like 150 000 k/n pairs...:rotfl: I really don't know.

EDIT: I have something like 43 000 k/n pairs reserved, that estimated for the server is way down. More than 400 000 k/n pairs. My average is one test per minute.

[Beyond]

Okay.

I was trying to get a feel for how long it would take to run all the lower range pairs. With 400k pairs, probably 3 weeks.

[em99010pepe]

Okay.

I was trying to get a feel for how long it would take to run all the lower range pairs. With 400k pairs, probably 3 weeks.

Probably less but it's something that need to be done. Thank you once again.

[IronBits]

Do you need me to move 4 cores over to port 100 to help out or just stick with 300?

[em99010pepe]

Do you need me to move 4 cores over to port 100 to help out or just stick with 300?

That might be a good idea...don't forget to first to clear the cache.

[em99010pepe]

user=Free-DC_Beyond
[02/16/08 11:26:18]
465*2^355428-1 is prime! ----->>>>Confirmed

user=Free-DC_Beyond
[02/16/08 20:13:46]
723*2^276579-1 is prime!

[IronBits]

That might be a good idea...don't forget to first to clear the cache.
Ok, 4 cores dedicated to clean up on port 100 :)
Let me know when to switch back to port 300. :hifi:

[em99010pepe]

Ok, 4 cores dedicated to clean up on port 100 :)
Let me know when to switch back to port 300. :hifi:

Yes I will, thank you.

In this page (http://www.rieselprime.org/nplb/NPLB_TeamDrive2.htm) it's possible to visual the cleaning progress. 320.0k to 333.2k is almost complete, a few ranges to confirm from LLRnet and one from me. 260.0k to 320.0k, the cleaning is about to begin.

Carlos

[em99010pepe]

user=Free-DC_Beyond
[02/17/08 03:10:39]
885*2^276858-1 is prime!

user=Free-DC_Beyond
[02/17/08 07:59:48]
567*2^277069-1 is prime!

[em99010pepe]

user=Free-DC_Beyond
[02/17/08 20:18:24]
979*2^357019-1 is prime!

user=Free-DC_Beyond
[02/17/08 22:43:19]
735*2^357135-1 is prime! (confirmed one)

Just to remain:

crus.dynip.telepac.pt port 300: Top 5000 candidates to test
crus.dynip.telepac.pt port 100: cleaning lower ranges, no Top 5000 candidates

[em99010pepe]

user=Free-DC_Beyond
[02/18/08 05:09:13]
549*2^277884-1 is prime!

user=Free-DC_IronBits
[02/18/08 09:45:15]
417*2^278058-1 is prime! confirmed

user=Free-DC_IronBits
[02/18/08 15:12:59]
861*2^278225-1 is prime! confirmed

user=Free-DC_Beyond
[02/18/08 15:30:38]
503*2^278210-1 is prime!

user=Free-DC_Beyond
[02/18/08 00:10:54]
895*2^357205-1 is prime!

user=Free-DC_Beyond
[02/18/08 16:59:29]
987*2^358020-1 is prime!

And alpha found one, please check your PM

[Beyond]

Racking them up now....Go Free-DC! :cheers:

For a change of pace, just bought the bits to put together an AMD Quad Core instead of the usual Intel Quad Core today. But will not have time to get it setup till the weekend. :(

[em99010pepe]

I already switched one core to port 100.

[IronBits]

And alpha found one, please check your PM :cheers: :|party|:

[IronBits]

For a change of pace, just bought the bits to put together an AMD Quad Core instead of the usual Intel Quad Core today.
But will not have time to get it setup till the weekend. :( :cry:

[em99010pepe]

A new page with a primes-overview can be found here: www.rieselprime.org/nplb/NPLB_Primes.htm

[alpha]

And alpha found one, please check your PM

Yay! Submitted (http://primes.utm.edu/primes/page.php?id=84136).

[Beyond]

Carlos, do not count on my production level for the rally this weekend or the days ahead, just find out the company is getting ready to shut down 2 of our 3 offices here in Houston. :(

Not sure where that leaves me. :confused:

[em99010pepe]

Carlos, do not count on my production level for the rally this weekend or the days ahead, just find out the company is getting ready to shut down 2 of our 3 offices here in Houston. :(

Not sure where that leaves me. :confused:

Don't worry, you already contributed a lot to the project. Meanwhile, could you move what you have to server-port 100 and help me clean the lower ranges?

