Erm... I don't know what went wrong...
9 Dec 2005 9:57:57 UTC
Topic 190337
(moderation:
I don't know what went wrong but I was granted 0.00 credit for this result (11930965) when the other two completed results were awarded credit... all I can assume is that it thinks the result I sent back is fake or something (the tweleve hours (ish) my machine spent processing it, however, is _not_ a figment of my imagination - I can see the logs!)
Why did this happen?
Also a more altruistic question... If my result is believed to be fake then why is it still used to allow the other two results to be given credit? Surely if I have posted a fake result then the requirement for three valid results has not been met and so credit should not (yet) be given...
If you want to see the logs I have I can provide them.
StrongARM
Erm... I don't know what went wrong...
)
That happens to all of us from time to time. More often to Linux/Mac
users though, because of the sheer number of Windows users. See
this post
for details.
Michael
Team Linux Users Everywhere
I'm not familiar with the
)
I'm not familiar with the output from Einstein 4.81 (Linux) so I'll paste it here to save someone else from having to dig:
5.2.7
Detected CPU type 1
Fstats.Ha: bytecount 811531 checksum 38893162
Fstats.Hb: bytecount 2743068 checksum 131451756
I don't see anything wrong with this - it's different from 4.82 (Windows), though almost identical to 4.79 (Mac), but there's no error messages as I was expecting to find. Both of the other computers in that quorum were Windows, so it is likely that when the validator compared the output of all three, yours was "just barely far enough" away from matching the other two that it decided it was invalid. This sometimes (less than 1% I think) happens, and it happens both ways; if there were two Linux (or Mac) boxes and one Windows box, the Windows box would have gotten the invalid flag.
There isn't any way around this without making the validators so open that people COULD slip through with cheated results. Windows just "does math" different, and every once in a while, we Mac or Linux people just have to put up with it. It's amazing that 99% of the time Windows _does_ match with the real computers... :-)
If it happens to you _regularly_, then I'd be concerned that your computer has some problem where it's throwing off math errors - over-overclocked, overheated, bad RAM, etc... but if it only happens to you once every 5,800 credits, I'd just mark it off as a fluke and forget it.
EDIT:: I type too slow, Michael beat me to it. Okay, the figure given in that thread is that it happens 0.3% of the time.
Another edit: To answer your "why did it count as one of the three" question. The rule is three SUCCESS results, according to the hosts themselves. Then the validation rule is "two VALID" results. If it hadn't had two that matched, it would have waited for the fourth. (Or fifth, or sixth...) But once three are in without flagged errors, the validator runs.
Thanks guys... I guess that
)
Thanks guys... I guess that answers both my questions pretty well... guess I'll just have to live with another issue of being not mainstream! :)
StrongARM