The RT mod works fine at my 280 ... 2D - 0,78v ;)
3D extra 1,35v > 840/1730 ...
I made one misstake, I interchange hex and dez one time :shocked:
Conclusion ... 1,65v unter full load by furmark, ... :eek:
Printable View
unwinder, i just have 2 questions:
is it possible to change voltage on G92 65nm reference nvidia boards using rivatuner exclusively?
not asking how, just asking is it possible and have you ever done it, btw?
best regards :)
So no fried card, cool.....and you can read IC2 aswell.....how could this have possibly happened? :p:
If you'd used a slider do you think you'd have accidentally selected 1.6v?
So there's no voltage slider control in RT, so you probed and calculated hex, ballsed it and could've killed your card.......and all the safe unwinder way :p:
I'm starting to see the sense in this now. Atleast you had to use math before frying your card.
:rolleyes:
No need to check BIOS, try to probe I2C device on that address instead. If there is some valid I2C device - you have at least a chance. But that's just a chance, the reality devends of the PX3544 programmability level (it can allow both I2C voltage override and monitoring as well as a single monitoring only).
Im trying this on my blue PCB Sapphire 4870 but with "C:\RivaTuner\RivaTuner.exe /wi3,70,18,44" nothing happens (im trying to increase from 1,2625 to 1.3v), how could i know if this doesnt work for my card?
Just look at the card - if there's a VT11xx controller on it. If not, then it won't let you program voltages.
I used the rivatuner method without issue on my GTX280. Fantastic tool! Wish I knew about it before.
All we need now is a simple guide on how to do it and everyone is happy. :up:
KELLS: you're drifting very close to the troll line. The reasons are given already.
His post makes good logical sense though, and you know it. I don't care about any "reasons" as the given reasons are weak and there are countless ways to prevent the scenarios outlined in the reasons why it won't be done. And I don't know about Kells, but I feel that I have plenty of room to complain seeing as I would be willing to pay money for this feature.
I got a valid device, it's device 2e...
So i've put the command /ri,2e,1A and it turned out to be not invalid, but 00.
Seems like a starting point, huh? :)
http://img142.imageshack.us/img142/1...6659rw6.th.jpg
Edit: Now, I think I need brains bigger than mine to figure out the reg values, to be written... For that, should I get a new (and bigger/better :p:) brain or is someone knowing willing to provide? :)
[mod hat on]Everyone can complain all they want about not having sliders....just not in this thread, or really anywhere on XS. :shrug:
This thread is about an upcoming eVGA tool that changes voltages and the thread evolved into how this functionality is already possible on certain cards with existing software.
If you want Unwinder to include a specific subset of functionality please do not address him via this thread on XS. All it's doing is cluttering this thread and annoying people--essentially trolling. I wouldn't recommend harassing him via PM on XS either, that's PM Spam and we don't take it too lightly ;)
He's made it very clear his stance is "no." As such, the persistent trolling of people trying to convince him otherwise needs to stop. :shakes:
[/mod hat off]
This is what is on my mind...
I figured out that my 280 is in the fourth bus, 03 I2C.
On my system, my 4 registers are set like this, using BIOS v62.00.0E.00.00 on my 280:
Register 15 - 1.1875v (Performance 3D Mode)
Register 16 - 1.0625v (Low Power 3D Mode.)
Register 17 - 1.125v (Unused voltage entry.)
Register 18 - 1.0375v (2D Mode.)
Using RivaTuner's monitoring tools:
When my Core is at 300, my voltage reads 1.04v, looks like register 18.
When my core is at 399, my voltage reads 1.06v, looks like register 16.
When my core is at 756, my voltage reads 1.19v, looks like register 15.
http://www.evga.com/forums/tm.asp?m=...55974;?
I did ask if my register 17 for shader voltage, or if it even worked that way.
Posted by Unwinder:
"No, it is not a shader voltage. In serial VID encoding mode (this mode is used on both 4800 and GTX200 series) VT1165 VRM allows generating just one of these 4 voltages defined with registers 15-18 by 2-bit VID defined by state of VRM input pins at the moment on time. Other words, it doesn't allow generating multiple voltages simultaneously and there is no such thing as shader voltage. VRM supports 4 different voltage entries in serial VID mode and your GPU needs only 3 of them (one for 2D, one for Low Power 3D and one for performance 3D mode. So that is just an unused voltage entry".
So on my system, register 17 is just an unused voltage entry...
He further confirmed for me that the voltage value: "applies only to GPU core (and stream processors aka shaders, as they are part of GPU). Memory chips are driven by differnt VRM (not software controlable)".
This tool will probably not help my memory run any faster....
My water cooled 280 current max settings are: Core=756, Shaders=1512, and Memory=1296. I currently can't go any further on shaders or memory, withought the system re-booting under load.
(My Core is limited by my shaders right now...)
Heat is not an issue, as my GPU only hist 53C max temp, with both Prime95 and Furmark running simotainiously on my system.
So my question is how high of voltage will be safe for a cool running 280?
I imagine we will only be setting the (Performance 3D Mode) with EVGA's tool?
I am currently at 1.1875v in my BIOS, how high should I go?
I have seen another guy go up to 1.35v on water with very impressive results, like 840Mhz core speed or something like that. It is posted in the water cooled GTX280 story thread under liquid cooling.
I'm limited by heat at 1.25v but it definitely helps, I can run 760mhz core on Furmark for about 60-70% of the pass before I start to get artifacts probably due to hitting near 80c. At 1.19v I'm limited to 740mhz, 750mhz gets me a display dropout.
FTW for Ray Adams, for implementing voltage control via ATI Tray Tools... Ray Adams is da man! Praise to him!!!
FTW for Evga doing it now.. Regarding the topic of this thread, I'm wondering if Evga has a more efficient method of Vcore increase that allows for greater overclocking than the VRM hex method.
Unwinder, thank you anyways for participating here on XSforums, giving us a few hints, etc.. Oh well, I'm sure you have your own reasons for whatever you're doing. If you could tell me how to add a GUI interface for adding the voltage sliders to RT, I'd gladly do it for you or for 95% of the RT users who love overclocking but have not come across this particular thread... Not to beat on your own dead horse, but would a programmer need your permission to add this "mod" to RT?
Here's a much better guide:
http://www.ocxtreme.org/forumenus/showthread.php?t=4427
The other question is on voltage, what is safe for a 65nm 280, is also safe for a 55nm 260?
One would think that 65nm -vs- 55nm 200 Series GPU, would be able to brunt more voltage before the chip would go south...
Will all GPU's have the same range on the slider, or will the available range be adjusted based on the GPU?
Thanks for the report, I hope all 280's can make good use of some extra juce. :D
So my early 280 BIOS version currently has me sitting at 1.1875v's in Performance Mode.
You have taken your's up to 1.25v's, and lived to tell the tail. ;)
And Wild Man OliverM84 volt moded his 280 to > 1.36v's... :p:
I need to bump my Performance voltage soon! Go EVGA!!
My Q6600 does better for me at 1.5 volts, maybe if my 280 is given more power, it will too?