help please...

[laguardia]

After installing SoB, I got a test from the server.
But after having done 2 blocks I closed my computer.
When I restarted a couple of hours later, my blocks were gone, and I got a new test with new blocks. I saw on my user profile that my 2 blocks were sent.

Has an another user got my previous test and is he continuing from were I left or are the blocks simply lost?

And what should I do, I would like to complete the entire test myself?

Not sending intermediate blocks doesn't help either, cause after rebooting I still got a new test??????

please help me, thx in advance

[Ken_g6[TA]]

This should be in a FAQ. There's an older thread about this (http://www.free-dc.org/forum/showthread.php?s=&threadid=4728), and I thought there was one even older.

[allio]

Ken's right, but what that thread doesn't say is that this happens after virtually every new install of the client. For some reason, it always drops the first test after you reboot. You should be able to reboot without losing your work, now :)

[dfamily]

Originally posted by laguardia
And what should I do, I would like to complete the entire test myself?


Notes on Continuing an Abandoned Test
-------------------------------------
The following applies to a windows machine NOT running SOB as a service...
(tested on Windows 2k/98/ME)

As long as it's listed in your pending tests, you can complete one of your abandoned tests but understand that in order to do so, you will have to edit the Windows Registry File. Unless you are totally comfortable with that, personaly, I say that salvaging 2 blocks versus a chance of messing up the computer isn't really worth it. Expire the test and don't worry about it.

I used to think that if the orphaned test was 1/3rd or less, it would be worth the trouble to salvage it. In reality, it's six of one and half dozen of another. No matter what the percentage is, you will still lose that amount of machine time whether you spend it re-working an old test or losing that x% and getting credit for x% of a new one...

But, if you're like me and there are times that you are determined to finish that test no matter what...

Make sure the abandoned test is listed in your pending tests. (Login, Preferences, Pending Test Management) If you press the "Expire" button, it will be reassigned to someone else immediately. Or if there is no activity on it for 10 days, it will be reassigned to someone else.

Assuming you have started a 2nd test and it is currently running...

You can save test #2 and work on test #1 but this example assumes that
you let test #2 finish and then work on the test #1...

it's easier to let test #2 finish first...but be sure and set it NOT to grab a new test #3 when #2 is done (click on Config and then check "On completion, exit w/out new test", Save).

Before test #2 is done, to be safe, you also want to make a note of the k and n of test #1 as well as a back up of the z file for test #1. The Z-file will be a file in c:\program files\sb\ that begins with Z followed by the N number of your test, no extension. For example, my K-55459 N-5098486 has a z-file of z5098486. I just copy it to the desktop.

(I might interject here that you may see an old z-file from a long abandoned test. That does not mean you can still process it and get credit for it. Unless it's listed on your pending file page, then delete it and forget it.)

Also don't be fooled. Once test #2 is finished, the icon may still appear in the system tray and at a glance, you're thinking it's still running...it won't disappear until you mouse over it.

So, when test #2 is finished, the client stops without grabbing a new test. All is well. Now make sure that the Z-file for test 1 is in the c:\program files\sb\ folder (double check, if not, it will start test 1 at the beginning)

Edit the registry file [HKEY_LOCAL_MACHINE\SOFTWARE\LhDn\sob\cache] and give
it the K and N for test #1. REBOOT. (did you note the REBOOT there?) If you have it on auto start then it should read the cache and start up where it left off...

and you want to be sure and un-check the exit without new test option...

good luck...

ps: this has worked for me more than once...use at your own risk, no warranties implied or expressed etc...

[Meloentje]

I'd like to add a little note to this.

When you start SB after it has finished a test with "On completion, exit w/out new test" checked it grabs a new test. This is because the program knows the previous test was completed.
As a result of this the registry entries for k and n will change, thereby overwriting the k an n values of test #1 you just entered.
To prevent this you also have to change the registry setting "cache" in [HKEY_LOCAL_MACHINE\SOFTWARE\LhDn\sob\cache] to 2.
:thumbs:

[vaughan]

My situation is a little bit different but along the same lines as the original post. When I stopped SB so I could reboot and run some other DC project for a few days I came back and SB restarted the same k=24737 and n=5046031 from the beginning. This caused me to go down in the stats by 80497. Why did it do this? I'm using v1.1.1 on an Athlon XP2000+ lightly overclocked from 1.67 GHz to 1.72 GHz with 512 MB DDR333 RAM on an MSI K7N420 Pro motherboard and I'm running Win XP Pro SP1.

From the SB.log:
[Sat Dec 20 03:04:10 2003] n.high = 2395592 . 270 blocks left in test
[Sat Dec 20 19:25:40 2003] got k and n from cache
[Sat Dec 20 19:25:40 2003] restarting proth test from cache (k=24737, n=5046031) [47.6%]
[Sat Dec 20 19:36:56 2003] resolving hostname
[Sat Dec 20 19:36:56 2003] opening connection
[Sat Dec 20 19:36:57 2003] receiving from server
[Sat Dec 20 19:36:58 2003] logging into server
[Sat Dec 20 19:36:59 2003] login successful
[Sat Dec 20 19:36:59 2003] n.high = 2405410 . 269 blocks left in test

snip

[Sun Dec 21 21:38:41 2003] n.high = 3210486 . 187 blocks left in test
[Sun Dec 21 21:56:05 2003] block processing paused
[Wed Dec 24 21:36:34 2003] got k and n from cache
[Wed Dec 24 23:01:57 2003] resolving hostname
[Wed Dec 24 23:01:57 2003] opening connection
[Wed Dec 24 23:01:57 2003] receiving from server
[Wed Dec 24 23:02:06 2003] logging into server
[Wed Dec 24 23:02:06 2003] login successful
[Wed Dec 24 23:02:08 2003] n.high = 9818 . 513 blocks left in test
[Wed Dec 24 23:15:11 2003] resolving hostname
[Wed Dec 24 23:15:11 2003] opening connection
[Wed Dec 24 23:15:12 2003] receiving from server
[Wed Dec 24 23:15:13 2003] logging into server
[Wed Dec 24 23:15:14 2003] login successful
[Wed Dec 24 23:15:15 2003] n.high = 19636 . 512 blocks left in test

snip

I have also seen this occasionally:

[Thu Dec 25 08:51:33 2003] n.high = 451628 . 468 blocks left in test
[Thu Dec 25 08:59:23 2003] internal computation error [mismatched sums]! check your memory/processor. test will restart in 5 minutes.
[Thu Dec 25 09:04:24 2003] restarting proth test from cache (k=24737, n=5046031) [9.0%]
[Thu Dec 25 09:14:21 2003] resolving hostname
[Thu Dec 25 09:14:21 2003] opening connection
[Thu Dec 25 09:14:22 2003] receiving from server
[Thu Dec 25 09:14:22 2003] logging into server
[Thu Dec 25 09:14:23 2003] login successful
[Thu Dec 25 09:14:23 2003] n.high = 461446 . 467 blocks left in test

vaughan
AMD Users
www.amdusers.com

[allio]

Are you using the service handler? We found that a lot of people using the service install with the +k flag set experienced the first problem, myself included. After going back to the vanilla programme my tests stopped restarted.

As for the second, it just means your overclock is high enough to be producing some dodgy results. It happens to me whenever I run my heater for too long, as my crappy memory overheats ;)

[vaughan]

Hello allio.
I'm using the plain vanilla v1.1.1 application.
Maybe I should get some better quality memory chips.

[vaughan]

Is this "install, run SB, shutdown, restart SB, lose all the work you've done on that task" bug fixed in the new v1.2.0 release?

NO. :swear: