This sounds bad... I thought at least my Win/Intel box would be save. It has gotten away clean until now... I certainly keep my fingers crossed that you guys get your problems sorted out asap, or have more luck with your WUs...
This sounds bad... I thought at least my Win/Intel box would be save. It has gotten away clean until now... I certainly keep my fingers crossed that you guys get your problems sorted out asap, or have more luck with your WUs...
If it's any reassurance, I run exclusively Win/Intel - and so far, touch wood, I have had no errors at all on S5R2, from Celeron 400 to Xeon E5320.
I have 5 computers all on XP or Vista. None have returned with error, yet. P4, P4HT, and 3 Duo Cores. Einstein is my main project, although I do crunch a couple other things, but in lesser amounts.
So, there is some stability out there. I am running 5.8.16 on them all, installed as Server, and do not run the BOINC Manager, nor screen savers.
One thing to keep in mind, which I pointed out in the other thread, is this could be as simple as some of the datapacks which have been produced are faulty, and not a host side problem at all.
The info I found regarding exit code 99 talked about faulty work coming down from the project as the cause, but those conversations were not about EAH so I'm not 100% sure how that applies here.
Up until this event, none of my WinBoxes has had any trouble with S5R2.
Up until this event, none of my WinBoxes has had any trouble with S5R2.
Repeated and bolded...
Hopefully it was just a bad set of work, however the initial result that started all of it has a Linux host with a SIGABRT... :( Could be that they are related with the OpenGL and/or video driver...?
As an aside, it is really troubling to me about the "recursive error" in the output...
Edit: Although looking at it more, that could be just a downstream error caused by a failure ahead of it with the XLAL error... :shrug: Dunno. Most of the stuff is over my head. I really need to get back into Java and C++ again....
Agreed, and I'm not going to rule out anything happening on a beta run. ;-)
The main reason I posted my event was I saw some similarities between it and some of the problems the Linux folks have been having.
From what I read it seems clear there is an OpenGL related problem for Linux, but it may be that it looks worse than it really is because of a general cross platform problem 'amplifying' the situation on Linux.
FWIW, I haven't seen or read about a Mac host getting bitten by either, but that mean it hasn't happened.
Alinator
LOL, yeah my programming skills are pretty rusty too. They're a lot sharper now than they were before I started playing with BOINC though. :-)
No, I haven't noticed a WU crashing on a Mac, either. But unfortunately that doesn't really help most of us. If it was between Win and Linux at least some of us could just switch, and it wouldn't cost money in most cases (assuming everyone has an old Windows CD lying in a corner, but with MS putting it on the majority of new computers I think most people do). But with Mac it's a different story, as you need the box to get the OS... So, we'll have to sit this out ;-) or hope we can improve our chances a bit with some clever workarounds.
LOL, yeah my programming skills are pretty rusty too. They're a lot sharper now than they were before I started playing with BOINC though. :-)
Sadly, I allowed myself to get into a comfortable position, thinking that I had a relatively secure job. I didn't keep myself current...
At any rate, what was posted indicates that they're passing by reference using C++, so that immediately brings into my mind memory problems or pointer problems... Of course then there's the whole "invalid argument" bit and then what was in your results..."input domain error"...
I may be enthused about looking into it more...after I finish my Java final... although without source code it's hard to really say...
Edit: I am assuming C++, but it could be straight old-fashioned C...
FYI, I hate semicolon languages... They may be powerful, but they get complex real quick... :sigh:
Yeah, I'm working this same
)
Yeah, I'm working this same problem on one of my WinBoxes in the 'Unpleasant Surprise' thread.
No definitive answer, but I have posted some post mortem info from when it happened.
You could take a look and see how it compares to your event.
Alinator
This sounds bad... I thought
)
This sounds bad... I thought at least my Win/Intel box would be save. It has gotten away clean until now... I certainly keep my fingers crossed that you guys get your problems sorted out asap, or have more luck with your WUs...
RE: This sounds bad... I
)
If it's any reassurance, I run exclusively Win/Intel - and so far, touch wood, I have had no errors at all on S5R2, from Celeron 400 to Xeon E5320.
I have 5 computers all on XP
)
I have 5 computers all on XP or Vista. None have returned with error, yet. P4, P4HT, and 3 Duo Cores. Einstein is my main project, although I do crunch a couple other things, but in lesser amounts.
So, there is some stability out there. I am running 5.8.16 on them all, installed as Server, and do not run the BOINC Manager, nor screen savers.
One thing to keep in mind,
)
One thing to keep in mind, which I pointed out in the other thread, is this could be as simple as some of the datapacks which have been produced are faulty, and not a host side problem at all.
The info I found regarding exit code 99 talked about faulty work coming down from the project as the cause, but those conversations were not about EAH so I'm not 100% sure how that applies here.
Up until this event, none of my WinBoxes has had any trouble with S5R2.
Alinator
RE: Up until this event,
)
Repeated and bolded...
Hopefully it was just a bad set of work, however the initial result that started all of it has a Linux host with a SIGABRT... :( Could be that they are related with the OpenGL and/or video driver...?
As an aside, it is really troubling to me about the "recursive error" in the output...
Edit: Although looking at it more, that could be just a downstream error caused by a failure ahead of it with the XLAL error... :shrug: Dunno. Most of the stuff is over my head. I really need to get back into Java and C++ again....
Agreed, and I'm not going to
)
Agreed, and I'm not going to rule out anything happening on a beta run. ;-)
The main reason I posted my event was I saw some similarities between it and some of the problems the Linux folks have been having.
From what I read it seems clear there is an OpenGL related problem for Linux, but it may be that it looks worse than it really is because of a general cross platform problem 'amplifying' the situation on Linux.
FWIW, I haven't seen or read about a Mac host getting bitten by either, but that mean it hasn't happened.
Alinator
LOL, yeah my programming skills are pretty rusty too. They're a lot sharper now than they were before I started playing with BOINC though. :-)
No, I haven't noticed a WU
)
No, I haven't noticed a WU crashing on a Mac, either. But unfortunately that doesn't really help most of us. If it was between Win and Linux at least some of us could just switch, and it wouldn't cost money in most cases (assuming everyone has an old Windows CD lying in a corner, but with MS putting it on the majority of new computers I think most people do). But with Mac it's a different story, as you need the box to get the OS... So, we'll have to sit this out ;-) or hope we can improve our chances a bit with some clever workarounds.
RE: LOL, yeah my
)
Sadly, I allowed myself to get into a comfortable position, thinking that I had a relatively secure job. I didn't keep myself current...
At any rate, what was posted indicates that they're passing by reference using C++, so that immediately brings into my mind memory problems or pointer problems... Of course then there's the whole "invalid argument" bit and then what was in your results..."input domain error"...
I may be enthused about looking into it more...after I finish my Java final... although without source code it's hard to really say...
Edit: I am assuming C++, but it could be straight old-fashioned C...
FYI, I hate semicolon languages... They may be powerful, but they get complex real quick... :sigh:
LOL.... Yeah I hear ya.
)
LOL....
Yeah I hear ya. With great power comes great responsibility (to not screw up the coding)! ;-)
Assembler takes that from the sublime to the ridiculous. :-)
It sure is small and fast when you do it right, though.
Alinator