This looks like a Computation error.
Please provide a link to the workunit.
Something beginning with https://www.worldcommunitygrid.org/contribution/workunit/…

Al, Mike is a good guy. I feel like I'm often trying to teach him precisely where to look.

EDIT: I think Mike hates me for that, because it seems he's never having a real conversation with me (and as long as there's no need, that's fine with me). It's just that I hate to see anyone providing false information or wrong explanations (and I think I know better.) In the meantime I'm just trying to learn from the situation, trying to express myself in the right way and trying to avoid treading on long toes (= stepping on toes).

Adri

Okay Blount, thanks. You have provided the link to the error log of your task, it helps a bit.

It would've been better if you had provided the link to the workunit. (You can find that link if you click on the name of your task on the Results page.)
Example


Now, from your Results log, what strikes one most is the exit code 3221225477. I believe I can only find two other posts on the forum with that exit code in combination with African Rainfall:

this Crystal Pellet's post and

this erich56's post.

Crystal Pellet's conclusion is: "The amount of memory is not the problem."
However, "I think running ARP's together with 4-core VM's for LHC-ATLAS is sometimes conflicting." he says.

And erich56 concludes with: "this a a fairly old PC. So I am not too surprised that it cannot meet the challenges which ARP is posing to a system."

I cannot see much more in my crystall ball, Blount, it's been hazy all day. I wonder why.

Adri,

Good follow-up post above regarding Blount's result error, but as another one has shown up I thought I'd just add a little bit of data regarding possible completion of the work unit(s) concerned :-)

Firstly, on Windows 3221225477 (0xc0000005) is specifically Access Violation (as per the report itself!); I suspect the scarcity of examples may have as much to do with how users post about errors as anything else (and the hex format report seems more commonly used here and elsewhere...)

Now, as for the two work units with Errors now showing up in this thread... Unlike the [usually] identical stack traces we're used to for genuinely stuck unit SIGSEGVs on Linux, these two produced very different diagnostics, and Kirel2's stack trace is really strange!

When I feel I need to reply to a post with only a result number, I've been using the "Swagger" GUI interface to the API to find out work unit names; in this case it seems Blount's WU is 152524445 (ARP1_0034797_127). Fortunately, that one seems to have a good chance of completing (only one No Reply!).

However, Kirel2's WU is on it's last chance (having 2 No Reply, 1 User Aborted and one Pending Validation as well as the Error...) so ARP1_0019859_126 may end up stalled.

I''ll watch out for these two work units in case either does stall...

Hope that's of some use to someone :-)

Cheers - Al.

P.S. Your edit to your reply to my earlier post exactly tags my feelings on that sort of thing. However, I feel there are situations where it just isn't worth adding to the post count, and others (usually on other BOINC sites) where one knows a correction will be met with a hostile response...

My crunching machine is just my home desktop; let me know if you need any hardware specs or any additional info about the system. This is the first time since the WCG test WUs started going out that I've noticed any kind of issue. My machine was able to process https://www.worldcommunitygrid.org/contribution/workunit/152550562 with no problems, for example.