• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Pascal series at > 1.2V vcore

Joined
Dec 12, 2020
Messages
1,755 (1.05/day)
If there had been some way to increase the vcore in Pascal series GPU's would this have allowed for better overclocking margins? It took 1.2V to get my 1080ti to 2164 Mhz. I'm just curious if more vcore would've allowed a significantly higher overclock.
 
probably yes but I'm not sure because I don't have a 1080ti to try...
 
From the little I've seen, GTX1080FE to 1.2V T4 VBIOS and GTX1050Ti from 1.118V to a little over 1.3V, the increase in voltage seems to result in a relatively small gain.
 
I too, doing tests, exceeded 1.1v, now I don't remember if I had 3 or 4 more voltage points... but my gpu is not an overclocking card and it has ridiculous cooling. if the 1080 is already capable of doing 1.2v maybe it can go even further.
unfortunately these are guesses because I don't have a 1080.
I will try again later.
 
Last edited by a moderator:
If there had been some way to increase the vcore in Pascal series GPU's would this have allowed for better overclocking margins? It took 1.2V to get my 1080ti to 2164 Mhz. I'm just curious if more vcore would've allowed a significantly higher overclock.
Most designs dont scale that well as voltage passes certain thresholds, and it would have required power circuitry redesigns to support it
 
this crap gpu has 0.975v as default,
1,1310v is already beyond what my cooling system can handle, so useless.

1.1310v.jpg
 
If there had been some way to increase the vcore in Pascal series GPU's would this have allowed for better overclocking margins?
ASUS LN2 Bios has no thermal trip point and no voltage limitation *
Flash under own responsibility.
Its what i run
Stock Curve
1693128885758.png


Boosts stock to 2000MHz (mem undervolted due to lackluster cooling on VRM)
* it's what i have been told - but my diy cooler also can't handle it.
Goal is to turn it into a more safe version, soo keep an eye on the temps with this one.
LN2 bios should have no clock stretching.

EDIT:
You can need custom msi afterburner entry for the voltage unlock.
 

Attachments

Last edited by a moderator:
I have only tested mine up to 2200 and 1.2V (palit card, xoc bios on water), just wanted to mention that at exactly 500W gpu load, my system hard resets. So there's probably some safety measure implemented :sleep:
 
I have only tested mine up to 2200 and 1.2V (palit card, xoc bios on water), just wanted to mention that at exactly 500W gpu load, my system hard resets. So there's probably some safety measure implemented :sleep:
That's pretty incredible, I think I only saw one 1080ti that could hit 2190 on Hwbot.

My MSI 1080ti Duke, w/liquid metal TIM, thermal tape applied between most of the backplate and PCB and two 127mm fans that put out 200CFM could hit 2152Mhz. in Metro Exodus with the memory downclocked by 500Mhz. where it drew in HWiNFO64 upwards of 546.8 Watts and according to afterburner: 468 watts. I used to get system resets as well in Metro Exodus but I eventually figured out that was my system memory o'clock causing that.

@Veii, in afterburner there are higher voltages than 1.2V available in the frequency/voltage curve (at least w/the Asus XOC VBIOS flashed) but selecting anything above 1.2V would result in only 1.2V being seen in HWiNFO64 and afterburner. I suppose that could've been a software issue though.
 
put a nice -100000 (fffe7960) negative offset at the bottom in overclock section... it works:D
Bottom , OC section ?
In what program.

The thing you quoted simply ment,
I got no cooling on them, mem vrm on this pcb is prone to explode
run an underclock, as i don't need that perf.
 
Back
Top