Carlos

[IronBits]

just find out the company is getting ready to shut down 2 of our 3 offices here in Houston. :(
Not sure where that leaves me. :confused: Say it ain't so :cry:
Good luck !!!

[Beyond]

We shutdown 90% of our East Coast operations in the last 6 to 9 months. The hard times have reached Texas now. We shutdown 6 offices across Texas in the last 2 months. So I guess it is our turn in Houston to take a hit.

I'm not sure if I will be affected. [knocks on wood]

I currently wear 2 hats the 1st (and most important) is managing our commercial construction division which has not seen a downturn in the market, yet. And 2nd is running the IT department for our Southwest Division.

Just hedging my bets, hoping for the best and planning for the worst.

[Beyond]

I have been kindly asked to stop finding primes, it seems the administrator(s) of the project are afraid that we are running out of potential primes. So I will be shutting down operations this evening. :(


:D

[em99010pepe]

Beyond,

I'm less than 12 hours to move the rest of my fleet to server port 100. So far I have 3/8 on it. Care to join me for a week? I'll stay there until I dry out the server.

Carlos

[Beyond]

Beyond,

I'm less than 12 hours to move the rest of my fleet to server port 100. So far I have 3/8 on it. Care to join me for a week? I'll stay there until I dry out the server.

Carlos

I'll join you tomorrow.

On the job front, I have it on good authority (the owner) that the office closings are not scheduled till next month and I personnaly have no worries, my job is secure. :hifi:

The only good thing about this mess, is I will have access to those other 2 offices machines when we move them to our office for storage. I have a "good" place picked out for storage (in my office). :)

[IronBits]

:rotfl: :thumbs: :rock:

[Beyond]

:rotfl: :thumbs: :rock:


I was sweating bullets for a while .... too old to start over now. :geezer:

[em99010pepe]

I'll join you tomorrow.



I'm already there with 7 cores, more to add.

Carlos

[Beyond]

I've added 5 of the 8 machines in my office at work. I'll start moving the home machines to port 100 when I get home in a couple of hours.

[em99010pepe]

IronBits, are you going to move the rest of your cores to server port 100?

[Beyond]

Carlos, I think your server has crashed ...... several people are unable to connect. :Pokes:

[IronBits]

IronBits, are you going to move the rest of your cores to server port 100? Yes, shortly. Have to empty the cache first, then switch the port...

[Beyond]

Yes, shortly. Have to empty the cache first, then switch the port...

I'd wait to see if the server recovers, port 300 server is still running, but port 100 is down. :(

[IronBits]

LLRnet-CRUS current status :
Sleeping ...
LLR network client 0.9b7 started
Based on LLR version 3.5.0
Loading library 'win32' ...
winloop thread
LLRMain started
server dialog thread started ...

I guess that's what you mean :bang:

[IronBits]

What does this mean?

LLRnet-CRUS1 current status :
Initializing ...
LLR network client 0.9b7 started
Based on LLR version 3.5.0
Loading library 'win32' ...
winloop thread
LLRMain started
server dialog thread started ...
StartServiceCtrlDispatcher returns 0

[em99010pepe]

I suggest to move all cores to port 300. I really don't know what's going on on port 100. That's the only server that is down at the moment.

Sorry.

[em99010pepe]

Everything back to normal, don't ask me what happened...

[Beyond]

So are we to use 100 or 300 for the rally?

[em99010pepe]

So are we to use 100 or 300 for the rally?

100.
llrserver has a memory leak so I need often to restart the servers. I forgot to do it last night.

[Beyond]

100 then.

[em99010pepe]

Thank you.
I'm sorry but the server went down at 2 am here and I went to bed at 1 am :bang:...

[Beyond]

Thank you.
I'm sorry but the server went down at 2 am here and I went to bed at 1 am :bang:...

I noticed ..... no problem, I had already pulled 1000+ exponents for my proxy.

Im noticing a lot of "server refused your new result" error messages

[em99010pepe]

I noticed ..... no problem, I had already pulled 1000+ exponents for my proxy.

