Compute Errors on New Global Correlations S5 Tasks
23 Mar 2010 2:37:43 UTC
Topic 194836
(moderation:
I have multiple Exactly Identical I7-860 Computers, one of them is failing every Global Correlations S5 task but all other tasks work just fine. I have reformatted the drive, reloaded Windows 7 Ultimate, I even replaced the I7-860 CPU with a brand new processor today, same compute error. I can tell it is going to fail, when the project starts the elapsed time starts but the Progress never moves from 0000, nada. I am at a loss - the only item I have not changed is the Motherboard. I also relocated the computer to several different network locations. The others are processing those file with no apparent issue. I need some new ideas or even a fix for this problem I do not know enough about how the program works to understand what is wrong. ONLY the Global Correlations S5 tasks fail.
Mike Kubanek
Rank = 44, Current Average = 20,260 Hours = 12,198,007
Mike Kubanek
Compute Errors on New Global Correlations S5 Tasks
)
Is the AV software the same on all those hosts?
If you'd post a link to the i7 in question, we could check the stderr output of the failed tasks.
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
It's this one. All with "-
)
It's this one.
All with "- exit code -1073741819 (0xc0000005)"
Which could be driver based. Make sure you have updated your motherboard drivers, video card drivers and other hardware drivers. Make sure they are compatible with your operating system.
RE: It's this one. All
)
Man I have never worked so hard on a simple problem in my life. After replacing the processor, then today the motherboard, then even the OS from a Win 7 Ultimate to a Vista Ultimate I finally saw one single project work, but all the rest of the 8 failed, so I had an indication of something related to timing. I then realized the memory was common from the old Motherboard to the new I7 - 860 & Mobo, and then hit on the actual timing issue. I changed from the 1333 memory to 1600 memory, the problem is solved. I appreciate the help it sort of made me realize it was probably hardware. The down side is I spent a lot to find this, the up side is I will have a backup system - something I needed. I try real hard to stay on the first page of the top performers (top 50) and with what I had one little hickup like these new projects not working on one system will knock me off the top 50 pretty fast. My sincere thanks to Jord & Gruß I was really stuck on this one!!!
Keep on Helping our friends who help so much to find answers to the Questions about our Universe.
Mike Kubanek
www.racing97.com
Mike Kubanek
That's what I get for giving
)
That's what I get for giving the condensed version of what the error message can be. :-(
I'm sorry. Now I do wish I'd pointed you to the FAQ I have on that, as there it gives you a checklist, which includes faulty RAM.
The dynamic duo. :-D
Too bad that it's Jord & Gundolf. Gruß is German for Greetings. :-)
RE: It's this one. All
)
I think this info originates from old times where the graphics was part of the science app and a failure in the graphics thread crashed the whole application. That shouldn't be the case anymore. But it could still very well be a hardware problem, as it was in the current case.
The current S5GCE Window App doesn't give more information in case of a "general access violation" (which is 0xc0000005). The next app version should, we are working on it frantically.
BM
BM
RE: I think this info
)
The "general access violation" (0xc0000005) is a Windows error as well, with causes such as I detailed in the FAQ I pointed to. Overclocking, corrupt device drivers, out-of-date DirectX, out-of-date hardware drivers, faulty RAM, malware/spyware are the most common causes for this error to pop up (in a blue screen of death).
While it may have been caused by the graphics worker thread in the past, it wasn't in this case. I looked through several of his stderr.txt reports and didn't find that the cause resided in either the graphics thread or the worker thread with houghmap.c (which are the two cases you seem to point to).
Still truly sorry I didn't point out all of the obvious causes. :-(