SBQueue v0.20a2 (the "yarn" release)

[Ken_g6[TA]]

SBQueue v0.20 alpha 2 is out! This version should correct the freezing problems many of you were experiencing, especially with the auto-transmit function.

To do this required a surprising number of threads (thus the name.) Of course, so many threads could get tangled up if I made a mistake - thus the continued alpha status.

Meanwhile, I've also released a version 0.19 "stable". Although I cobbled together different versions of classes to make it, it should work and will only freeze for the reasons previous versions froze.

Download either one from the link in my sig.

[Haranaka]

Great, I've just installed the new version and I'm now running it with the -t option. I'll keep an eye on it and keep you posted!

[Joh14vers6]

With this -t option activate I got this strange error;

[Thu Aug 26 17:28:31 CEST 2004] Connection received from /127.0.0.1
[Thu Aug 26 17:28:31 CEST 2004] User Joh14vers6 logged in.
[Thu Aug 26 17:28:31 CEST 2004] Work unit 99.95% complete
[Thu Aug 26 17:28:31 CEST 2004] All data received.
Connecting to server...greeted...Connected.
Logging into server...Accepted.
Transmitting block...Transmission complete.
1 of 1 blocks flushed.
[Thu Aug 26 17:53:40 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:00:07 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:05:42 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:11:17 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:16:52 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:22:27 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:28:02 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:33:37 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:39:13 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:44:48 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:50:23 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 18:55:58 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:01:33 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:07:08 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:12:43 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:18:19 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:23:54 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:29:29 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:35:04 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:40:39 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:46:01 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:51:35 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 19:57:24 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:03:00 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:08:35 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:14:10 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:19:45 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:25:20 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:30:55 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:36:31 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:42:06 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:47:41 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:53:16 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 20:58:51 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:04:26 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:10:02 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:15:37 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:21:12 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:26:47 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:32:22 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:37:57 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:43:32 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:49:07 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 21:54:42 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:00:02 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:05:53 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:11:28 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:17:03 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:22:38 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:28:13 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:33:48 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:39:23 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:44:58 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:50:34 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 22:56:09 CEST 2004] Connection received from /127.0.0.1
Error reading stream.
[Thu Aug 26 23:01:15 CEST 2004] Connection received from /127.0.0.1
[Thu Aug 26 23:01:15 CEST 2004] User Joh14vers6 logged in.
[Thu Aug 26 23:01:15 CEST 2004] Work unit 100.0% complete
[Thu Aug 26 23:01:15 CEST 2004] All data received.
[Thu Aug 26 23:01:15 CEST 2004] Listening for fetch.

I have one more computer running at the same time and at the same internetconnection that had no problem.

[Ken_g6[TA]]

Were you doing something very CPU intensive (like playing a videogame, not like running SB) on either machine at the time? That's the only repeatable case I've found when this can happen.

The only other time I've seen it happen was when I was running ZoneAlarm on one machine. I'm not sure I ever did figure that one out.

[Joh14vers6]

Originally posted by Ken_g6[TA]
Were you doing something very CPU intensive (like playing a videogame, not like running SB) on either machine at the time? That's the only repeatable case I've found when this can happen.

The only other time I've seen it happen was when I was running ZoneAlarm on one machine. I'm not sure I ever did figure that one out.

I't's my PC at work and that time I was at home, so it was doing nothing else than SoB.

[Mystwalker]

Just a (very) wild guess, but weren't there (under seldom circumstances) some errors with WinXP w/ SP2 and the localhost loopback?
It's unlikely (maybe even most unlikely, when Joh14vers6 isn't using XP SP2 ), but who knows?

[Death]

Ken, how's the current progress?
Wanna something new... =))

And, well, I try it at SP2. Maybe tomorrow....

[Joh14vers6]

Originally posted by Mystwalker
Just a (very) wild guess, but weren't there (under seldom circumstances) some errors with WinXP w/ SP2 and the localhost loopback?
It's unlikely (maybe even most unlikely, when Joh14vers6 isn't using XP SP2 ), but who knows?

The machine in question is just a P3 450MHz running at Win98SE and will be replaced soon.

[Ken_g6[TA]]

I'm back in college, so I'm not actively working on this anymore. When I had the problem the first time, I couldn't find any bugs in my code. I did find an anomaly, but it turned out to be an inconsequential bug in Kugano's server. I think at the time I blamed the problem semi-arbitrarily on ZoneAlarm, and gave up.

The machine hosting my SBQueue is my fastest machine, and it's never had this problem. The machine that had a problem with a fetch is almost done with its current WU. If it has a problem this time, I'll let you know.

[Ken_g6[TA]]

Well, the remote computer flushed with no problems. Maybe this only crops up the first time a client sends a result?

