Trying to pin down (memory?) error

Urban Reflex · 21 Feb 2009 at 13:31

Okay guys, so my computer has a tendency to freeze every so often. It seems to happen most when I'm doing something stressful, like playing a game or encoding a video. So I figure it's either memory, CPU, or motherboard.

I've run several scenarios of Memtest (chips in different slots on the motherboard, 1 chip at a time, etc) and only two tests came up with errors. Each test had only one error in it, and a friend seems to think it might just be a random floating point error or something.

The failures were on the same chip of RAM, but it was in a different slot in each test. The "failing address" was the same: "0000c1ae82c - 193.8MB", and so was the "Err-bits": 80000000.

Other than this, I did have one test that locked up the PC, but no errors were shown. This was with both my RAM sticks in at the time. All tests have been over 8 hours long.

So any ideas what to do next?

Urban Reflex · 21 Feb 2009 at 21:08

Aye, done that. It was setting it to 1.6 I think, upped to 1.8 as it should be.

Urban Reflex · 22 Feb 2009 at 01:52

It's 2GB of this: http://www.ocztechnology.com/products/memory/ocz_ddr3_pc3_10666_platinum_edition

Also, it's an Asus Striker II Extreme board with a Core 2 Duo E8400. Corsair TX750 PSU so plenty of power. CPU temp isn't going over 50-60 after a few hours of Orthos. Everything is stock, no overclock of any kind. Volts are within recommended levels. Windows Vista HP 64-bit (Though has frozen outside of Windows) and an 8800 GTS 512MB.

It seems to happen a lot when I set the 32M test running on Super Pi Mod. Though Task Manager isn't showing any significant leap in resource usage.

Urban Reflex · 24 Feb 2009 at 13:35

Nah, it's designed to run at 1.8V :/ I've had it at 1.8 ever since I got it. I might try upping the voltage a little more, see what happens. 1.9 or something...

I thought this issue may be what was stopping me from overclocking. However, I tried doing a little OCing with only the one "good" RAM chip and and I'm getting freezes like I did before. This is what lends me to think it might be something to do with my motherboard being overloaded somehow. Is that even realistic?

Urban Reflex · 26 Feb 2009 at 13:55

Hmm. I've tried setting the RAM voltages to auto, 1.6, 1.7, 1.8, 1.9, 2.0, but still the same thing happens. Even tried altering the north and south bridge volts but again, I'm getting freezes.

I tried to boot my PC with the one stick of RAM I figured had no issues, but I got my first freeze with that as soon as it got into Windows. Considering it's my first freeze with just that one stick in, it might be coincidence. However, it's exactly the same type of problem I'm having as with the other stick, so that seems to coincidental...

I'm running out of ideas now :/

Urban Reflex · 26 Feb 2009 at 20:19

Unfortunately, no. That would be the downfall of DDR3 - it's expensive, so no one has it

OCUK have been kind enough to issue me an RMA number though, so we'll see what their techs make of the RAM. If it's not faulty, I'll be RMAing my motherboard next!

Urban Reflex · 7 Mar 2009 at 02:13

I'd just like to say a big thank you to everyone who helped me on this thread. Your input is as ever, appreciated.

OCUK were lightening fast with their RMA. I sent the RAM off on Monday, and received my replacement on Friday. Honestly, I am just stunned by the speed of you guys. I know it helps that the RAM came back in stock at just the right time but in all my other, non OCUK RMA experiences it's taken about 2 weeks at least to go through all the red tape.

So yes, a class experience. And the new RAM is actually running at timings faster than it's advertised at. I love you OCUK