memory errors

Question

Last night, while playing World of Warcraft, I recieved a sudden game shutdown, with an error stating that "somesuch" memory block: could not be accessed. This caused a corruption that's going to require me to reinstall the game.

I ran 7 passes of memtest, and that turned up the same two blocks each time. I ran the Windows Memory test and it told me all was fine. I've experienced inexplicable issues with this machine; random bluescreens, various shutdowns, etc. I've never had anything specific enough to pinpoint the memory.

My question is: short of swapping in one board at a time, is there any way to pinpoint which memory board is failed? With the way the machine is set up, it's a huge pain to open it, swap a board then hook it all up to test over and over. I've already spent far too much time inside this machine.

As always, I really appreciate the input I get on these message boards. Thanks in advance!

Boyd B · Answer

Hello Raelf, If you ran Memtest and it showed errors, the memory is defective. The Microsoft Memory test is not nearly as good as Memtest. I have had the same results as you mentioned, Memtest revealed defective memory, Micosoft Memory Test did not. I would remove all memory modules except for one and run Memtest. If no errors are found, remove that module, install another and run Memtest and continue until you find the module(s) that is defective. Replace it with the same as the original and run Memtest on the new module. When all the memory is verified to be free of defects, then the OS would need to be reinstalled and not just the application. With defective memory, all software on the system, and Windows in particular, can and does become corrupted. Hope this helps.

Raelf · Answer

Thank you, Boyd!

Yes, it is as I feared... a total reinstall. In order to use my warrantee, Dell is requiring that I run their memory utility to test the boards and the sockets... good practice, just annoying. Upside is, Dell will send me 4 new boards if any fail... downside is, I don't trust them. I am going to have to swap in one at a time, and test each of the new ones.

I'll post back when it's complete. Thanks again for the input!

Raelf · Answer

Just a final update for those who responded to this post, and anyone else experiencing intermittent, or a single, memory error(s).

I called Dell and they insist that if the dell diagnostic doesn't show it, it didn't happen. So I ran the Dell tool, and of course, no error. For laughs, I re-ran memtest... 10 passes overnight, and no errors!

The only thing I can figure is heat. The error did occur after a 6+ hour gaming session in a rather graphics intense encounter. The 720 H2C leaves the memory in a small passively cooled area with lots of unruly cables and no direct air flow. If not heat, an errant driver that might have locked down a cell or two on the memory (yeah, it's a stretch)...

Given that I wanted to do it anyhow, I re-routed all the cables I could and tiewrapped them back, reseated the memory, tuned the OC back to factory for CPU and memory, zero'd the drives with DBAN, repartitioned with Acronis Disk Director, reloaded Vista 64-bit and SP1, reloaded all drivers (including the Creative drivers from heck), left the PCCillin until afer I loaded all apps, and re-ran memtest. Clean. Probably extreme, but I know my hardware is ok, or I am as sure of it as I can be. My partitions are pretty now, too! :smileyvery-happy:

Before I re-up the OC, I am installing the corsair cooling fans over the memory. I just do not trust the air-flow in the memory area, even though I already installed an exhaust fan. This design begs for another 120MM fan behind the front drive slots blowing directly on the memory and rear drive slots, but as it would have a slightly obstructed intake and outflow, it would probably roar and need to be MB controlled... and I've been told there is no port on the MB to add a fan sync'd to the lower one, so the corsair cooler is my last option.

As always, comments are welcome and encouraged. If you read this far, I appreciate it!

Message Edited by Raelf on 05-30-2008 08:37 AM

Message Edited by Raelf on 05-30-2008 08:41 AM

Boyd B · Answer

Hello again Raelf, I am glad to hear you found no errors with Memtest and all the other hardware appears to be operating as it should. While you can never count on anything 100%, I have never had Memtest to render false or misleading test results. If this is the case, you may have narrowed the issue down as being heat related and taken steps to resolve the heat issue on the memory. If all this corrective measure does not correct the problem, then it may be software related as you mentioned. I have found errant drivers before that work for a while and then act up. If all else fails, even if you have the latest ones, you might download another copy of the drivers and reinstall them. A problem as yours really comes down to trial and error and is very difficult to locate and resolve. But it can be done with perseverence and patience. I have found that trying just one attempted correction at a time, you will be more likely to locate and resolve the issue, rather than applying 2 or more corrections at the same time. This takes much longer since it is an intermittent problem, but will be the best way to determine what is causing the problem. Good luck of resolving the problem. Hope this helps.

Locked Topics – Desktops General

memory errors

Was this post helpful?