Windows vs Mac Validation
10 Jul 2006 15:46:39 UTC
Topic 191554
(moderation:
I have had a workunit fail validation against a Mac.
The third result was also sent to a Mac; I wonder whose result would've been validated if it had gone to a Windows box?
All three machines in the quorum used stock apps, and my host has never had validation problems.
If all machines are running stock apps why are different (invalid) results being returned?
Could this be a problem with the science apps?
EDIT : I also put a post in Crunchers Corner
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
Windows vs Mac Validation
)
Here is the other thread with a few more problem reports Mac vs. Windows
The bug hits only now and then, the affected Mac computers as well as the affected Windows boxes have mostly valid results except for those few.
RE: I have had a workunit
)
I CANNOT confirm this! I have 2 Macs running for weeks with the new client and only 1(!) WU produced an invalid result (the other client errors come from stopping this specific WU because the computer was overcommitted due to large downloads from Rosetta).
RE: I CANNOT confirm this!
)
That one invalid result is exactly what I talking about though. You were first paired up with a Windows host, the results were dissimilar so a third WU was sent out. This third host was another Windows host so naturally the two windows results were more alike and yours got tossed out.
If the different apps are returning slightly different results is the integrity of the research being compromised? Also it leads to wasted computer cycles and electricity for the result that gets tossed.
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
There is a problem indeed.
)
There is a problem indeed. It's not scientifically relevant, but confuses the validator. It only occurs with a few percent of WUs, and only if one task ran with a PPC Mac App and the other with an x86 App (on a SSE-CPU). I'm currently testing a PPC Mac App that should fix this.
BM
BM
RE: There is a problem
)
Thanks Bernd, you guys are always on top of things! Best of luck with the new app, and I hope this validation error doesn't cause too much of a headache for you.
Dave
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