Overclocking help please!

zyntaxable · August 4, 2015

Hi,

I'm very new to overclocking so please go easy on me

I purchased a Nepton 240M and a GTX 970 the other day, as an upgrade for my existing computer. My cpu is an FX-6300.

I bought the cooler so I could overclock the cpu to reduce bottleneck and get as much out of it as possible before and upgrade is absolutely neccesary. So I installed everything and it worked fine; Great, I thought. Let's get to overclocking! (At this point the GTX 970 isn't installed, because the delivery time took longer than the watercooler, so all these tests have been performed with an R9 270X DCUII)

So in the beginning everything went alright, I followed a guide here on the forums and everything seemed to go well. I started out with 1.45V and 4.5GHz. That ran fine so I dropped the voltage down to 1.375V, that also went well. But I noticed a weird thing, while stress testing with Aida64 and monitoring with CPU-Z, the voltage was lower than that I had set. (It got to a max of 1.288) I don't know if that affects performance at all, but the stress test went fine so I thought nothing of it. I increased the clock with and additional 100MHz, and keeping the voltage at the same level for the time being. Again, the voltage in CPU-Z was lower. This time it was max 1.296. Again, after 5 mins of testing everything still seemed okay, so I went ahead and added 100MHz more, again keeping the voltage set to the same.

Here I ran into my first problem, after 10 seconds of torture testing with aida there was a fatal error and the test was stopped. I didn't find this strange because I still had the same voltage I had with 200MHz lower. So I added 0.125V. It booted up fine, and again the 5 minutes of testing went flawlessly. This time there was a max voltage of 1.304 in CPU-Z.

I found that last very strange, because it could have reached that amount of voltage even when I had set it to a max 1.375. But since the test went fine I decided to try 4.8GHz, with the same 1.400V. This test also failed, but at this point I was getting really confused, and kinda scared that I might break something, so I reverted the core clock back to 4.7GHz and decided to test it with prime95. So now the clock was set at 4.7GHz, with a voltage of 1.400. As soon as I started the prime95 stress test (with small FFT's) core #2 failed. It said something like: rounding was 0.5, expected less than 0.4 Strange, I thought, since the aida64 test went perfectly fine without even reaching 1.4V. I decided this was enough, because I didn't want to increase the voltage any further since I don't want to harm the CPU.

I then tried to run another aida64 torture test, and again, it went fine for as long as I ran it, without ever getting higher than around 1.304V. I thought maybe the error had fixed itself, so I tried running prime95 again. It gave me the same error as before, but instead of 0.5 it was 0.492 (and alot of numbers after that).

So I am wondering, what am I doing wrong? My processor has not reached higher than 1.312V, according to CPU-z, but my core #2 keeps getting errors in prime95.

I'm also wondering why just one of my 6 cores gets an error, 1, 3, 4, 5 and 6 seem to work without any problems at all.

Here is my complete build:

CPU: FX-6300, cooled by Nepton 240m

mobo: ASRock 990FX Extreme3

GPU: Asus R9 270x (no OC)

RAM: 2x4GB G.Skill ValueRAM, 1600MHz (I think it is running at 800MHz, because that is what CPU-Z says in the screenshot but I have put it at DDR3-1600 in the BIOS)

PSU: Sharkoon WPM500, 80+ Bronze

And some random hard drives.

Any and all help will be GREATLY appreciated, as I'm at a complete mystery here.

- zyntaxable

ApolloX75 · August 4, 2015

@zyntaxable

That's a mighty strange one you've got. Have you disabled Cool n Quiet in the BIOS? That is often a culprit for lower than expected voltage.

And you're sure the voltage settings are on manual not adaptive?

Not all modules are created equal, so several may run fine on a lower voltage than the others. That is to be expected. But let's focus on getting your voltages straightened out before troubleshooting something that is most likely just a symptom.

zyntaxable · August 5, 2015

@ApolloX75 I have disabled all of those things, al lthe voltage settings are on manual, but it still does it.. Any other tips perhaps?

ApolloX75 · August 5, 2015

@ApolloX75 I have disabled all of those things, al lthe voltage settings are on manual, but it still does it.. Any other tips perhaps?

Hmmm. Going through your manual to see what I can find. Try disabling C6 State and CPU thermal throttle under CPU Config in Advanced. The C6 State is directly related to CPU power consumption so that may be it right there.

You can also set the CPU VDDA Voltage to about 2.8v which acts the same as the VCCIN setting for Intel. It keeps the core voltage (which runs at 1.4, 1.45, etc whatever you set it at under maximum CPU voltage) From flagging under load. It's a basic kind of load line calibration.

Give that a try, it's all I can really find in the manual for you.

zyntaxable · August 5, 2015

I was trying some things out before you posted your latest comment, and while trying those things, i got BSOD with a certain error code. I tried everything to fix it buy my pc landed in a loop of giving error and restarting. So I think somehow my windows is corrupt. I managed to find an old hdd with windows 8.1 on it, so I booted from there. i reset the bios and am going to reinstall windows on my main ssd

Thanks for the help tho, really appreciated!