Folding Forum

over the last few days I have had 0X12 Errors with three WU:
2605 (10, 272 Corrected *271 Sorry 7im, 33) on 3/7/08
2605 (8, 291, 34) I think I got credit for this one on 3/9/08
2605 (14, 429, 34) on 3/10/08

The first one and the third one listed I did not get credit for.
All three finished to 100% and errored at the end (0X12)... then the client started them again, and once finished again, the client folded on different WUs.

I am running Linux/Ubuntu and SMP client.

Thanx in Advance for checking.

Peace

We prefer if you cut and paste the WU data straight from the fahlog to post in the forum. It's much easier for us to put in to the search tool. Thanks.

I was afraid you were going to say that
being noob at Linux, I need to know where to go to find it (Which folder)
I can probably figure out how to copy/paste it once I know where to look.

Thanx for quick response 7im

Peace

First WU, only one upload, but not by anyone named anything like MoneyGuyBK

Your WU (P2605 R10 C272 G33) was added to the stats database on 2008-03-10 06:18:18 for 1760 points of credit.

2nd WU: Hi MoneyGuyBK (team 80856),
Your WU (P2605 R8 C291 G34) was added to the stats database on 2008-03-10 16:19:23 for 1760 points of credit.

No one has completed the 3rd WU.

Thanx again 7im for confirming my suspicion As I thought.

Question, if the first WU was crunched by me only 2 times, why was it assigned to someone else? I thought a particular WU would crunch 3 times before client gives up on it.
That WU 2605 (10, 271, 33) was first assigned to me on 3/7 at 12:40 am PST, Errored and restarted again on 3/7 at 9:21pm
*** OOPS, I jotted the wrong C Number the first time 7im it was 271 not 272, I will edit and correct my OrigPost.

*** Could you check again for me please. my Old Age must have caused me to flaw I am truly ashamed and apologize.
and again same Q about the third one, running only 2 times before I got a new assignment, Am I correct that a WU used to crunch 3 times before it was re-assigned?

Peace

MoneyGuyBK wrote:Question, if the first WU was crunched by me only 2 times, why was it assigned to someone else? I thought a particular WU would crunch 3 times before client gives up on it.

The system (including the client logic, the server logic, and all the data managed by them) is designed to complete as many WUs as possible.

If a WU has an error, there are many possible causes.
If a WU fails, maybe there was an error during download so try it again.
If it fails again, maybe it wasn't really downloaded again, maybe it came from cache, so try it a few more times.
If it fails again, maybe the problem isn't the WU itself -- maybe the FahCore_*.exe is corrupt, so download a fresh copy and try again.
OK. Maybe it's none of those, but there could be a hardware problem on this machine (including heat, overclocking, and plain hardware errors) so assign it to somebody else.
What are the chances of two different clients both having hardware errors, so assign it to yet another person.

In many cases, errors contain some scientific value and receive partial credit, so there's no reason to skimp on the number of retries. (v5.91beta SMP is a notable exception) In fact, we do see a fair percentage of WUs which award partial credit to one person and are later completed without error by another person.

The PandeGroup has not explained this to me (or anywhere outside of Stanford, for that matter) so I really don't know if that's what they were thinking when they designed FAH, nor do I know how many times each error is retried. Eventually, a bad WU is removed from circulation.

MoneyGuyBK wrote:
*** OOPS, I jotted the wrong C Number the first time 7im it was 271 not 272, I will edit and correct my OrigPost.

Hence the strong recommendation to cut and paste!

Hi MoneyGuyBK (team 80856),
Your WU (P2605 R10 C271 G33) was added to the stats database on 2008-03-08 18:20:31 for 1760 points of credit.

To sum what Bruce said about repeats, there are different circumstances that cause the WU to repeat, and thus it may repeat a different number for times for each cause. And Pande Group may change that repeat logic to better tune the performance of the project, and we really don't keep a close track on all of those possible causes, and their repetition numbers because they do change.

Thanx to both 7im and bruce for sticking with me in this.
Thanx very much 7im for checking my stats.
Thanx to my neighbor bruce for the explanation.

God Bless both of you. I am still crunching away here.

Peace