I'm noticing a lot of "server refused your new result" error messages

Don't worry, they will be re-assigned but for the last 3 hours I received only one rejected pair on port 300 and 9 on port 100 where 6 were from you. That's normal, it's because I tried to put the server online for several times...

The client idles connections when it can't connect and tries again to connect every 1 min or so, the server will get all results. So far on more than 500 000 k/n pairs we had two misses.

[alpha]

Should port 300 still work? Mine has seized up.

[em99010pepe]

Should port 300 still work? Mine has seized up.

Yep. Your last result was at:
user=Free-DC_alpha
[02/23/08 11:06:33]

It's 12:21 here...what's your cache size?

[alpha]

Restarted the client and it is working again now.

Cache size is just the default of 10. Doesn't really matter because I'm running it alongside dnet. For most of the time I've contributed to this it has only been half a core but I gave it a full core overnight last night.

[em99010pepe]

Restarted the client and it is working again now.

Cache size is just the default of 10. Doesn't really matter because I'm running it alongside dnet. For most of the time I've contributed to this it has only been half a core but I gave it a full core overnight last night.

Ok. Got 5 results from you.

[em99010pepe]

Ironbits,

Please check all your cores running on server port 100. Last contact before the server went down was from one of your machines. A few years ago RieselSieve had a similar problem and it was a faulty client that jammed the server.

Thank you.

Carlos

[em99010pepe]

Beyond,

I'm pulling off my machines from LLRnet and feed them with manual work from the lower ranges. I reserved a range that will put my cores busy for two weeks.

Carlos

[Beyond]

I have not geared up yet and you are pullng boxes off?

[em99010pepe]

I have not geared up yet and you are pullng boxes off?

I'll do the work not with LLRnet client but with LLR standard client.

[em99010pepe]

I changed one of the server settings allowing more number of concurrent connections. The server port 100 had 8 and server port 300 had 16, changed both to 30. Let's see what happens.

[Beyond]

I'll attempt to get everything switch over to running off my proxy in the next hour or 2, should reduce the number of clients connecting to your server but it will increase the amount of time I'm connected.

Proxy is set to cache 1000 test and uploaded every 100 results. I can increase this if needed?

[em99010pepe]

I'll attempt to get everything switch over to running off my proxy in the next hour or 2, should reduce the number of clients connecting to your server but it will increase the amount of time I'm connected.

Proxy is set to cache 1000 test and uploaded every 100 results. I can increase this if needed?

You can keep your proxy server on, just reduce the number of uploaded results to 50 at time (proxyTosendThreshold = 50).

[IronBits]

My clients upload each wu as it completes.
Is there a setting I can change so that it caches 10 wus, finishes 8 and starts looking for the server to upload?
Also, is there a windows or OSX version of the llrnet server software I can use that might help out?

[em99010pepe]

My clients upload each wu as it completes.
Is there a setting I can change so that it caches 10 wus, finishes 8 and starts looking for the server to upload?

I think this is the setting:


-- 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



Also, is there a windows or OSX version of the llrnet server software I can use that might help out?


Give me more 2-3 hours to see if the servers are ok, meanwhile you can start to download this:

http://stats.rieselsieve.com/download/llrnet-windows.zip
http://stats.rieselsieve.com/download/llrnet.linux.tgz

And start changing a few settings on llr-serverconfig.txt file, see bold text.


--
-- 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)
--
-- SERVER CONFIGURATION FILE
--
-- WARNING: comments start with two dashes
--

-------------------
-- general settings
-------------------

-- server name
-- used in particular to identify the server into the sql stats table
-- (change this name if you have several servers using the same sql database)
serverName = "CRUS.dynip.telepac.pt"

-- port that the server listen to (make sure clients are configured
-- with the same number)
port = 100

-- maximum number of concurrent connections
maxConnections = 30

-- maximum allowed time for a user to complete a job (in seconds)
jobMaxTime = 4 * 24 * 3600 -- 4 days

-- various files used by the server
knPairsFile = "knpairs.txt"
jobListFile = "joblist.txt"
resultsFile = "results.txt"

-- if you want the server to regularly prune your knpairs and joblist
-- files, then uncomment the following line and set the period of
-- pruning in second.
prunePeriod = 0.5 * 24 * 3600 -- 1/2 days