[russkris]

I have had a look at the forum to see if there was a answer to my question.
I have downloaded the SBQueue from Ken_g6[TA] sig, but could someone please tell how to make it work. do you have to make a .bat file or something like that. .

I have Java 1.4.2 installed if you need that, I only say this because I read something about it in the "readme.txt". I don't know much as I'm new at 17oB, and when I read it was made for Dial-up user and that I am. So tucking blocks away until I would like to U/L sounds great. Good for stealth THWACKING to....

Thank you very much for taking the time

Rusty

[Death]

russkris, if you have one PC you absolutely don't need a SBQueue..

it's intended for users with a lot of boxens.

just run your sb.exe normally...

[russkris]

Originally posted by Death
russkris, if you have one PC you absolutely don't need a SBQueue..

it's intended for users with a lot of boxens.

just run your sb.exe normally...

Thank you for your reply Death, but I'm just about to increase my production to 5 boxens, can you help now..

Rusty

[Mystwalker]

From the readme file:

*** Usage ***
There are two ways to start SBQueue:

Command-line:
java -jar SBQueue.jar [-rst] [-a serveraddress[:serverport]]
[-p localport] [-f user] [-l user]
[-i ImportFile]

-r: Report (flush) all completed blocks.
-s: Run a command-line server, after completing all other tasks.
-t: Auto-Transmit blocks received
-a: Set the address and port of the remote Seventeen or Bust server. Default is sb.pns.net:1717.
-p: Set the local port to listen on. Default is 1717.
-f: Fetch work for the given user (runs before any list command.)
-l: List queue contents for the given user.
-i: Import work from a .sbq file.

GUI:
java -jar SBQueue.jar
java -jar SBQueue.jar -g [-st] [-a server_address[:port]] [-p local_port]
-g: Force GUI; default is command-line unless no options are given.
All other allowed options do the same thing in the GUI they did for the command-line.

So "java -jar SBQueue.jar" should basically do the job.

[Ken_g6[TA]]

Thanks, Mystwalker.

I beg to differ with Death, though. SBQueue can be useful on just one machine, if that machine is on dial-up. In fact, that's why I originally created it.

[Death]

well, if you on a dial up, you can just call your ISP at the end of test, submit it and recieve new...

if you have a test at sbqueue cache, it can expire before your client get it.

[Ken_g6[TA]]

If you have your computer auto-dial your ISP:

1. It could interrupt an actual *phone* call you're making.
2. I can't find a way to make it hang up immediately after the client is done. This could use up all your ISP minutes (I don't have an unlimited plan), and block incoming phone calls.

Work unit expiration is now 30 days after download. It's unlikely it will take that long to get around to at least starting the test. And SBQueue caches intermediate results, so you can flush anytime you're online.

[Haranaka]

Originally posted by Ken_g6[TA]
If you have your computer auto-dial your ISP:

1. It could interrupt an actual *phone* call you're making.
2. I can't find a way to make it hang up immediately after the client is done. This could use up all your ISP minutes (I don't have an unlimited plan), and block incoming phone calls.

Work unit expiration is now 30 days after download. It's unlikely it will take that long to get around to at least starting the test. And SBQueue caches intermediate results, so you can flush anytime you're online.

The problem in my opinion is not that you can flush when ever you want but when you can fetch a fresh test. If you come 7 days after the client finished its test, the computer is sitting on his lazy ass doing nothing. That's a shame...
Perhaps it is an option not to led the client dial for an internet connection, but just wait and keep trying if the user makes the connection? Then it would just flush/fetch when someone fetches his email.

[Matt]

Quick question about sneakernetting. I have an offline PC with several units completed. Do I just copy the whole folder to a networked PC and run the client and tell it to flush? Or what?

It says sneakernetting is enabled through the merge and export operations but it's not very clear exactly how. Can somone explain it to me please?

EDIT: I copied the .sob files to a networked PC and flushed them, however I'm not sure what happens to part completed blocks? I have 2 days left to work on them apparently and I don't know if they'll be finished in time, is it possible to extend the time or to flush part of the block?

[Ken_g6[TA]]

Haranaka, my queue is supposed to try to fetch a new work unit from the main server if it doesn't have any queued (although I haven't tested that feature lately). If you set up your computer to auto-dial when an internet connection is requested, it should fetch a new work unit.

That is, unless I set the timeout time for connecting to the server to be too short. I'll check when I get home.

Matt, yes, I probably do need to be more clear about sneakernetting, though I think I might have added some more detail in the readme.txt file.

