How To Overclock A Video Card: NVIDIA & AMD Graphics Card (GPU) Overclocking Guide

[WLockIV]

So I edited the BIOS with VBE7. For some reason the interface of VBE7 was jacked up (please don't make me explain) on the machine in question, so I transfered the BIOS to my laptop via USB and edited with VBE7 on there. I was only able to select the desired voltage from a drop-down menu, but it's been bumped from 1.18 to 1.2. Successfully flashed it (whew) using atiwinflash. 

 

After restarting the machine in question, I've upped the clock from 1080 mhz (stable @ stock voltage) to 1100 (tested unstable at core voltage), and began the OCCT test. It's about 10 minutes in with no errors at this point. So far, an unstable clock would have thrown a ton of errors by now, but we'll let it finish at 30 minutes. Will report back afterwards. 

 

I used the following link for reference: https://www.reddit.com/r/ScryptMiningRigs/comments/1xi194/bios_flash_guide_r9_270_and_270x/

 

Though it is referring to UNDERclocking, it's still a similar concept. Except I used GPUZ or whatever to grab the BIOS.

 

THANK YOU FOR ALL YOUR HELP! 

[WLockIV]

After all that, I was only able to add another 30 mhz to 1110 mhz. I don't plan on bumping the memory (yet). Seems like I went thru an awful lot for only a little bit more, but oh well. Not to shabby considering I got this card used for $100!

 

Thank you for your help and support. If you or anyone else needs help with solving the issues I outlined above feel free to give me a shout.

[Constantin]

Better spit on the floor than OC any kind of GPU!!!!

[CC268]

Can Valley Benchmark be used instead of the two you recommended in your thread?

[BigDay]

Can Valley Benchmark be used instead of the two you recommended in your thread?

 

i wouldn't recommend it

[CC268]

Alright - in your thread you mention using the EVGA software for higher end GPUs...I have a 980TI - should I give the EVGA software a shot over the OCCT?

[BigDay]

Alright - in your thread you mention using the EVGA software for higher end GPUs...I have a 980TI - should I give the EVGA software a shot over the OCCT?

 

i would experiment with both. sometimes, evga precision x is able to stress higher end gpu's harder, which will help give you a better understanding of stability in game etc.

 

try using occt first. if your gpu usage stays around 100%, which is what you want, you should be fine. if it goes under 90%, then you have a problem. it could be the fact that adaptive mode is on in the nvidia control settings or that occt just can't stress the card enough. or, it could be something else.

 

let me know what your gpu usage is when you follow the guide and stress with occt

[CC268]

Thanks for the info! I think I got a pretty conservative overclock setup right now so I probably won't do any further testing - if I do I will try OCCT and let you know.

[CC268]

Oh something I noticed in your guide - you say to check the "Force constant voltage" before you do any of the testing, but don't you disable this after testing? I didn't see that in the guide.

[BigDay]

Oh something I noticed in your guide - you say to check the "Force constant voltage" before you do any of the testing, but don't you disable this after testing? I didn't see that in the guide.

 

yes, i wrote it somewhere in the guide that after you finish testing, you should disable "force constant voltage".

[CC268]

I have a few questions for you if you don't mind.

 

So my current stable OC in MSI AB is:

 

Core Clock: +100 (1454 MHz max in Unigine Heaven)

Memory Clock: +500 (4005 MHz max in Unigine Heaven)

Voltage: Stock (1.19V)

 

So...I am thinking I should up the voltage 10mV to see if I can get to 1500MHz core clock, but I don't want to put so much wear on my card that it is dead in a few years. I know in your guide it says 1.212V is your recommended limit. Should I up my voltage a bit?

 

Someone also told me that my stock BIOS will only allow 1.187/1.205/1.230 - although I don't know how to verify this.

[BigDay]

I have a few questions for you if you don't mind.

 

So my current stable OC in MSI AB is:

 

Core Clock: +100 (1454 MHz max in Unigine Heaven)

Memory Clock: +500 (4005 MHz max in Unigine Heaven)

Voltage: Stock (1.19V)

 

So...I am thinking I should up the voltage 10mV to see if I can get to 1500MHz core clock, but I don't want to put so much wear on my card that it is dead in a few years. I know in your guide it says 1.212V is your recommended limit. Should I up my voltage a bit?

 

Someone also told me that my stock BIOS will only allow 1.187/1.205/1.230 - although I don't know how to verify this.

 

don't be afraid to push the card to 1.212. you should be fine. which 980 ti model do you have? overclock that sucker.

[CC268]

MSI 980TI (it is in my signature as well). Thanks for your feedback!

[astranger200]

Got 150mhz on my 660 ti

[waggythegeek]

My AB settings page does not have nearly as many options as this one- What did I do wrong?

[BigDay]

My AB settings page does not have nearly as many options as this one- What did I do wrong?

