Overvolting titan x through nvidiamaxwell editor

Hello forum i am trying to have a 1.3v unlocked on my titan x. right now i am maxing out at 1.23 with +112mv on evga precision x. afterburner will not allow me to control the voltage whatsoever and on evga 112 is the max . i went to the bios editor and put 1.3v but nothing has changed
what could i do

Mikesamo said:

Hello forum i am trying to have a 1.3v unlocked on my titan x. right now i am maxing out at 1.23 with +112mv on evga precision x. afterburner will not allow me to control the voltage whatsoever and on evga 112 is the max . i went to the bios editor and put 1.3v but nothing has changed
what could i do

Click to expand...

the max voltage for the titan X is 1.274v as measured by Vdimm, which is reached by setting the voltage to 1.281 in bios or higher. There is no bios that will reach 1.3v.

You have 2 choices to change volts to what you want
1) locked volts which will boot up at whatever you lock them at and stay there, no software necessary. Disadvantage you cant change the voltage unless reflash bios. To do this I changed P00 and P02 volts to 1.218 min and max (P00 is necessary). then I changed my boot table, so I am always at 1400 core with 1.218v, I use that for 24/7 bios. No software needed, and on any load im at 1400 core/1.218v, and downclocks like normal on idle. You also need to raise power limits to avoid throttling.

2) you can have adjustable volts from stock up to 1.274v like you want, by increasing P00 minimum voltage to 1.156-1.16. You can leave P02 volts at stock or increase that min as well, doesnt seem to matter. But I increase it if increasing boost table, if dont mess with boost table then no need to increase P02 (if increase boost table, then P02 speed increases and I increase P02 volts for stability). And increase power table, I increase mine so doesnt throttle even without use of slider, up to 360W. max is just less than 400W (hard limits of card, it will throttle just before 400W regardless of bios settings). the only disadvantage is this requires software to use, when you first boot up the voltage will be 1.274v. Once you hit default on precision X or whatever software u use, it will change volts to stock. Also note that as you increase volts, it climbs the boost table. If you have precx for example to start with windows, then it should keep your settings from then on. There is no way to boot up at stock voltage without using software to first set it, unless further editor improvements allow more options. see pic of this bios changes:

I think number 2 is the better choice . more efficient . Ill try and get back to you
here are my settings by the way . im sorry for being nooby but this is my first time doing this .

hey man my clock speed went down signifcantly after increasing the voltage . yes it runs at higher than what i had but the core clock now is 1392 instead of 1437

not sure what you are saying.

But when you hit apply or default for the first time, it will lower the voltage and clock to stocks. Then you increase core and voltage from there using whatever you want.

If you mean you are throttling/decreasing on load, that can be from temps if you are on air, or wattage. btw...if you are on air, you probably wont be able to use volts over 1.25ish, or you will throttle based on temps during load.

your power limits are fine, though you need to use the slider on your bios limits.

rge said:

not sure what you are saying.

But when you hit apply or default for the first time, it will lower the voltage and clock to stocks. Then you increase core and voltage from there using whatever you want.

If you mean you are throttling/decreasing on load, that can be from temps if you are on air, or wattage. btw...if you are on air, you probably wont be able to use volts over 1.25ish, or you will throttle based on temps during load.

your power limits are fine, though you need to use the slider on your bios limits.

Click to expand...

yes i fixed it by increasing the boost limit . it was set lower for some reason . now i am maintaning 1450 mhz .. but theres still a problem . why am i crashing with such high power limits . power limit usage isnt even close to being fully used
in games under full load im around 80-94 % power usage . i have up to 114. im still getting crashes

Increasing power limits only keeps you from throttling (assuming you dont temp throttle which can start around 60C, issue only for air cooling), nothing to do with crashing/instability. Crashing means unstable overclock, driver issues, setting issues, or game itself issues.

