4770k and OC

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

Guest
I am currently running an overclock at 4.5GHZ with 1.175v and running Prime now for about 20 minutes with a MAX core temp reaching 97c with my H100i at 100% Load

Should I be concerned?

Also, is it safe for me to lower voltage to lets say, 1.15? I want the best temps with the best stable OC and I don't know what to exactly watch for.
 
I would read this. http://www.overclock.net/t/1411077/haswell-overclocking-guide-with-statistics
The AVX instructions in p95 is probably bringing your vcore up to 1.275, which is still safe. Those are good temps and a good overclock on that voltage. 1.275 gets me 4.3Ghz without AVX.
Lowering the voltage will not hurt it, but it cause the overclock to become unstable and BSOD your PC. It is ok in that situation though. 😛
 
That's an excellent voltage for that OC. However you'll need to run Prime95 a little longer to know that you're stable.

As for the temperature, that's pretty good as well. The AVX instructions used by Prime95 really heat up the cores beyond any actual application. So if you are looking for an absolute worst case scenario, you won't find a better test than Prime95. However the load is one specific instruction set, so it's possible to have it Prime95 stable and have other applications crash.

As for lowering your voltage, you can't do any harm lowering your voltage. It will just become unstable and you'll have to raise it again.

One thing to be careful of when overclocking Haswell is that if you are using Adaptive or Offset mode for voltage control, the motherboard can go overboard (go beyond what you set) when you apply AVX loads such as Prime95. It's recommended that while dialing in your stable voltage for your OC, to use Manual. Once you're done dialing in your voltage and you're finished testing, then switch it to Adaptive or Offset (assuming you want Intel SpeedStep to work).
 


97c really isnt that bad for that voltage? Maybe I'm just over-reacting.
 


Changed it as I was testing and updating. My mistake.

I'm just really concerned about raising voltage. I don't want to kill my cpu in a few months by raising voltage too high.
 

Yeah, don't raise voltage at those temps. Check to see that is is seated correctly and isn't broken. Even my NZXT T40 is keeping my 4670k under 70c with .1v more.
 
a common misconception if people say too much voltage will damage your CPU or "kill", no matter how high you put it, it shouldn't effect it, though it's not best to (putting to the extreme of high, as that is unknown). Increase your voltage by 0.5V (only if you have stock voltages), then if not working gradually increase by [0.01V], run Prime95 torture test which will run your CPU @ 100%, you will soon know if something is wrong (screen flicker and PC will crash)
 

Um, that's insane. It can (and will) damage the CPU and possibly the motherboard power delivery system. 1.65v is not in any way ok for a current gen CPU.

 

Most will. Mine gladly takes 1.5v. (XF) I don't like much over 1.4v anyways, and if he is seeing temps over 97c anyways than that's a horrible idea. I have to go to sleep now, but let's keep him from killing CPU please? At least while I'm gone?
 
More voltage equals more heat. Heat is the big consideration when adding voltage. Another thing that is overlooked is electromigration. Essentially the deterioration of the conductive pathways in the die. The more voltage supplied, the higher the current. The easiest analogy is when a stream swells during run off. With the added current, the banks erode where the stream changes course (bends). Essentially the same happens for the conductive paths in the die. This can completely open the path, or increase the resistance of the path due to reducing the path volume. It's definitely something you want to avoid.

 


My error. When I was typing this Op I was running 1.5v.

I just raised it to 1.75v and I'm getting cooler Temps with a max cote of 73c on 100% load. Better?
 
Not saying to increase voltage further if he has already done at least 0.5V, then no he shouldn't.

On all modern day motherboards, especially for the 4th GEN CPU's, BIOS does NOT apply dangerous voltages past a certain point for a certain speed, due to most people who don't have a clue how much to increase it by.
 


yes, much better, Applying High Voltages won't do such damage when motherboard have limits to what is "safe".
 

Not always true, if voltage is not enough, it means more heat in some cases.
 
OK guys to wrap this up I just want a few answers.

When doing stress tests, what temp should I stay away from? Like is 85c bad or is it OK for 100% load?

What range of voltages are safe?

How long should I run Prime95 or IBT to make sure I have a stable overclock? I heard people say run it for 24 hours but I don't feel. Comfortable with my cpu reaching like 80-85c overnight. Makes me feel like I'll wake up to a fried component.

Thanks!!! These are my main final questions and I hope I get a response!
 

A) I like to stay under 90c when stress testing. If you use AVX daily I would move more towards 80c as a good limit, but that is personal preference. 90c should be safe in theory.

B) I don't like over 1.45v at all, and prefer to stay under 1.3v. Anything over this and efficiency drops like a rock. If your chip wasn't doing 4.6Ghz stable at least under 1.5v then the chip isn't worth overclocking much IMO. I would drop it to 1.3v and overclock from there. If you get 4.5-4.6Ghz then you might bump the frequency up a little along with the voltage. Many people are around 1.4v with haswell, but I'm not sure on that.

C) People seem to agree on 24 hours of p95 or 20 passes of linx. I just loop whatever I will be doing for ~12 hours. I doubt mine is game stable, but it is rock solid in game as video editing. (That's what I do.)