Issues during undervolting

[Queelis]

Hello there.

This a part of my undervolting experience, which was described here:
http://forums.hexus.net/hexus-hardwa...adventure.html

So, I've had some issues during the testing I'm hoping you might help me solve.
First was after completing the 8 hour Linpack stability test. After it had completed, I was playing some Heroes III (oldschool, I know ) (or was it when I was converting some audio? Can't remember now), that the system BSODed on me. Ookay, rebooted, got the message about an event "BlueScreen" and some error codes, along with links to some files (a .dmp, .xml and .txt. The .xml and .txt files weren't there, and opening the .dmp file with notepad didn't go very well). So, I've no files which "help to describe the problem" (as the error report says), but did find the .dmp file, so if anyone knows how to open it and get some info out of it, fire away


Second issue (or, issues) came up when I was getting the same-day temperature comparisons. When the CPU was loaded (again, with OCCT's Linpack test) I was also browsing and playing music. At times the music would be, well, "invaded" with static-like noise (which stopped with the music, and it wasn't that the files would be bad, later on they played cleanly), and eventually WinAmp shut down with an (0xc0000005) error.

While browsing, Firefox's tab or status bars would sometimes black out and the pages wouldn't be displayed properly (distorted). Other browsers would do the same. When I had finally rebooted later on I found out that my NoScript exceptions (the whole whitelist) were gone.

Later on my mistake was to look up the error on google, and I've gone to some phoney site pretending to be Microsoft, and then Avast also went with an (0xc0000005) error. I immediately thought this was some kind of a virus and restarted Avast, which showed that the shields (active scans) were set to off. I didn't do that! Then I set it to scan the whole computer, but in the beginning of the test it shut down with the same error again.
Rebooted with a diagnostic boot (where only "necessary" services are started), Avast scan completed with no infection, and nothing bad has happened ever since. Hope that's over.

So, if any of you have any ideas about the issues, I'd like to hear them

[Halk]

It sounds as if the Linpack stability test wasn't testing your system in the same way as a 3D game was. It's not uncommon for a system to be Prime stable but not 3D stable, even though it's only a CPU overclock.

The reason for the disconnect between this is because CPUs have evolved over time. It used to be the case that the CPU was just the processing and overclocking it affected little else. However now overclocking it affects the stability of a larger number of things in the system, which may not be exposed by a simple CPU stress test.

[scaryjim]

Just thinking out loud here, but it sounds to me like it might have been a memory issue from the symptoms. It could be that the reduced core voltage + the additional extra load was interfering in some odd way with the memory controller, perhaps? *shrug* Hope it stays away from now on

[Halk]

That's very possible, the memory controller is integrated into CPUs now.