And newer games like far cry 4 and GTA V are very taxing on overclocks and require more voltage for stability, just like some programs (like prime 95) are more taxing for cpu overclocks. For example I am any 3d bench stable at 1405 core with stock 1.169v, ie firestrike extreme, valley, heaven, fsu, etc. I am 3d bench stable also at 1480core with 1.21V

However to game with far cry 4 (30 + hours without any crashing) and GTA V over 50+ hours without a single crash, I require 1.21v for 1400 core stable. Both far cry 4 and GTA V will intermittently crash if I run 1400 core at bench stable settings. I need to be about 75mz lower overclock for a given voltage than 3d bench stable to run both of those without crashes. But depends on the game, I can run some other games near benching stable settings. And apparently, per others posts, COD AW is also taxing on overclocks and requires similar lower OC or higher voltage like far cry4/GTAV.

Also, I ran Display driver uninstaller in safe mode, then reinstalled all drivers (newest), changed performance mode to max from adaptive in nvidia panel, and then need to maintain enough voltage for a given overclock, then shouldnt crash at least in those 2 games, assuming you also have a stable OC, and assuming you are not crashing from P02 voltage mismatch with boost states tab settings.

rge said:

Increasing power limits only keeps you from throttling (assuming you dont temp throttle which can start around 60C, issue only for air cooling), nothing to do with crashing/instability. Crashing means unstable overclock, driver issues, setting issues, or game itself issues.

And newer games like far cry 4 and GTA V are very taxing on overclocks and require more voltage for stability, just like some programs (like prime 95) are more taxing for cpu overclocks. For example I am any 3d bench stable at 1405 core with stock 1.169v, ie firestrike extreme, valley, heaven, fsu, etc. I am 3d bench stable also at 1480core with 1.21V

However to game with far cry 4 (30 + hours without any crashing) and GTA V over 50+ hours without a single crash, I require 1.21v for 1400 core stable. Both far cry 4 and GTA V will intermittently crash if I run 1400 core at bench stable settings. I need to be about 75mz lower overclock for a given voltage than 3d bench stable to run both of those without crashes. But depends on the game, I can run some other games near benching stable settings. And apparently, per others posts, COD AW is also taxing on overclocks and requires similar lower OC or higher voltage like far cry4/GTAV.

Also, I ran Display driver uninstaller in safe mode, then reinstalled all drivers (newest), changed performance mode to max from adaptive in nvidia panel, and then need to maintain enough voltage for a given overclock, then shouldnt crash at least in those 2 games, assuming you also have a stable OC, and assuming you are not crashing from P02 voltage mismatch with boost states tab settings.

Click to expand...

1450 mhz here. it starts at 1463 but a vdroop happens and it sits at 1450 even if i game for 8 hours . some games crash some dont .
so more voltage is what i need ?

yes, you simply have an unstable overclock, likely need more volts. What voltage are you using for 1450. What are your max temps gaming? There are 3 people that have found, just like me, that several games you need to be about 70mhz less than bench stable at a given volts. So if you can bench at 1520 firestrike and valley at a given voltage, than that should be close to voltage you need to run 1450 24/7 for most taxing games without crashing.

rge said:

yes, you simply have an unstable overclock, likely need more volts. What voltage are you using for 1450. What are your max temps gaming? There are 3 people that have found, just like me, that several games you need to be about 70mhz less than bench stable at a given volts. So if you can bench at 1520 firestrike and valley at a given voltage, than that should be close to voltage you need to run 1450 24/7 for most taxing games without crashing.

Click to expand...

1.230 for 1450 . temperatures are 75c at max with 85 % fan curve . usually 68 in less demanding games

Your temps will cause some throttling, so your core and voltage will drop down based on temps at certain times...and actually that can cause instability by lowering volts/depending on throttling logic and how it lowers core speed. But again for newer games that really tax gpu, you will need to have a higher gap between bench and game stable as well.

I would run 1.21-1.22v and 1400 for a while (assuming you are bench stable at 1450-1470 with 1.21-1.22v. And then if dont crash in any games for a couple days, then you can increase as tolearated, but on air your kind of high voltage for 24/7 use. Once you get above 60C, you are going to have throttling issues. There is no means to stop the temp throttling.

rge said:

Your temps will cause some throttling, so your core and voltage will drop down based on temps at certain times...and actually that can cause instability by lowering volts/depending on throttling logic and how it lowers core speed. But again for newer games that really tax gpu, you will need to have a higher gap between bench and game stable as well.