-- (for gurus only)
if WIN32 then
no_shell = 1
end

-- format used to display numbers in result file
-- (uncomment the appropriate line)
--displayFormat = "%s*2^%s+1" -- use this for PRP type test
displayFormat = "%s*2^%s-1" -- use this for LLR type test (default)

-- uncomment the following line to remove username check when
-- accepting a new result
noUserCheck = 1

-- you can customize the way the date is displayed
-- (in joblist.txt and results.txt)
--function DisplayDate()
-- return date("%d/%m/%Y\ %r")
--end

-----------------
-- PROXY settings
-----------------
-- uncomment the following line to make this server a proxy server
--proxy = 1

-- address of the parent server
proxyMasterAddress = "localhost"
proxyMasterPort = 7000

-- name of this proxy (change 'nobody' to something else)
proxyName = "nobody"

-- size of the proxy cache
proxyCacheSize = 10

-- minimum number of results to send back to the main server
-- at a time
proxyTosendThreshold = 10

-- name of the tosend file of the proxy
tosendFile = "tosend-proxy.txt"


-----------------
-- User CALLBACKS
-----------------

-- below is an example of callback to log all primes as they are
-- found into an additional file

--function OnPrime(t, k, n, job)
-- local fileprime = openfile("primes.txt", "a")
-- if fileprime then
-- write(fileprime, format("[%s]\n", job.resultdate))
-- write(fileprime, format(displayFormat.." is prime! Time : %d.0 sec.\n",
-- k, n,
-- Seconds() - job.seconds))
-- closefile(fileprime)
-- end
--end


-- similarly, you can have another callback called OnResult,
-- beware that it takes one additional argument (the residue)

--function OnResult(t, k, n, residue, job)
-- insert whatever you want here
--end


------------------------------------------
-- SQL settings (mySql only at the moment)
------------------------------------------

-- uncomment the line below to use SQL
--sqlUsed = 1

-- sql general configuration
sqlServer = "localhost"
--sqlServerPort = 4000
sqlDriver = "mysql"
sqlUser = "llrnet"
sqlPassword = "llrnet"
sqlDbName = "llrnet"

-----------------
-- SQL user table
-----------------
-- uncomment the line below and set the name of the table of users
--sqlUsersTable = "users"

-- name of the fields containing the username and the password
sqlUsernameEntry= "username"
sqlPasswordEntry= "password"


------------------------
-- SQL pairs and results
------------------------
-- uncomment the line below and set the name of the tables
--sqlPairsTable = "pairs"
sqlJobsTable = "jobs"
sqlResultsTable = "results"
sqlRejectedTable= "rejected"


------------
-- SQL stats
------------
-- uncomment the line below to allow llrnet server statistics and infos
-- to be regularly updated into the given sql table
--sqlStatsTable = "stats"


--------------------------
-- WIN32 specific settings
--------------------------
-- uncomment the line below to hide the tray icon on startup
--hideTrayIcon = 1

-- change this value if you want to run several client as services
serviceName = "NPLB_LLRserver"

-- under WIN32 platform, uncomment the following line to log stdout and stderr
-- to a file called "stdout.txt"
--logOutput = 1


-- MY TEST
--port = 7001
--proxy = 1
--knPairsFile = "knpairs-proxy.txt"
--jobListFile = "joblist-proxy.txt"
--proxyCacheSize = 10
--proxyName = "proxy1"

--sqlUsed = 1
--sqlUsersTable = "users"

[Beyond]

I think this is the setting:

That's wrong .... that is the setup for the main server.


