Overclock.net › Forums › Graphics Cards › AMD/ATI › Polaris Bios Editing ( RX5xx / RX4xx )
New Posts  All Forums:Forum Nav:

Polaris Bios Editing ( RX5xx / RX4xx ) - Page 36

post #351 of 3354
Proud owner of a Saphire RX 480 8Gb card, already put my H90 on it, strapped some getto fan under it, runs like a dream vs my old 290 Asus OC.
I have one question for any of the more expert-minded people here, does anyone know how I can increase the max power draw via I2C commands? (like the increase in max voltage)
I feel the performance could be considerably higher with a +100% power limit. I see that for now a custom bios wont do the trick, but if there is another way, I would love to try it.
With the extra 100 mw I can easyly get to 1400 stable, would love to try to push the card to its absolute max for fun biggrin.gif
Got 14334 in Firestrike -> http://www.3dmark.com/3dm/13368742
post #352 of 3354
You can't increase the power limit through I2C. It is a driver limitation and the VRM controller only provides OCP and OVP protections.
Edited by The Stilt - 7/19/16 at 7:48pm
post #353 of 3354
Quote:
Originally Posted by The Stilt View Post

You can't increase the power limit through I2C. It is a driver limitation and the VRM controller only provides OCP and OVP protections.

Well you kinda can. One can mess with the current sensing calibration and make the VRM think that the output current is lower than its actual value. Since the card reads this value to check if it is in the power limit this effectively increases the limit but of course also undermines the OCP and might kill your VRM section if the power draw is getting too high. Hence doing so is really not recommended. In case someone really wants to do that try the tool I hacked together: http://www.overclock.net/t/1605757/vrmtool-a-simple-tool-to-read-and-write-to-i2c-vrm-controllers
post #354 of 3354
Quote:
Originally Posted by deeper-blue View Post

Well you kinda can. One can mess with the current sensing calibration and make the VRM think that the output current is lower than its actual value. Since the card reads this value to check if it is in the power limit this effectively increases the limit but of course also undermines the OCP and might kill your VRM section if the power draw is getting too high. Hence doing so is really not recommended. In case someone really wants to do that try the tool I hacked together: http://www.overclock.net/t/1605757/vrmtool-a-simple-tool-to-read-and-write-to-i2c-vrm-controllers

Yeah, but for obvious reasons that's a really stupid thing to do as you lose OCP. With the phase current bias set to the maximum, it is not impossible to exceed the safe limits for phases 1-3 even on air cooling.

No need for such hack job when you can have the same effect by increasing the driver limits.
post #355 of 3354
Quote:
Originally Posted by deeper-blue View Post

Well you kinda can. One can mess with the current sensing calibration and make the VRM think that the output current is lower than its actual value. Since the card reads this value to check if it is in the power limit this effectively increases the limit but of course also undermines the OCP and might kill your VRM section if the power draw is getting too high. Hence doing so is really not recommended. In case someone really wants to do that try the tool I hacked together: http://www.overclock.net/t/1605757/vrmtool-a-simple-tool-to-read-and-write-to-i2c-vrm-controllers

Perfect, big thank you, that's what I was looking for, time for testing biggrin.gif
post #356 of 3354
Quote:
Originally Posted by Felix39 View Post

Perfect, big thank you, that's what I was looking for, time for testing biggrin.gif

Please only press any of the buttons in the tool if you do know how a VRM works! As The Stilt and I said it's possible to kill your card - so be careful. Also I do not take any responsibility for what you/others do with the tool at all.
post #357 of 3354
Quote:
Originally Posted by The Stilt View Post

Yeah, but for obvious reasons that's a really stupid thing to do as you lose OCP. With the phase current bias set to the maximum, it is not impossible to exceed the safe limits for phases 1-3 even on air cooling.

No need for such hack job when you can have the same effect by increasing the driver limits.

I am willing to risk it, first card I go this far, usually much more cautious, but I almost have a full second system in spare parts, so, for the love of science, I will try out everything I can throw at it. Exciting times...or frying times, we shall see biggrin.gif
Btw, thanks for the EMBD, good stuff!
post #358 of 3354
Quote:
Originally Posted by deeper-blue View Post

Please only press any of the buttons in the tool if you do know how a VRM works! As The Stilt and I said it's possible to kill your card - so be careful. Also I do not take any responsibility for what you/others do with the tool at all.

I know mate, the thanks where for you providing the possibility, my actions are my own, not gonna run around blaming other for my mistakes, but hey, lets hope it doesn't die and I can see where this little card can go. Btw, about killing it, after yesterdays reading from HWinfo64, my card can take millions of watts and volts literally eek.gif talk about weird readings biggrin.gifbiggrin.gifbiggrin.gif

post #359 of 3354
Thread Starter 
Anyone got past 1.4ghz at rasonable voltage? I can do 1.4ghz@1162mv but even 1405mhz is very unstable at 1175mv.
post #360 of 3354
Quote:
Originally Posted by generaleramon View Post

Anyone got past 1.4ghz at rasonable voltage? I can do 1.4ghz@1162mv but even 1405mhz is very unstable at 1175mv.

Not me, at 1400 I need at least 1.20 to get it stable in everything, so your lucky I would say wink.gif Will try next week to see what I need for 1450 and maybe 1500, but at what I need for 1.4, doesn't look to promising. Praying that AMD will allow custom bioses
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › Polaris Bios Editing ( RX5xx / RX4xx )