Using V/F point only instead of combining it with offset voltage?

[bobert]

I think most people have been using offset voltage to change the voltage for all frequencies because that's how it's been done for all these years.

But now that V/F point is available, wouldn't it be better to just only use that?

For example, here if I just cap turbo to 34 and then lower and test V/F Point 4, then I don't have to worry about the cpu crashing at lower frequencies because the voltage is too low right?

Has anyone tried just using V/F point and unchecking unlock adjustable voltage?

 

[unclewebb]

wouldn't it be better

What is your definition of better? It is possible to only use V/F Point tuning and not use any offset voltage values. What do you gain by doing this?

Having to enter 11 individual V/F Point values seems like a lot of work. It would also require a lot more testing to make sure all of the V/F values are stable.

Setting an overall offset value plus a single V/F Point 1 value seems easier to me. I use the Windows High Performance power plan and only run my CPU at one speed so your V/F Point only method might actually work quite well. Most people that are using the Balanced power plan and running their CPU at a wide variety of speeds are probably not going to like your method.

 

[bobert]

Why not just set the V/F point at the highest turbo value you have set? Then you don't have to set the others right?

 

[unclewebb]

I usually run at 5200 MHz. For V/F Point tuning there is a V/F point at 4800 MHz and the next one is at 5400 MHz. Do I need to adjust both of those or only one of those? Do I also need to adjust all of the V/F Point values higher than 5400 MHz? I have never done any testing so I do not know the answer to these questions.

Intel says the voltage curve has to continuously increase or all of the V/F points will be ignored. Based on that, you might have to adjust the values higher than 5400 MHz.

This is interesting but how is it any better than the old method that everyone is familiar with? What is the advantage?

 

[bobert]

This is interesting but how is it any better than the old method that everyone is familiar with? What is the advantage?

Because the thing about having to set the V/F point at 800 mhz got me thinking. What if your undervolt is crashing because you need to set a higher voltage at all the points between 800 mhz and your max frequency?

We are assuming high undervolt is causing a problem at 800 mhz, and not all the other points inbetween like 1700 and 2400 which could be a mistake.

So what if we just leave the lower frequencies unchanged and only change the top frequency? Then you could be assured that the lower frequencies aren't crashing your cpu.

I usually run at 5200 MHz. For V/F Point tuning there is a V/F point at 4800 MHz and the next one is at 5400 MHz. Do I need to adjust both of those or only one of those? Do I also need to adjust all of the V/F Point values higher than 5400 MHz?

Based on the website you linked from throttlestop, you'd have to set both, and the voltage you get would be interpolated between them.

Intel says the voltage curve has to continuously increase or all of the V/F points will be ignored.

If this is true then doesn't only boosting V/F point 1 negate any undervolting you do?

 

[unclewebb]

What if your undervolt is crashing because you need to set a higher voltage at all the points between 800 mhz and your max frequency?

That is possible. You can adjust the do this by changing the main offset value in the FIVR window.

So what if we just leave the lower frequencies unchanged and only change the top frequency?

This might work. One would have to do some testing to see how the interpolation works.

doesn't only boosting V/F point 1 negate any undervolting you do?

When you boost V/F Point 1 a large amount, the CPU seems to automatically boost V/F Point 2 and any other V/F points that need boosting. This helps to make sure that the voltage curve is ever increasing as the MHz increases. You would have to increase V/F Point 1 a crazy amount before it would negate your undervolt at high MHz.

You have some good ideas. Do some testing. Adjusting only a couple of the high MHz V/F points might become a new and simpler way to undervolt. You could lower a high MHz V/F point to reduce max power consumption at full speed without the voltage getting too low at low MHz and causing crashes.

 

[bobert]

I will try it but it seems hard to even see if anything changed because intel removed the vf curve visualizer from XTU.

If it works the way you say it does, then all you would have to do is change the top and bottom points. I guess its about the same as the current method but you can do it in the same window. But then again you won't be able to use a top frequency that isn't in the v/f settings.

 

[Ashlander]

I've also been experimenting with using only the VF points with my 14900HX, using the same method, only I started at the top end and am working my way down in frequency. The main benefit I'm finding is not being locked down to the lower UV allowance in the higher mhz range. Mine can run stable with -190 at 4800 and -160 at 5400, but anything above -110 crashes at higher mhz. If I did just the global offset, I'd have to keep it at -110 (with a small 800mhz boost), which leaves massive room for improvement precisely where it tends to sit during an all-core load.