How did you determine that you have 2 days left to work on those blocks? If they're still in the queue, and you don't get to them, they will be dropped and you'll just have to fetch new ones, but no work was done on them. If you started working on the blocks, and flushed the .sob files (which is exactly what you're supposed to do, BTW), their timeout is reset to 30 days from that point, but no timeout should show up in the Queue or client windows.

[vjs]

Is there a way to specify/choose k/n pairs to be tested and qued, rather than SOB server assigned?

[Joh14vers6]

Originally posted by vjs
Is there a way to specify/choose k/n pairs to be tested and qued, rather than SOB server assigned?

Nope

[Ken_g6[TA]]

Originally posted by vjs
Is there a way to specify/choose k/n pairs to be tested and qued, rather than SOB server assigned?

I actually have a system put together that would allow this. But I'm to sure what the admins would think of being able to alter client communication this way. I sent it to both Louis and David, but I never got their blessing, so it's still sitting on my hard drive.

What I made is actually a Java API, so if and when it's approved, several other projects could benefit as well!

[vjs]

I also made a program using reg.exe from the nt4 cd.

It's a batch file which will run specified k/n pais but one basically has to modify the batch etc. It's a pain.

I always wonder about the co-ordinators of projects, I can see them trying to stop cheating etc.

But doing k/n pairs of your choice is frowed upon??

Sure it's not the best choice for the project, but hey it's your time, your machine, if it keeps you happy and gives you fuzzy feelings great.

Alot of people do 2^n-1 just to find the largest prime, if people would like to choose some of our numbers around 30m great. If we find a prime up there great. Wouldn't it be better to find a prime or have people do their own k/n pairs even if it results in a double check, tripple check, quadrupple check, for us rather than going elsewhere???

Sure repeating some k/n pair 3 or 4 times is of no benift to the progect but niether is leaving and running folding at home.

I think a user should be able to do anything that they want as long as it doesn't corrupt the project.

[Mystwalker]

Originally posted by vjs
I also made a program using reg.exe from the nt4 cd.

It's a batch file which will run specified k/n pais but one basically has to modify the batch etc. It's a pain.

As the server has no record of you getting this test, I'm quite sure it will discard it.

Alot of people do 2^n-1 just to find the largest prime, if people would like to choose some of our numbers around 30m great.

People searching for a 10M digits prime should use GIMPS, as
1. GIMPS has only one "k": 1 (which definitely speed things up),
2. it is possible to automatically get 10M digit tests there and
3. the 30M range for SoB's numbers is completely unsieved, thus the likeliness of a test resulting in a prime is magnitudes lower here.

If we find a prime up there great. Wouldn't it be better to find a prime or have people do their own k/n pairs even if it results in a double check, tripple check, quadrupple check, for us rather than going elsewhere???
Sure repeating some k/n pair 3 or 4 times is of no benift to the progect but niether is leaving and running folding at home.

The Sierpinski problem only considers IF a k leads to a prime (sloppy description, I know). When one is found, there is no need to find another one (except maybe trying to figure out if it is the lowest prime). Running tests with high n takes more time and the chance of a prime is lower. So, for the project, it is best when the smallest numbers are tested first.

Sure, the project benefits even from arbitrary work to a certain percentage (likely more than 50%). But personally, I'd prefer to have a higher "overall DC projects output" instead. When those people leave SoB and join GIMPS because they want to find 10M digit primes, it IMHO gives GIMPS more benefit than it takes from SoB (although SoB definitely is an interesting project - I still devote a lot of my time to it; it just has a slightly different goal than GIMPS, though).
Plus, those project members themselves would be better off, as I showed above.
Again, I really like SoB (to avoid misunderstandings), but I think it's not the best option to search a 10M digit prime (at least ATM - things will change in a few years, I guesstimate).

I think a user should be able to do anything that they want as long as it doesn't corrupt the project.

It at least makes it harder to coordinate...

[Ken_g6[TA]]

Originally posted by vjs
But doing k/n pairs of your choice is frowed upon??

Not at all! Changing the values in the registry is just fine. The issue I'm not sure about is whether a program that allows you to modify the client's data stream (though not connecting to the main server) would be frowned upon.

[Joh14vers6]

http://www.free-dc.org/forum/showthr...?threadid=7827

I don't if read this thread. But my question is: is it possible that you can make SBQueue let make a logfile? So we can see afterwards what tests it send to the mainserver and if it is sending tests more than once?

Arnoud
DPC

[Ken_g6[TA]]

I've been working on that, but I'm not there yet for the GUI. So for now, you can use the command line server:

java -jar SBQueue.jar -s > logfile.log

Or if you're lucky and have Linux or Cygwin:

java -jar SBQueue.jar -s | tee logfile.log

[Ken_g6[TA]]

As promised in another thread, the logging version of SBQueue is up! See my site linked in my sig.