PRIME95 error after oveclock

deeper02

Honorable
Aug 13, 2012
15
0
10,510
Hi,

I'm quite new to overclocking but after I've read about it I changed my GHZ from 3.1 to 3.5 and then to 3.7 and now to 4.0.

After 20 minutes of prime95 it was stable on 3.5 and 3.7 as well.
But after I changed it to 4.0 today, I get this:
"ROUNDING WAS 0.5, EXPECTED LESS THAN 0.4" And when I get this error
prime stops testing core 8.
I do have a good cooler and the heat don't go above 45c while testing.
I also disabled everything that Is required to overclock and to have it stablisied,
I didn't change my voltage so it's on stock voltage which is 1.35v

My specs:

FX8120
HD7850 2GB
8GB RAM
AsRock 970 extreme3

I really need help because I'm really worried about this.
Thanks!
 
Solution
it means your oc is unstable, try tuning up your voltage or lowering your oc, and try oc use manual voltage .. everything Chip will different get voltage ... get as low posible, if get BSOD up voltage little untill you can use it, min 1hours with no problem / get BSOD again after that ran prime95 again .. just a suggestion

Buzz247

Honorable
Mar 18, 2013
962
1
11,360
i might try an offset of ratio to DRAM. But a couple questions - what size PSU are you running? Air cooled? (Do not OC over 4Ghz if aircooled) What multiplier are you using? on stock voltage, you SHOULD be able to reach or be near 4Ghz by adjusting multiplier. I'd try 200x20 1.37v. that board/CPU is better off OC the multiplier than BUS. Use BUS and vcore to stabilize but not ramp - in other words VERY small changes on those 2

Don't be afraid to actually DROP voltage to reach stable either - up is not always best. IIRC 1.25-1.4v is considered safe range for that chip

2 Things to keep in mind on that board - 1)Temp sensor is skewed cool due to sensor being embedded in the board - allow up to 15C variant 2) Phase power design - shut down power control and set BIOS at high performance

Also make changes thru BIOS - not software if you are not already
 

MEMOFLEX

Distinguished
What Buzz247 said is very true in regards to voltage and mor enot always being best although it depends on the chip in question. Only reason I am mentioning this is I have been running a 4GHZ OC on an i7 930 for nearly 3 years at 1.35v which was all I could get stable at the time.

I recently upgraded my bios in regards to an SSD issue so had to manually do my OC again and I managed to get much lower voltage settings across the board. This may have been linked to the bios update but it also could be directly tied to the fact that I spent a considerable amountof time slowly building my OC rather than going all out for 4GHZ. I am currently running 1.29V on the CPU and temps are so much better both idle and OC.

I can actually get to 4.2GHZ with a little more juice but imo the benefit is not worth it for the heat and volts required.

I will admit that my real experience lies with Intel based systems but the logic still applies in terms of OC. What I would try and do is explore some OC tutorials and see how best you could dial it back a fraction at a time, test again and go from there. I know you have already done some of this but it can reap rewards.

Again all of this is dependent upon having a chip that is willing to work at these settings but it is defo worth the effort for the longevity of your chip