 

what kind of gpu do you have? post a screen shot. do you have the latest version of msi ab? which version of windows are you running? make sure you download msi ab from the official msi page

[waggythegeek]

what kind of gpu do you have? post a screen shot. do you have the latest version of msi ab? which version of windows are you running? make sure you download msi ab from the official msi page

I have a r9 280 windforce. I just told AB to update itself, it said it was up-to-date. I am running windows 10, and I downloaded AB from the link you provided. Here is a screenshot:

[BigDay]

I have a r9 280 windforce. I just told AB to update itself, it said it was up-to-date. I am running windows 10, and I downloaded AB from the link you provided. Here is a screenshot:

MSI broken.PNG

 

what seems to be the problem? it looks like you have all of the options that i have. are you referring to options like core voltage, memory voltage, power limit, etc?

 

try enabling different user interfaces by going to:

 

settings > user interface > user interface skinning properties > default msi afterburner v3 skin - big edition

 

if you've tried using all of the skins and you still haven't been able to adjust your core voltage, your gpu may be voltage locked

[waggythegeek]

what seems to be the problem? it looks like you have all of the options that i have. are you referring to options like core voltage, memory voltage, power limit, etc?

 

try enabling different user interfaces by going to:

 

settings > user interface > user interface skinning properties > default msi afterburner v3 skin - big edition

 

if you've tried using all of the skins and you still haven't been able to adjust your core voltage, your gpu may be voltage locked

I mean with all the option tabs at the top- like in settings, for the on-screen display.

[BigDay]

I mean with all the option tabs at the top- like in settings, for the on-screen display.

 

that's because you DON'T have riva tuner statistics installed along with MSI AB. uninstall MSI AB completely and then reinstall along with riva

[waggythegeek]

that's because you DON'T have riva tuner statistics installed along with MSI AB. uninstall MSI AB completely and then reinstall along with riva

heeey thx. I thought that was spam.

[waggythegeek]

Okay, so i sorted out my AB problems and began experimenting with core clock speeds today. One MHz over or one MHz UNDER the clock speed it came at results in huge temperature spikes- the test stops itself within ten seconds due to overheating. I set the temp max to 85, and let it cool down into the high 60's before each test. what am I doing wrong?

[BigDay]

Okay, so i sorted out my AB problems and began experimenting with core clock speeds today. One MHz over or one MHz UNDER the clock speed it came at results in huge temperature spikes- the test stops itself within ten seconds due to overheating. I set the temp max to 85, and let it cool down into the high 60's before each test. what am I doing wrong?

 

so, what you're saying is that when you set your core clock exactly 1 mhz higher or lower than the default core clock your gpu temperature increases by a large amount? how much exactly? are you referring to the gpu temperature or the vrm temperature? your temperature increases regardless of a positive or negative change to the core clock because you're stress testing the gpu; that's normal

 

stress testing increases the overall temperature of the video card. oftentimes, it will increase very quickly. are your fans running?

 

which test stops itself within ten seconds due to overheating?

 

what kind of video card do you have? have you checked to see if it's voltage locked?

 

the more info you give me the better...

[waggythegeek]

so, what you're saying is that when you set your core clock exactly 1 mhz higher or lower than the default core clock your gpu temperature increases by a large amount? how much exactly? are you referring to the gpu temperature or the vrm temperature? your temperature increases regardless of a positive or negative change to the core clock because you're stress testing the gpu; that's normal

 

stress testing increases the overall temperature of the video card. oftentimes, it will increase very quickly. are your fans running?

 

which test stops itself within ten seconds due to overheating?

 

what kind of video card do you have? have you checked to see if it's voltage locked?

 

the more info you give me the better...

GPU, not vrm. fans are running at max. any change at all, up or down, any amount, in clock speed, results in GPU temps going from high 60's to 85 in about fifteen seconds once a benchmarkis started.. Both OCCT and the EVGA one stop. I have an r9 280 gigabyte windforce, which I can't change the voltage on.

[BigDay]

GPU, not vrm. fans are running at max. any change at all, up or down, any amount, in clock speed, results in GPU temps going from high 60's to 85 in about fifteen seconds once a benchmarkis started.. Both OCCT and the EVGA one stop. I have an r9 280 gigabyte windforce, which I can't change the voltage on.

 

i think it's normal for your temperatures to increase like that especially since your running a stress test that forces your gpu to go under 100% load. the r9 280 gigabyte windforce edition is most likely voltage locked. your overclock will be limited as you won't be able to increase the core voltage

 

apparently, the r9 200 series cards run very hot by nature. they can supposedly run up to a safe max operating temperature of 95 degrees according to the article below:

 

http://www.pcgamer.com/amd-radeon-r9-290x-review/

 

as long as your temps stay below 95 degrees or even 90 degrees, i would continue pushing an overclock and trying to find the optimal settings for your card