Machine Check Exceptions when OCing: What does it mean?

D

djBon2112

Supreme [H]ardness
Joined
Jul 29, 2006
Messages
5,279
I know what they mean from a technical standpoint, but what does it really MEAN to get one of these BSODs when you're overclocking? I bumped my Q6600 up to 3.8 last night, and first run it did 28 minutes of Prime95 multicore without error or locking up (I stopped it); however, the second time I tried it locked up with this BSOD. Changing the voltage up or down didn't make a difference; I'd still get this BSOD at about 3 minutes into Prime95.

What does this mean in terms of stability? Is there anything I can do to help the situation/make it stable?

I'm on an Abit IP35 Pro, clock was 423*9 to get 3.8 (it was 1.5 volts before vdroop average; i tried 2 settings up and down to no avail; 1.5 was what got the 28 minutes), currently running at 400*9 stable (1.375 volts), RAM always at 1:1 with very loose timings for the clock. My temps were never above 60 degrees at the highest voltage I used (they hit 59 and quickly to 60 but didn't stay there).

Any advice would be appreciated!
 
Anyone, any ideas? I pointed a fan (well, two) at the Vregs and Northbridge and it did 3.8 stable for an hour (it's now my 24/7 clock; temps are fine), but 3.9, EXACT same thing; any voltage results in this BSOD. Since I can't check the temps of my NB in Windows (and hence under load) I'm shooting blind in this area.
 
I am curious about this as well. If you get a good look at the error message before the computer reboots you could diagnose it. At that speed i would imagine its the FSB, try a little more juice to the chipset and FSB.
 
I wrote down the codes, but I have no idea how to debug them. FSB volts does sound like a possible cause; that kind of thing is all I'm really looking for. :p Kinda like how a Page Fault in Nonpaged Area means "RAM timings too tight/voltage too low".

EDIT: Tried bumping up the FSB (MCH) volts, didn't do anything. Still got the error. Oh well....
 
djBon2112 said:
I wrote down the codes, but I have no idea how to debug them. FSB volts does sound like a possible cause; that kind of thing is all I'm really looking for. :p Kinda like how a Page Fault in Nonpaged Area means "RAM timings too tight/voltage too low".

EDIT: Tried bumping up the FSB (MCH) volts, didn't do anything. Still got the error. Oh well....
Click to expand...

Its because nobody loves you, least of all the computing, and oc'ing gods :p

<3, now get you ass back home!
 
Yeah dude, you're not stable, simple as that, no troubleshooting besides bios adjustments, no window debug will make an unstable CPU, stable.

I know you like 3.8Ghz, but if it isn't stable, then back it down, cause you're not gonna find a solution to a BSOD if it's being caused by your cpu, especially after bios adjustments to try to get it stable.

So go ahead, change your little sig, back it down to a stable speed, run prime, get your stable speed and put it in your sig and enjoy :)
 
I saw this error and it usually mean the board is faulty or overstressed.

 
I pointed a single fan at the NB heatsink and now I can do 3.8 stable for 4+ hours (read the update post), instead of BSODing after 5 seconds without the fan. I think that points to something OTHER than the CPU clock.

Guys, I KNOW that "turn the OC down" is the simple answer, that's not what I'm looking for. I'm not a complete idiot. I'm looking for reasons WHY it's BSODing instead of simply failing Prime, etc. If you're not going to aid in that area, please stay out of the thread.
 
Does anyone think it might be the NB temp? I can't actively check my NB temp (it shows up in BOIS but not in Windows), so I don't know how hot it's getting. A fan pointing at it obviously helped, so it must be something in this area. Do you guys think watercooling my NB will help this?
 
Improving the NB cooling helps somewhat, as you've already shown us. Why not try upgrading the cooler to say, a Thermalright HR-05 SLI IFX?
 
djBon2112 said:
Does anyone think it might be the NB temp? I can't actively check my NB temp (it shows up in BOIS but not in Windows), so I don't know how hot it's getting. A fan pointing at it obviously helped, so it must be something in this area. Do you guys think watercooling my NB will help this?
Click to expand...

By pointing a fan directly on it and getting the system more stable, I'd say that would be a good presumption.

Watercooling helps everything... :D

Doesn't the IP35PRO have heatpipes for the North and south bridge chips??? Not sure watercooling the Northbridge by removing the heatpipe/heatsink on it will help the other components that heatpipe system cools.

How would one watercool the NB on an IP35? (without screwing with the heatpipe/heatsink setup for the rest of the board?
 
^^ I've got some heatsinks from my old board that I can use for the SB/Vregs; just hoping they're good enough!
 
The IP35 Pro has a issue with poor heatpipe installation. Sometime, they are crooked so it's not in full contact with the chips so you need to remove it, straighten it and reinstall with AS5. I bet this will help you improve the temps :)

 
You might give a try for stability with a one step down on the voltage to the north bridge. Maybe strike a balance. I find it odd they'd have such an incapable cooler for the north bridge.
 
djBon2112 said:
I know what they mean from a technical standpoint, but what does it really MEAN to get one of these BSODs when you're overclocking? I bumped my Q6600 up to 3.8 last night, and first run it did 28 minutes of Prime95 multicore without error or locking up (I stopped it); however, the second time I tried it locked up with this BSOD. Changing the voltage up or down didn't make a difference; I'd still get this BSOD at about 3 minutes into Prime95.

What does this mean in terms of stability? Is there anything I can do to help the situation/make it stable?

I'm on an Abit IP35 Pro, clock was 423*9 to get 3.8 (it was 1.5 volts before vdroop average; i tried 2 settings up and down to no avail; 1.5 was what got the 28 minutes), currently running at 400*9 stable (1.375 volts), RAM always at 1:1 with very loose timings for the clock. My temps were never above 60 degrees at the highest voltage I used (they hit 59 and quickly to 60 but didn't stay there).

Any advice would be appreciated!
Click to expand...



Yes,it means your pushing it a little to much,back it down a notch.
 
Transistors literally require that a tiny tiny tiny spark jump a gap when they're switched on. The higher your clock rate the more often they switch. The more often they switch the more power required. If there's not enough power they don't switch properly and you get computational errors. Sometimes they're minor, like the sort Orthos will stop on if you have a semi-stable OC. Major errors cause blue--screens and sudden reboots.

The above is greatly oversimplified but that's pretty much what's going on.

Of course, the issue is that you can overpower the chip and damage it, and the more power you run the hotter things get, so it's a balancing act.
 
You must log in or register to reply here.
Back
Top