Why can't I get this kind of credit???
28 Jan 2005 19:05:34 UTC
Topic 187382
(moderation:
Check out this result...the high credit requested was what was granted. It was about 70x the credit requested by the fastest (valid) cpu.
http://einsteinathome.org/workunit/308066
Seti Classic Final Total: 11446 WU.
Why can't I get this kind of credit???
)
A credit lottery?
That's BOINC v4.13 + einstein
)
That's BOINC v4.13 + einstein v4.71.
That's why everyone should update to BOINC v4.19, einstein v4.72 should be downloaded automatically with new WUs.
- Marco
Team Canada
But you've gotta love the
)
But you've gotta love the AuthenicAMD Pentium :)
Still, doesn't it seem strange that a Pentium 1500MHz completed the WU in 1,367.86 seconds while the XEON 2.00GHz took 62,293.36 seconds and both results were deemed canonical? I guess I just can't figure out a scenario where those kind of result could possibly match up as valid results...
www.clintcollins.org - spouting off at the speed of site
A main bug of the 4.71 app is
)
A main bug of the 4.71 app is that it only reports the CPU time since the last resume (when removed from memory), thus sometimes claiming very low credit. The result, however, should be the same, regardless of how often and when the app was interrupted.
The Pentium 1500MHz probably didn't complete the WU in 1,367.86 seconds, but it surely completed it 1,367.86 seconds after the last resume.
BM
BM
I don't expect to get 279
)
I don't expect to get 279 Credits on any WU because it has to be way out of whack to start with, but what upsets me a little is to see something like that & then dig through my Validated WU's and see something like this WU that is at the other end of the spectrum. It isn't the first time this has happen in my Account nor do I expect it to be the last.
What bugs me about the whole thing is why does the one WU get 279 Credits and my WU only get 00.65 credits for 10 hr's of crunching time. 3 of us turned in close times and credit request but we all end up with diddly squat. All I ask is that it is done the same for everybody & each time, not give the high one time and the low another time ...
Take a look at the discussion
)
Take a look at the discussion in this thread:
http://einsteinathome.org/node/187359
> A main bug of the 4.71 app
)
> A main bug of the 4.71 app is that it only reports the CPU time since the last
> resume (when removed from memory), thus sometimes claiming very low credit.
> The result, however, should be the same, regardless of how often and when the
> app was interrupted.
>
> The Pentium 1500MHz probably didn't complete the WU in 1,367.86 seconds, but
> it surely completed it 1,367.86 seconds after the last resume.
Thanks Bernd,
I had read that a couple of times but just didn't put 2 and 2 together. It makes more sense when you put it in context.
The good thing is that this beta test appears to be clearing problems like that. I'm excited to see this project go into actual production!
www.clintcollins.org - spouting off at the speed of site
The "AuthenticAMD Pentium"
)
The "AuthenticAMD Pentium" from the first post who claimed 279.43 credits did probably not really use 394,953.31 seconds (=109 hours or 4.5 days). Probably this host is running more projects and is suffering from the "win98/Me time bug" or something like that I assume.
The same happens to one of my hosts (5000), that one is also claiming far to many credits.
It's actually the leave
)
It's actually the leave applications in memory bug. I turned that on once when i was running CPDN (don't anymore) and when a WU was paused the CPU timer kept increasing even when it wasn't the active unit. Yes.. the 98/ME bug. (I run ME)
That's the one I mean. The
)
That's the one I mean. The strange thing is, when the applications are paused when still in memory the time is increasing for Seti/Predictor "live" on the work tab. The time of Einstein look to be stopped, but when closing and restarting the Einstein application (or closing at 100% to be reported) the "paused" time is added after all.
This is at least the case for Boinc 4.60-4.62 and Einstein 4.71/72, running on a PIII 1GHz (WinME).