What are your ambient temps? That'll decide if you really need an after market solution & whether you can crank the card further up.By the way, I did an hour of BF3 at 1050/1300 and 1.25V
Max Temps
Core: 71C
VRM1: 70C
VRM2: 74C
Is that something to worry about enough that it's worth buying those heatsinks and disassembling the card for?
What are your ambient temps? That'll decide if you really need an after market solution & whether you can crank the card further up.
Those temps are perfectly fine @ 1050 if that is the max core temp during actual game play and vrms temps are fine too.
That card does not have vrm cooling. It sounds like in your previuos post you think you have to replace (heatsinks) there is no heatsink cooling so nothing to replace. By having no vrm cooling it will limit the amount of oc'ing clocking you do to a degree due to vrm temps. At the oc your and your posted game play temps your fine not even hot if you stay at that oc so don't worry.
Just don't use occt to test just use UH 4.0 Extreme and play games for monitoring temps if no problem in either of these you are fine.
Yeah, but I'd have to remove the GPU heatsink to install those. That's what I meant.
If your VRMs are staying under 90C they are fine. VRMs are generally designed to be fine up to 120C +/-
If you are paranoid, you can add a small undervolted fan blowing at them to drop temps.
If you go liquid cooling, you will likely need to add thermal pad + heatsinks and active cool the VRMs to get acceptable temps.
Example from my OC trials with a Powercolor 7950 reference PCB:
http://forums.anandtech.com/showthread.php?t=2317696&p=35648205
So what is the voltage and temps you are looking at here?
Maybe could try that, just increase the voltage to like 1.25, and crank up the fan to 100%. Then set the card to 1200 core, leave the memory stock, and run Unigine Heaven or Valley while you keep an eye on the temps and looking for artifacts - let it loop maybe twice too see if the temperature levels off.
When you installed the card did you remove all old drivers and run a driver sweeper?
By the way, I did an hour of BF3 at 1050/1300 and 1.25V
Max Temps
Core: 71C
VRM1: 70C
VRM2: 74C
Is that something to worry about enough that it's worth buying those heatsinks and disassembling the card for?
Because there isn't one program that can accurately test a "fully" stable OC. The best thing I could find was playing Crysis 3 for an hour or more. You could be stable with 3DMark, Heaven, BF3 and more with 1050Mhz -- the second you load up tomb raider you could crash from an unstable OC. It's just the nature of the beast as of right now.