Unlucky validation error :(
9 May 2007 18:10:08 UTC
Topic 192713
(moderation:
See here:
The first host successfully finished the WU with the first(buggy) app which caused validation errors quite often.
The second host didn't finish in time.
Then my host(the third one) got the WU and crunched it with the Linux beta app.
The validator found a difference, so it sent it out again to the fourth host, that didn't finish so far.
Some hours later the second host finished the WU(with a buggy app too). The validator found, that buggy results from host no. 1 and host no. 2 are equal/similar.
These two got credits, my host didn't. Well that's wasting time, because another host ist crunching this WU again for hours and also will get no credits for it, as it is using the newer bugfixed app.
In this special case the validator should wait until all results are back and then decide which results are valid. 2:2, ok send it out to a fifth host. ;)
No not really. ;)
I know, this is a pretty rare situation and the results of the first app are not necessarily buggy, just different.
I hope, the wasted time in the S5R2 run will shortly tend to zero.
cu,
Michael
Unlucky validation error :(
)
i have something similar, but it's invalid due to using another OS
see here
i have XP home and they crunched it using darwin 8.9.1
well, darwin=darwin=>Valid ; XPhomeDarwin=>InValid :(
GOD is only word
Yep, Darwin and XP don't seem
)
Yep, Darwin and XP don't seem to go too well together, especially with the old Windows app. I had a case when I crunched with a Darwin and a Linux host with my XP box and their results validated while mine didn't.
Michael, that is really quite an extreme case you describe there...
another workunit with
)
another workunit with validation problems
im lucky, because third WU was sent to WinXP
but i have another WU (not crunched yet) waiting to be validated with Darwin
should i abort it or crunch it and hope, that third WU will be sent to Win?
GOD is only word
RE: but i have another WU
)
Hi!
I'd let it crunch. It's not that Darwin and XP never agree on a workunit, my Darwin system has validated successfuly against XP systems running 4.17 and even 4.13 of the app.
CU
BRM
same problem, also on May
)
same problem, also on May 8....
It does say success, eh? No credit was granted.
D
application Hierarchical all-sky pulsar search
created 25 Apr 2007 2:49:19 UTC
name h1_0253.75_S5R2__74_S5R2c
canonical result 83600203
granted credit 396.83
minimum quorum 2
initial replication 3
max # of error/total/success results 20, 20, 20
Result ID
click for details Computer Sent Time reported
or deadline
explain Server state
explain Outcome
explain Client state
explain CPU time (sec) claimed credit granted credit
83600202 812888 25 Apr 2007 2:49:21 UTC 25 Apr 2007 19:06:24 UTC Over Client error Compute error 42,508.57 231.53 ---
83600203 696547 25 Apr 2007 6:50:16 UTC 29 Apr 2007 4:12:32 UTC Over Success Done 99,855.94 396.83 396.83
83625203 834175 25 Apr 2007 23:10:00 UTC 8 May 2007 9:31:30 UTC Over Success Done 355,149.33 396.83 0.00
84067790 864718 8 May 2007 11:03:16 UTC 10 May 2007 7:33:13 UTC Over Success Done 128,015.72 396.83 396.83
Success is everything before
)
Success is everything before validation. It does not mean valid. It's not the best english word to use, but for translating purposes to other languages it's the better one that Berkeley found.
They have been trying to change that state if a result is invalid in the last couple of backend revisions, but as you see it does not always work.
http://einstein.phys.uwm.edu/
)
http://einsteinathome.org/workunit/33671808
Paired my windows with a Mac not good, now a Linux got it. Wonder who get the credit and who get none. Hopefully the staff will correct this soon, don´t care so much about the credit but think the wasted time both for me and the project sucks.
Sorry for me the Mac and
)
Sorry for me the Mac and Linux got it and not my Win.
Here's a good one. All three
)
Here's a good one. All three hosts are Core architecture, one running XP, an other OSX, and mine LINUX 2.6.20.
http://einsteinathome.org/workunit/33738056
Guess which one doesn't validate? ;-)
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
RE: Here's a good one. All
)
This points to a problem with the mathlib linked to the Win-app.
But this also can mean, the Win mathlib is the more accurate one. ;-)
cu,
Michael