Sign in with
Sign up | Sign in
Your question

GTX 570 Overclocking for people who KNOW WHAT THEY'RE TALKNIG ABOUT!!

Last response: in Graphics & Displays
Share
January 23, 2012 3:30:40 AM

Hey all, I have a MSI GTX 570 Twin Frozr III, and I have a few questions about overclocking. Yes I know, don't overvolt, don't undervolt, etc. And to all of you, before you post- this card is non reference and has 6+1 Volting not 4+1 or whatever.
Okay so the real question: I was playing Battlefield 3 the other day. I was getting a little lag on a HUGE server with a lot going on, so I bumped it up to 900 core and 1812 shader, and put Vcore to +126 or so. It improved gameplay A LOT! But later I fired up Furmark and did some tests, and Furmark gets better FPS at stock than at my OC! Can anyone explain or elaborate on this? Thanks, -AW
(PS, I didn't mean to sound like a jerk at the beginning, I was just telling you what I already know, apologies if it came off wrong)

More about : gtx 570 overclocking people talknig

a b K Overclocking
January 23, 2012 5:43:29 AM

Review here: http://benchmarkreviews.com/index.php?option=com_conten...
Apparently that card just doesn't have much overclocking headroom, per the review.
I don't think of Furmark as a benchtest for FPS, more of a stress test. Try something like Unigine Heaven 2.5 to make a direct comparison between stock and clocked.
m
0
l
a b Î Nvidia
January 23, 2012 10:32:14 AM

Well, wasn't the furmark a burn-in test not an fps test? I think you are not suppose to say this as a strange thing, considering that you know furmark is a burn in test instead of a FPS benchmark. So it is I think pretty obvious to see when you have a lower fps on furmark when overclocked. If you have your games running much faster as you said when overclocked, I think it's just the furmark. Try running benchmark like 3dMark11, 3dMark Vantage, Unigine heaven, etc. You should be able to see a performance increase when you overclock your card then you compare it to stock speed. Good luck :) .
m
0
l
Related resources
January 23, 2012 10:55:24 AM

refillable said:
Well, wasn't the furmark a burn-in test not an fps test? I think you are not suppose to say this as a strange thing, considering that you know furmark is a burn in test instead of a FPS benchmark. So it is I think pretty obvious to see when you have a lower fps on furmark when overclocked. If you have your games running much faster as you said when overclocked, I think it's just the furmark. Try running benchmark like 3dMark11, 3dMark Vantage, Unigine heaven, etc. You should be able to see a performance increase when you overclock your card then you compare it to stock speed. Good luck :) .


+1.

And one more thing. Furmark and Kombustor may damage your gpu, so avoid using them if you care about your gpu:


source: http://www.geeks3d.com/20091209/geforce-gtx-275-vrm-dam...
m
0
l
a c 87 Î Nvidia
a c 166 K Overclocking
January 23, 2012 10:58:52 AM

I'm not sure why the benchmarkreviews author had so much trouble. The base card comes in at only 850 Mhz just 30Mhz over the reference design speed. Most other factory PC'd cards come in at 900MHz or better outta the box.

The BMR folks only managed to OC to 900 MHz, again right where most of the competitors cards start with outta the box. I have seen 560 TI's reach much higher OC's ...... for example:

1070 MHz http://www.pureoverclock.com/review.php?id=1201&page=17
1017 Mhz http://www.overclockersclub.com/reviews/nvidia_asus_gtx...

And my son was running an SLI rig with two cards OC'd to 1020 that we just clocked back to 980 MHz to keep fan speeds down at the inaudible level.

I'd suggest using Afterburner 2.20 Beta 11
http://www.guru3d.com/news/msi-afterburner-220-beta-11-...

I was able to get almost all (11 outta 12)of the factory OC'd cards in builds I have done to 1000MHz w/o a voltage bump ... the last needed just 1 notch but again, all these were Asus cards.

The 6+1 you are referring to defines how many phases the Voltage Regulation Module has been provided with ..... more phases, the better.
m
0
l
a b Î Nvidia
January 23, 2012 11:03:13 AM

It damages your card very rarely, so it's a FUD (Fear, Uncertainty and Doubt). But you should not break your graphic card when you have a proper cooler and enough PSU. I ran Kombuster on my Laptop and there is no such issue, the creator of furmark is just warning you for a 1 out of 1000 cases I think.

But still that might happen, but really rarely and mostly happens when you play volting. I don't know if this is 100% correct but I think it is. Good luck OP :) .
m
0
l
January 23, 2012 2:32:42 PM

Furmark isn't really a benchmark. Try unigine with everything cranked up...that will torture the GPU and give you a realistic set of data between the stock and OC setups.
m
0
l
a b K Overclocking
January 23, 2012 4:11:23 PM


I may be wrong, but seem to remember when the fermi cards were first released, that reviewers had to dump furmark as the new drivers would downclock when furmark was detected in order to not overheat or something to that effect.
m
0
l
January 23, 2012 7:14:00 PM

If you wanted help you SHOULD HAVE JUST USED ALL CAPS IN THE TITLE OH WAIT YOU DID OMG :p 
m
0
l
January 23, 2012 9:48:27 PM

DelroyMonjo said:
Review here: http://benchmarkreviews.com/index.php?option=com_conten...
Apparently that card just doesn't have much overclocking headroom, per the review.
I don't think of Furmark as a benchtest for FPS, more of a stress test. Try something like Unigine Heaven 2.5 to make a direct comparison between stock and clocked.


That link led to me a Radeon card... It was TFIII but not GTX570
m
0
l
January 23, 2012 9:49:48 PM

Thanks everyone for clarifying what Furmark was for I didn't know that a stress test would do worse on an OC (which makes total sense now that I think about it lol), I'll continue to try and find that sweet spot.
m
0
l
!