• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Stuck on power limit 125% on GTX 780

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ntnlloverclock

Registered
Joined
Feb 13, 2012
Hi to all,
i'm getting crazy trying pass through 125% wall in AB power limit overclocking a EVGA GeForce GTX 780 sc ACX. Without achieve this, I cant push up the overclock as I wish.
I'm on liquid cooling, and on full load temps are quite low at 54°c, but if I push up the power limit over 125%, the system crash with a black screen and I must restart with an hardware reset. Seem to be related with some energy issue, because it is not the classic driver crash behavior. I red somewhere maybe due to psu spikes, is this correct? There is any way to solve? The PSU is a Cooler Master 850 Silent Pro GOLD.

Currently i'm using the skyn3t-3A-ACX-rev3a bios, and so far, i'm current stable on +440 (core clock) and +380 (memory clock), using msi after burner beta19.

I cant find any solution searching the forums, any advice?

Many many thanks
Anto
 
I don't think you need more than 125% for these clocks. It's more like a voltage/temp limit than power limit issue. Every card is scalling in some other way so maybe you've reached that step when voltage has to be higher or temps have to be lower.
Even at 50*C sometimes it's still too hot for much higher voltage/clock.
I was benching my GTX780 up to 1600MHz core with ~130% power limit and voltage was above 1.5V.
 
Sorry, I forgot to tell that i'm stable on 1.325v, though, I cannot quiete understand if my voltage read is correct: in Msi Kombustion and GPUZ i see 1.15v, but on HWInfo64 and Msi Afterburner I see stable 1.325v over full load...
 
I would check voltage on a pins of a coil or capacitor next to voltage regulator. GPU-Z or any other 3rd party soft sees voltage ID under which voltages are registered in BIOS. It's never real voltage. Also voltage under load and idle is different. However if you set voltage in system then you can expect +/- 0.02V accuracy.

So how high is your card's max clock in 3D at these 1.32V ? I forgot about one more thing. There is something like OCP/OVP so generally protection against burning your card due to voltage spikes or overvoltage. Most GTX780 that I saw had it set to about 1.3-1.35V and above were happening weird things without additional mods.
 
Last edited:
Assuming 90% efficiency your computer is using 585W, that should be easy work for your PSU.

Are you using a universal or full-cover block? Link and/or model number?
 
Looking the specs of the PSU I red
80 PLUS Gold certified: up to 93% efficiency @ 50% load
For cooling i'm usinga a fullblock Swiftech Komodo-NV-GTX
 
My thought is that you're either overheating (not likely with the full cover) or hitting OCP on the VRM for the GPU.

Edit: Just realized that your PSU was in the first post the whole time. Whoops!
 
Every card is scalling in some other way so maybe you've reached that step when voltage has to be higher or temps have to be lower.
Even at 50*C sometimes it's still too hot for much higher voltage/clock.
Sure sounds like this^^^^^^
 
I should have a bit of temperature margin raising the radiator fanspeed and the pump speed, I'll make a test later and will report, thanks you all very much for help!!
 
I should have a bit of temperature margin raising the radiator fanspeed and the pump speed, I'll make a test later and will report, thanks you all very much for help!!

Should is always the key word, whether it's a cpu or gpu there will become a point that in order to push X amount of voltage you will need to get it colder even if the temps are below what would be considered too high or even warm. Meaning even at 50c which is not hot at all for a Gpu, yours may not be able to handle more voltage until it's a lot cooler, it's just the nature of the overclocking beast!
 
Increase the fans and pumps seem not change the behaviour.
I tryed to lower fans and pump, and also if the temperature pass the 58°C, the system doesn't crash! The system crash only over the power limit 125%. I can't understand..
 
The system crash only over the power limit 125%.
You can thank Nvidia for making that happen...:p

So, to be clear, you have a flashed bios that the power limit goes over 125%. You have that limit set over 125%, but when it hits 125% actual use (shown in MSI AB or GPUz) you freeze/reboot, whatever, righT?

Yeah, 1.30v is alot for air regardless...especially for a reference model 780... those power bits are probably begging for mercy. I wouldn't go that high with a reference board.
 
Last edited:
You can thank Nvidia for making that happen...:p

So, to be clear, you have a flashed bios that the power limit goes over 125%. You have that limit set over 125%, but when it hits 125% actual use (shown in MSI AB or GPUz) you freeze/reboot, whatever, righT?

Exactly.
Now the GPU is working at about 1306-1345 Mhz (oscillating between these values). For get to an higher frequency, I must increase the power limit, but whenever I pass the 125%, the system crash. It's strange, because if I slide the power limit to the max, on 150%, and I click to apply, the system crash immediately (under load)
 
Last edited:
You may want to try a reflash or perhaps a different bios...

But, really, you should not be at 1.3v on a reference 780 in the first place... Hell on the Lightning I reviewed they said 1.35+ = almost immediate death on air/water, and that power section is one of the most impressive and robust around...

I think its time you figured out what the clocks are at 1.25v and enjoy your card... that is, unless you are benchmarking. :)

the system crash immediately (under load)
VRMs can't hack it...
 
Last edited:
Hi guys, I did some test tonight....definitively was the voltage, but not the tempearature:
I tryed to decrease voltate down to 1.288v and now, though I can't push up the core clock like before, FPS increases a bit and seem to be more stable (I can see the stable values in gpuz without any type of throttle) and the system does not crash, even with power target to 150%....my gpu seem don't like higher voltage :)
Thanks for your help guys!

edit: was not the temperatures because now i'm on 59°C and i have not any crash
 
Last edited:
When you hit OCP/OVP limit then 1st stage on these new cards is throttling or driver crash, when you push it too high then is black screen, freeze or system restart. ASUS DC2 cards have OVP at about 1.3-1.32V. I believe that reference cards have it slightly higher but there is no software mod for that. EVGA Classified or MSI Lightning can be set higher without additional mods but you pay extra for them. I doubt that anyone needs to run it 24/7 at 1200+ clock , not to mention 1400+.
I have my card set to 1.1V 1100MHz for gaming and it's more than enough for 1080p.
 
When you hit OCP/OVP limit then 1st stage on these new cards is throttling or driver crash, when you push it too high then is black screen, freeze or system restart. ASUS DC2 cards have OVP at about 1.3-1.32V. I believe that reference cards have it slightly higher but there is no software mod for that. EVGA Classified or MSI Lightning can be set higher without additional mods but you pay extra for them. I doubt that anyone needs to run it 24/7 at 1200+ clock , not to mention 1400+.
I have my card set to 1.1V 1100MHz for gaming and it's more than enough for 1080p.

You're right, but i'm fighting against flight simulators beasts, like x-plane and prepar3d 2, and with my rig the sim work at 25-30fps, any extra frames does the difference! :D
But i'm pretty satisfied of my overlclock, now. Thanks you all
 
Back