Overclock.net › Forums › Intel › Intel CPUs › Should I reverse my overclock?
New Posts  All Forums:Forum Nav:

Should I reverse my overclock? - Page 2

post #11 of 17
Quote:
Originally Posted by INCREDIBLEHULK View Post

What do I do? Do i run this chip 24/7 at 1.430v in bios with a droop always or do i run this chip at 1.370in bios with an occasional vrise due to Turbo LLC?!

Having a rise in load voltage is not ideal, but it you are truly in a situation where one notch of LLC changes you from needing 1.37V to 1.43V, then I would take the lesser of two evils and go with the 1.37V settings. That Gigabyte board has very good power delivery circuitry, so I doubt the long-term risk to the VRMs is going to be all that severe. The reduction in voltage and, more importantly heat, in the CPU is going to be more beneficial in the long-run.

Are you sure it is stable with the 1.37V setting though?
post #12 of 17
Thread Starter 
Quote:
Originally Posted by Forceman View Post

Having a rise in load voltage is not ideal, but it you are truly in a situation where one notch of LLC changes you from needing 1.37V to 1.43V, then I would take the lesser of two evils and go with the 1.37V settings. That Gigabyte board has very good power delivery circuitry, so I doubt the long-term risk to the VRMs is going to be all that severe. The reduction in voltage and, more importantly heat, in the CPU is going to be more beneficial in the long-run.

Are you sure it is stable with the 1.37V setting though?

Yes, I will try again when I get home, I was under the impression from my other threads to never have a rise because the VRM's have to correct it and this causes added heat and damage.

For PLL I really dont want to mess with the other settings on a 4.5, seeing as what this other gentlemen mentioned about users going from a cool chip with just vcore to modifying system agent, pll... To me auto is working fine smile.gif I see in hw monitor that it goes from 1.6-1.8 most of the time stays at 1.7


With turbo LLC, my vcore almost completely stays the same with whatever vcore I use.. My worry is that under stressing loads it does +0.010 to whatever vcore, (extreme does +0.020)

So I could be stressing IBT or prime, and in these constant case scenarios it will add +0.010 and pass everything, when using the computer randomly i'll be doing a bunch of things and I'll occasionally see cpu-z jump up +0.010 when a few applications use a lot of cpu+ram


This was my whole worry in the other thread!

Do I stay at 1.430v with a drop from 1.424 to 1.416 and never experience a a rise or do i stay at a 1.37 with a rise of +0.010 in load situations! frown.gif
OC'ing is the devil!!biggrin.gif
  
Reply
  
Reply
post #13 of 17
Quote:
Originally Posted by INCREDIBLEHULK View Post

Yes, I will try again when I get home, I was under the impression from my other threads to never have a rise because the VRM's have to correct it and this causes added heat and damage.

To be honest, you're overthinking it. Yes, you don't want a Vrise, but you also don't want to pump 1.43V through the CPU when you could be using 1.37V instead. So it all comes down to which way you want to go on the trade-offs. Neither situation is going to cause an immediate failure in your system, or probably even noticeable degradation. On a cheaper motherboard, or one with sketchy power regulation and delivery, I might err the other way, but I really don't think you have anything to worry about in your specific situation.

If you want to mess with CPU PLL or VCCSA you can normally reduce them a notch or two without affecting stability (which helps a little with CPU temps) but Sandy chips don't seem to be as sensitive to those changes as Ivy chips are. I'd just leave it in Auto.
post #14 of 17
Thread Starter 
Quote:
Originally Posted by Forceman View Post

To be honest, you're overthinking it. Yes, you don't want a Vrise, but you also don't want to pump 1.43V through the CPU when you could be using 1.37V instead. So it all comes down to which way you want to go on the trade-offs. Neither situation is going to cause an immediate failure in your system, or probably even noticeable degradation. On a cheaper motherboard, or one with sketchy power regulation and delivery, I might err the other way, but I really don't think you have anything to worry about in your specific situation.

If you want to mess with CPU PLL or VCCSA you can normally reduce them a notch or two without affecting stability (which helps a little with CPU temps) but Sandy chips don't seem to be as sensitive to those changes as Ivy chips are. I'd just leave it in Auto.

Thanks mate

You are right, I seem to take things too serious when sometimes I read too much about it or it's mentioned.

The way people made it seem was if I have spikes of vcore rise it's like a nuclear threat to my chip and that it is not acceptable at all

This whole time i've been questioning myself if the constant 1.43vcore outweighs the 1.37 with spikes here and there.

If it's not going to take a massive toll on my VRM's and harm my chip the way I thought you are definitely right.

At the same token, would you say I should bump LLC to extreme instead of high and lower my vcore even more than it is now or would i just be really pushing it in this case?
  
Reply
  
Reply
post #15 of 17
Raising your LLC even higher shouldn't lower your Vcore much, if at all, since you've already eliminated the Vdroop. That's what is letting you run a lower voltage, the fact that the voltage doesn't drop as much under load with Turbo - if you go to Extreme it might look like you can run with a lower Vcore (set in the BIOS) but what is being delivered to the chip is likely the same either way. So with Turbo you have 1.37 set and 1.37 is being delivered, but with Extreme you set 1.35 and still end up with the same 1.37 delivered (because of Vrise) - that's an example, but you get my drift.
post #16 of 17
Thread Starter 
Quote:
Originally Posted by Forceman View Post

Raising your LLC even higher shouldn't lower your Vcore much, if at all, since you've already eliminated the Vdroop. That's what is letting you run a lower voltage, the fact that the voltage doesn't drop as much under load with Turbo - if you go to Extreme it might look like you can run with a lower Vcore (set in the BIOS) but what is being delivered to the chip is likely the same either way. So with Turbo you have 1.37 set and 1.37 is being delivered, but with Extreme you set 1.35 and still end up with the same 1.37 delivered (because of Vrise) - that's an example, but you get my drift.

I hear ya smile.gif +rep for your help mate

Frankly, I'm still trying to put it in my head that rise might be okay and to ignore that "any droop is acceptable and any rise is unacceptable" mentality.

I found 1.430 ridiculous just to have a droop instead of a minimal rise
  
Reply
  
Reply
post #17 of 17
Thread Starter 
While revising my OC, I noticed Internal CPU PLL Overvoltage is disabled and I can't remember why I did this.

Looking through some guides now to see if I should have this on or not
  
Reply
  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Should I reverse my overclock?