PDA

View Full Version : LW Blackout Hard Crash



Surrealist.
07-10-2007, 03:27 PM
This has happened on a few different scenes even simple ones.

At first I thought my CPU was overheating because it was happening on a large scene. Now it even happens on a smaller scene. My CPU is well ventilated. I have the box open with a large fan blowing cool air in.

Wondering what I should do. Any ideas? Has my CPU taken a dump? Is it bad RAM? My MB? Graphics Card? Not sure.

I am using:

AMD FX55
3Gigs Ram
Asus 8NSLI Delux MB
WinFast 128 meg Graphics card
LW 9.2

Just wondering where to start troubleshooting. Anyone ever have this problem?

Thanks.

Extent
07-10-2007, 03:32 PM
Checking for failing ram is one of the simplest things, and it's what I would do first

http://www.memtest86.com/download.html

Surrealist.
07-10-2007, 05:13 PM
Definitely looking like a RAM problem so far. First time Test 5 crashed on errors. I removed two new modules that where possibly incompatible - different vendor.

Did the test again (I had been having this problem before installing new modules) and the original RAM crashed on a Test 6 with errors.

Thanks for the idea and the link. We'll see how it turns out.

Extent
07-10-2007, 07:08 PM
Wow, I've never seen memtest actually crash on errors, even after hundreds of occurances on really bad ram.

Surrealist.
07-10-2007, 11:26 PM
Wow, I've never seen memtest actually crash on errors, even after hundreds of occurances on really bad ram.

Yeah. In fact the first time it was a hard crash blackout so I was like, "OK, smoking gun!"

Thanks again for the tip. I would have never known how to handle it.

Surrealist.
07-13-2007, 12:50 AM
Well, as it turns out, it seems to be a CPU overheating problem, probably something with my MB. I did more RAM tests and they passed and I was still getting a power down trying to render a scene. Went down and got a new board - so far so good.

Extent
07-13-2007, 04:11 PM
Ya heating can cause corruption too, usually that's easy to spot with a temp monitor and test with a box fan. I've had memory controllers go out on me in the past as well.

Sensei
07-13-2007, 09:01 PM
Once upon a time one single hair landed on my old motherboard memory slot and giving problems with RAM.. It was pushed perpendicular to slot, and visible just after unscrewing motherboard from the case..

Surrealist.
07-14-2007, 12:22 AM
Yes. I pulled out the RAM and put it back in and I was no longer getting the errors on the RAM test. Must have been a bad connection. This led me to believe I had fixed it then I had a crash again. After talking to a few tech people I surmised that it was overheating. With my new MB - and CPU - things are going well, and much cooler.

Thanks for the help. :)

Extent
07-14-2007, 02:31 AM
How long did you let the memory test run? If I'm real suspicious about something I'll let it burn in at least overnight if I don't see any errors pop up. If you ever see even one error, even in a very long time, it's not ok.

I'd keep your old setup around and try and get it working for a render node if you can. Try reapplying your thermal paste, make sure that the chipset cooler is seated properly. Run some high load stuff and see where the heat is too much. If it was just overheating there's probably still some life left in it.

Surrealist.
07-14-2007, 06:17 AM
Yes, I have looked at that as a possibility and I agree on the RAM. I am in the middle of a project and I don't not have the time to fuss, but plan to perhaps build it in a real cheep box when I get the time.

Thanks again.