IB, this all you really need to change


-------------------
-- general settings
-------------------

-- server name
-- used in particular to identify the server into the sql stats table
-- (change this name if you have several servers using the same sql database)
serverName = "Free-DC_Beyond"

-- port that the server listen to (make sure clients are configured
-- with the same number)
port = 7000

-- maximum number of concurrent connections
maxConnections = 8

-- maximum allowed time for a user to complete a job (in seconds)
jobMaxTime = 4 * 24 * 3600 -- 14 days

-- various files used by the server
knPairsFile = "knpairs.txt"
jobListFile = "joblist.txt"
resultsFile = "results.txt"

-- if you want the server to regularly prune your knpairs and joblist
-- files, then uncomment the following line and set the period of
-- pruning in second.
prunePeriod = 1 * 24 * 3600 -- every days

-- (for gurus only)
if WIN32 then
no_shell = 1
end

-- format used to display numbers in result file
-- (uncomment the appropriate line)
--displayFormat = "%s*2^%s+1" -- use this for PRP type test
--displayFormat = "%s*2^%s-1" -- use this for LLR type test (default)

-- uncomment the following line to remove username check when
-- accepting a new result
noUserCheck = 1

-- you can customize the way the date is displayed
-- (in joblist.txt and results.txt)
--function DisplayDate()
-- return date("%d/%m/%Y\ %r")
--end

-----------------
-- PROXY settings
-----------------
-- uncomment the following line to make this server a proxy server
proxy = 1

-- address of the parent server
proxyMasterAddress = "CRUS.dynip.telepac.pt"
proxyMasterPort = 100

-- name of this proxy (change 'nobody' to something else)
proxyName = "Free-DC_Beyond"

-- size of the proxy cache
proxyCacheSize = 2000

-- minimum number of results to send back to the main server
-- at a time
proxyTosendThreshold = 50

-- name of the tosend file of the proxy
tosendFile = "tosend-proxy.txt"

[em99010pepe]

Beyond is right, you can set up a proxy, sorry, forgot about it.

[em99010pepe]

Increased number of concurrent connections to 60. Rally is under way.

First result in on port 300:


user=Anonymous
[02/23/08 18:59:41]
837*2^365204-1 is not prime. Res64: 475EF4E87375EFE4 Time : 38999.0 sec.
user=Free-DC_Beyond
[02/23/08 19:00:01]
1001*2^365554-1 is not prime. Res64: 30C65979AFA98D2E Time : 169.0 sec.
user=Free-DC_Beyond
[02/23/08 19:00:02]
775*2^365547-1 is not prime. Res64: AB9DB898FF4862D2 Time : 1518.0 sec.and on port 100 (draw):


user=gd_barnes
[02/23/08 18:59:55]
639*2^284613-1 is not prime. Res64: D03D672FFF1B01F0 Time : 756.0 sec.
user=Free-DC_em99010pepe
[02/23/08 19:00:03]
401*2^284650-1 is not prime. Res64: 03F684C4F1B1AF3F Time : 436.0 sec.
user=gd_barnes
[02/23/08 19:00:03]
533*2^284614-1 is not prime. Res64: C52191D6256872F1 Time : 740.0 sec.
user=ArsTechnica_BlisteringSheep
[02/23/08 19:00:05]
979*2^284555-1 is not prime. Res64: BBED63F2CBFBFEFE Time : 1693.0 sec.

[Beyond]

If Mr. Barnes continues to have problems connecting I'll back off the rally, with 60-70 cores running I'm afraid I am flooding the server with connection request.

[em99010pepe]

If Mr. Barnes continues to have problems connecting I'll back off the rally, with 60-70 cores running I'm afraid I am flooding the server with connection request.

He can run the manual LLR, and you?

[em99010pepe]

I pull out 6 cores from LLRnet (now manual LLR'ng the lower ranges), at least the connection is free from here to the world.

[Beyond]

I could, but I won't. Too much work involved and too many machines, I do not use the manual LLR on my personnal prime searches, just llrnet.

I cut-off access to your server from my end....will wait till the cache is drained brfore reconnecting.

[em99010pepe]

I could, but I won't. Too much work involved and too many machines, I do not use the manual LLR on my personnal prime searches, just llrnet.

I cut-off access to your server from my end....will wait till the cache is drained brfore reconnecting.

Please keep your machines online, at NPLB forum we are trying to pull out a few cores to free up the connection for you.

Carlos

[Beyond]

Carlos, just now uploading 4000+ results. I will attempt to download 15000+ tests to keep the machines crunching thru the night.

[em99010pepe]

Carlos, just now uploading 4000+ results. I will attempt to download 15000+ tests to keep the machines crunching thru the night.

Ok.

[Beyond]

It did not work out, was unable to grab that many tests. Probably a good thing. I have been unable to get everything connecting to my proxy, it's being overloaded and dropping connections too.

