I have an MSI 760GM P35 board and an Athlon II x2 250. After overclocking to 3.7Ghz with this board it'd get random errors like IRQL_not_Even or other seemingly random memory related errors. It didn't start happening until the OC. This was with 32bit XP on one HD. I decided to put Win 7 64bit on my blank HD to take full advantage of my 4 gig of RAM and possibly update to 16Gb in the future as the board supports it. Now, I can rule out drivers as both HD's have different drivers installed. With the Win 7 I can hear the system "shutting off" or making the sound it makes when the system is shut off. Like a TV being turned off. But nothing on screen is noticed. So, with XP it'd crash with BSOD, with Win 7 it seems to ignore the errors and continue running (but every couple hours or so it makes that sound where it sounds like it just powered off).
It took FOREVER to get this to 3.7Ghz on this board. I tried using a backup 3.6Ghz OC Profile to see if it'd correct the issue, but with Win 7 I can still hear that random "shut off" sound occurring. I had this chip to 3.72 Ghz on my old Biostar A780L3L board with no issues. Any idea what might be wrong?
The Win 7 install was as clean of an install as I'm gonna get. It was a blank HD with a CD install. The XP setup is on another HD, completely independent of each other. It seems the errors are related to my OC, or possibly the DDR3 settings of the OC? Again, the errors on the XP HD were completely random and were almost never the same. But seemed to point to memory related errors. The Win 7 setup has NOT crashed, but will make a shut off sound once in awhile. I have the clocks for memory set to AUTO. But the 1066 was turned down to 800. Would limiting the clocks on the RAM sticks help?
The OCing was done on the XP set up. At 3.7Ghz, the BSOD started. I upped the voltage from the 1.408 to around 1.450 to get it stable enough to enter XP. It passed Prime95 for over an hour and passed Memtest86+ with no errors. After awhile while seemingly doing nothing, it'd crash with BSOD and random errors. I decided to go ahead and load Win 7 on my BLANK HD as I'd been meaning to do so. And also see if it would rule out drivers causing the BSOD. I'm 100% confident that the error that's happening is related to the OC and not the setup. I'm wondering if others have experienced the same symptoms after overclocking and how they remedied it?
Also, I tested the setup on Win 7's rating system and DDR3, GTX 550Ti, and processor all got 7.3 for ratings.
Overclock need high stability system, I think you need choose one OS win7 or XP, don't mix your oc use dual OS . That can cause crash. Just use one system OS.
Also reading the Ram will different for xp and win7,..
If you want use dual OS ran at default (auto)
I wasn't running a dual boot. Disconnecting the XP drive while using the Win 7 drive stopped the power down sound. Setting the HD from "20min" to "Never" did NOT fix the problem. My guess is something with Win 7 (Vista does it too apparently from what I've read) accesses the HD and shuts it off whether or not your accessing files directly or not. There might be a way to manually stop it (like the remove hardware safely) but I just disconnected the SATA driver as there's no need to waste the lifespan by running it unnecessarily.
My system was crashing due to a fluctuation in CPU voltage that wasn't present before, therefore I didn't realize it was throttling. Thinking Cool and Quiet or something else that throttles CPU voltage isn't turning off. Possibly MSI's "Phase Control" as it's acting like it's on. I've read where C&Q has had issues like that. All my other settings are remaining the same, and volt meter confirms power supply is constant as well as the CPU speed. I'll post in a new thread as this one can be closed now.
I really don't know where to start with this question. but I know that if you have that chip on one board and burned it in on that then it is that. If you put on new board the chances of getting same performance is likely and better not likely. In fact you may have to use some of the same settings you used on your old board. and then of coarse alway's inspect you socket and cpu processor for physical damage if there is no visual damage you can more than likely you can rule out bad bad cpu if the board does not have visual damage you can rule that socket thoery out. What it sounds like to me is you have a voltage regulator issue there. Are you sure the switching sound is not coming from you *PSU?
UPDATE: It seems the voltage drop was from CPU-Z not reporting correctly in certain ranges. As if it had "holes" where it'd fluctuate. I never had issues with CPU-Z at lower, near stock voltages. I downloaded Everest and it is indeed showing a non-fluctuating (it is refreshing the value), correct voltage when put side by side to CPU-Z. I've managed a fairly stable 3.9Ghz that hasn't crashed yet, but has issues with Prime95. I'm testing this setting and monitoring temps before upping the voltage any higher. http://valid.canardpc.com/show_oc.php?id=2010732 and CPU-Z is "out" of it's hole as voltage is right at 1.445 in the BIOS.
I read somewhere on a Microsoft link that (a program I can't remember at the moment) they have a feature/program (might be the indexing service) that actively monitors drives for changes, so I'm assuming it was shutting the spare drive down after accessing it as a safety or energy saving feature. I just disconnected the SATA connector as I'm not using that drive for the time being. I'm thinking an easier workaround to avoiding that "sound" it was making would be using the feature that you use to "Safely Remove USB" sticks like what was mandatory on the old Windows ME for removing a USB stick. That way it wouldn't be trying to read a spare drive constantly for updated info.
EDIT: Running both PSU at once, one is running VC, other is handling fans, CPU, HD, etc.
I decided to run a spare 2nd PSU to take the load off GTX 550Ti to see if any difference, so I don't believe the Thermaltake 600w PSU I've been using is bad as no difference occurred. I do believe the HD was being turned off/on by Win 7.
I don't believe there's any socket damage, nor any chip damage either. I bough the board new, the chip took to it as I finally managed a fairly high OC. It's really sensitive to the RAM timings, which were holding it back at 3.72Ghz forever.
The IRQL_NOT_EVEN errors were related to too tight RAM timings. Had nothing to do with drivers. Going from a 6-6-6-15 to a 7-7-7-17-22 1T fixed this and allowed for stability.