Click to expand...

would running the fan at 100% solve the problem then . I don't mind.
Or should I increase the voltage. Btw having 122+mv on precision causes instant crash. I have to have it at 81mv and even then it crashes but after an hour of gaming

upping the fan may help, would be trial and error to see, but your boxed in because of your temps. And not just gpu temps which you see, vrm temps and vram temps get really hot on air with higher volts and can cause stability issues. But you cant monitor them directly, some have measure them in various ways however. Also helps vram/vrm temps to hit back of card with fan air.

On water you could just increase volts. On air you may have to decrease core speed to get stable at volts your at.

I havent been on air, mine only on water, so dont know what to make of +122 causing instant crash, unless temp related, power related, or bug in bios you are using.

rge said:

upping the fan may help, would be trial and error to see, but your boxed in because of your temps. And not just gpu temps which you see, vrm temps and vram temps get really hot on air with higher volts and can cause stability issues. But you cant monitor them directly, some have measure them in various ways however. Also helps vram/vrm temps to hit back of card with fan air.

On water you could just increase volts. On air you may have to decrease core speed to get stable at volts your at.

I havent been on air, mine only on water, so dont know what to make of +122 causing instant crash, unless temp related, power related, or bug in bios you are using.

Click to expand...

I contacted arctic . Asking if their custom triple fan cooler is compatible with my titan .
according to the website it is
I'm considering it . Do you think it would help alot compared to the stock blower

What are the temps and are you throttling because of them? If the answer is yes, then a better cooler would help... again IF you are currently temp limited and throttling.

EarthDog said:

What are the temps and are you throttling be caused of them? If the answer is yes, then a better cooler would help... again IF you are currently temp limited and throttling.

Click to expand...

hey earthdog .
The fan is really really loud at 85% . And even then I can't stay below 74c .
i notice that after exceeding 65c a vdroop happens where I drop from 1.25v to 1.23
core clock goes from 1463 to 1450 which is still a great core for me . But the thing runs loud and hot
75c at full load. However it does maintain 1450 no matter how heavy the load is
it is not stable and down the road it crashes. Usually along with a bsod

The Titan X isn't supposed to throttle from temps until much higher, 83C to be exact..

Are you hitting power limits there and not heat?

EarthDog said:

The Titan X isn't supposed to throttle from temps until much higher, 83C to be exact..

Are you hitting power limits there and not heat?

Click to expand...

Nowhere close to hitting powerlimits
I usually sit around 90% power usage.
I have up to 115% power limit and 90c temperature target

So the question is why is it throttling...

What does GPUz say about it on the sensors tab? There is a OCP/OVP section there I believe...

There is throttling logic that above a certain power, you will get mild throttling in 60's temp, then heavier throttling as temps/power increases. Below a certain power, for example stock bios and stock volts, you may not get temp throttling until higher temps. Many on OC.net on air have demonstrated throttling in 60's at higher voltage/higher power. If however on water, you will not see any throttling on same modded bios/high volts, because you cant hit the 390-400W hard limiit even with 1.274 max volts and running any game or benchmark and temps will be below 60C, unless you run furmark. But on air you will see first throttle point at 60's C with high enough voltage in modern game/benches. If your a reviewer on stock bios, you wont see throttling until you hit 83C, because you cant get above power limit where you will get throttling at lower temps, because you will throttle on power alone....until you remove those limits via modded bios.

But after market cooling will depend on how well it cools, and if it helps cool vrm/vram. I dont know how well other coolers compare to stock ones.

Part of issue, memory modules on air, no backplate, getting over 100C, second review I have seen with thermal temps, mem modules and vrms get very hot.
http://www.tomshardware.com/reviews/nvidia-geforce-gtx-titan-x-gm200-maxwell,4091-6.html