I've pulled a few machines off to see if the proxy can catch up. If that works fine, if not that is fine too, I'm tired of messing with them, I've had enough for one day.

I think the best bet for the next time is to reserve several large ranges and split them between several llrnet servers running on my systems. :idea:

[Beyond]

It was a good try Carlos, but your server is down again. I have 2000+ results (as of now) ready to upload when you come back online. :Pokes::cheers:

[IronBits]

Ok, going to try to get a proxy server running.
If all works out well, I may open it to the outside.

[Beyond]

Well I'm completely out of work, but still a boat load of results to be uploaded. Going to bed and will worry about it tomorrow.

[IronBits]

It's spotty at best. I got lucky, managed to get my proxy server up (thanks!) and snagged 100 just to keep my 8 cores busy for a little bit and see how it works out.
I have 4gb of ram so that's a non-issue.
Now Carlos can't blame me for crashing his port 100 server :p ;) :D

[em99010pepe]

The rally is off, I'm tired of all this stupid server. 19c2c462 client is stalling the server...

[em99010pepe]

Clients connected on server port 100 are:

22489c1 and 200a8c0 Free-DC_em99010pepe
829ea741 gd_barnes
75242c52 Flatlander
ee651e4c Free-DC_Beyond
b6ccb9d8 XtremeSystems_glennpat
5498c850 Free-DC_tnerual

[em99010pepe]

Server is up and running.

[em99010pepe]

19c2c462 is you Beyond, must be one of your clients, I suspect the proxy server...lol
We need to find a way to minimize the traffic load...

[em99010pepe]

I think the best bet for the next time is to reserve several large ranges and split them between several llrnet servers running on my systems. :idea:

Yep, your own llrnet server. Then you can send the results to Karsten and he will parse them to see if all candidates were tested or not.

[Beyond]

It's Alive! Work is flowing again. :cheers:

Hang in there Carlos, we will get the lower ranges knocked out. :)

[em99010pepe]

It's Alive! Work is flowing again. :cheers:

Hang in there Carlos, we will get the lower ranges knocked out. :)

I'm trying hard...since my 6 am to 8 am I rebooted the server several times, then I gave up until 9 am, went to bed instead. When I woke up I managed to turn it up again when you turned off your machines...lol

[Beyond]

I was trying to just start up a few at a time to keep from flooding the server this morning. It does not seem to be working. even though I got a bunch of results uploaded and a few tests downloaded it has locked up again.

I'm going to clear the caches on all the clients while I have access cut off to your server, then shut the clients down. At that point I will open access to your server again to upload the results.

Sorry for all the problems caused, I would have never dreamed llrnet could not handle the load. :(

[em99010pepe]

Beyond,

Tnerual pointed me something useful that I though I already had. Microsoft introduced Win XP SP2 and limited the number of possible TCP connection attempts per second to 10 (from unlimited in SP1). I applied http://www.lvllord.de/ patch and increased that number to 100.
More info in here (http://www.speedguide.net/read_articles.php?id=1497).

I hope now the server can handle more connections per second. It's XP SP2 limitation and not llrnet one.

Carlos

[Beyond]

That is good to know. It seems to be working now. I will give it a while before ramping back up. I'm sure you realize that the higher we go in "n" value the less the load will be on the server connections, it is just a matter of time and all will smooth out.

[em99010pepe]

That is good to know. It seems to be working now. I will give it a while before ramping back up. I'm sure you realize that the higher we go in "n" value the less the load will be on the server connections, it is just a matter of time and all will smooth out.

I wasn't here but the server went down from 18:51 to 20:03 and got up by itself...:looney:

Yes, as n goes high the number gets larger and so more time is needed to process and therefore less communications between clients/server.

[tnerual]

I wasn't here but the server went down from 18:51 to 20:03 and got up by itself...:looney:

maybe it was the "pruning" job ... if llrnet is at a higher priority than llrserver, pruning job is going very slowly and server appears off during the whole time.

how to check that ? take a look at the knpairs.txt details (last modification time)

if time of knpairs.txt is around 20:03, increase priority of llrserver

[Helix_Von_Smelix]

Does anyone know how to run this from behind a proxy server??

:cheers: