Kernal power 41 after OC

lambomb

Reputable
Nov 14, 2015
22
0
4,510
Hello all I over locked my gtx 680. And I've ran heaven and 3D and nothing crashed no artifacts or anything. Well after about hour or more playing fallout 4 ran good then it froze my screen and had to hard restart my computer. I go to event viewer and looks like kernel power 41. I just check my power cord re seeted it. And changed my outlet. Havnt tried playing yet. But my temps where fine it was about 64c under max load. I'm running memtest right now. Also this happened when I do the performance setting on saber tooth z77 motherboard. I did manual OC and it just happened again. Any suggestion would be so greatly appreciated just trying to get it all set up correctly.
 
Solution
Well, I've never seen a k-41 error caused by a gpu. It's a cpu error if anything. Lack of voltage usually to the cpu. This can be caused by several things although a unstable OC is the most common. An underpowered psu is also common, where the psu doesn't actually output what it claims by either age or cheap psu, then you slap on a high power draw such as an OC on a gpu and game with it, drawing power on cpu and gpu (doesn't affect stress testing as that's mostly testing mainly cpu or gpu).

99% of the time it is directly tied to power loss. This doesn't necessarily mean the psu. It could also be a bad power strip/surge protector, or even bad outlet in the wall somewhere in the line, even back to the house electrical panel. A psu can...

Karadjgne

Titan
Ambassador
That error is usually indicative of not enough vcore voltage. Since you've done a manual OC, I'm assuming you lowered vcore as much as possible to maintain low cpu temps. Bump it up a notch. Just means your OC isn't as stable as you thought it was.
 

lambomb

Reputable
Nov 14, 2015
22
0
4,510
I'm using evga precision. They have the power target wich I chose 132% and they have voltage wich mine is 1000 think stock is 975 or something so I didn't raise it much. Are you talking about the voltage that I should raise? Thanks so much for your time!
 

Karadjgne

Titan
Ambassador
Well, I've never seen a k-41 error caused by a gpu. It's a cpu error if anything. Lack of voltage usually to the cpu. This can be caused by several things although a unstable OC is the most common. An underpowered psu is also common, where the psu doesn't actually output what it claims by either age or cheap psu, then you slap on a high power draw such as an OC on a gpu and game with it, drawing power on cpu and gpu (doesn't affect stress testing as that's mostly testing mainly cpu or gpu).

99% of the time it is directly tied to power loss. This doesn't necessarily mean the psu. It could also be a bad power strip/surge protector, or even bad outlet in the wall somewhere in the line, even back to the house electrical panel. A psu can only change AC-DC with what it has, so if you are supplying limited voltage, you are going to get limited output, especially on a cheaper psu that doesn't have good, active power correction. Add to that a high draw cpu/gpu under gaming loads, and something's gonna give.
 
Solution

lambomb

Reputable
Nov 14, 2015
22
0
4,510
Thank you for the reply, so I just OC my CPU. And I upped my voltage up and few other things. So far I played about an hour, nothing so far. But I did realize my CPU gets about 80 wich I realize is hot when running memtest 100% load but if i play a game such a fallout it sits about 58-70 general area. So I figured I wouldn't be pushing my CPU that hard in games. I've been trying to work on OC GPU CPU for like 4 days and I'm starting to go crazy with it all. My PSU is hx750w corsair.
 

Karadjgne

Titan
Ambassador
Nice psu. Try a program called MSI Kombuster. It's good for gpu testing and has online test results so you can compare to others. It's not as brutal as firestrike but it does run a variety of different tests.

Pick one to OC first, doing both together would drive anyone nuts and you can't be sure if errors or instability are due to one or the other.
 

lambomb

Reputable
Nov 14, 2015
22
0
4,510
I'll give that a try. Yeah I started with my GPU clock then moved on to CPU. But doing them both within this week has me going crazy trying to figure it out, it's my first OC experience.
 

Karadjgne

Titan
Ambassador
Heh, got the bug and forgot (or didn't know) the one and the only absolute rule of OC. Baby Steps. It can easily take weeks to dial in an OC, and thats including a 3 second change and an hour of testing, where you actually just go watch TV. Over night memtest, 12 hr repeats of different tests. It's how it's done, there is no easy answer, no quick switch. Which is why those motherboard software oc's aren't worth the cd they come on.

Start from scratch, get 1 stable, wait a week to make sure you were successful, then move to the next. You'll own this pc for a while, there is no rush, research every setting, every switch, every name.

OC is a hobby at best, treat it as such and the end result is a lot more satisfying than 'OC Now'
 

Karadjgne

Titan
Ambassador
I oc'd my Asus 660ti. Currently is pulling numbers greater than most post for a Gigabyte 960 G1. That's not too shabby.

Bumped clocks as high as I could, 1 at a time. Added a little voltage, played with power target, kept bumping clocks. No worries on anything except power target for some odd reason. 0-9% all good. 10% crashed, even at lowered clocks. Same with 11,12,13%. 14% that gpu settled down and started really moving. 15%-25% crashes. Weirdest thing I ever saw in an OC, the span of crashes like that and one little island paradise, right in the middle.

Stock boost/memory 1058/6008
Oc boost/memory. 1190/6830 @1037mv 114% PT

Took 2 weeks of guessing with that stupid power target, but I'm happy with that OC, and it still runs cool and quiet.