Core i7 4770K OC Voltage?

[theonerm2]

What is the max voltage I should use with my Corsair H105 cooler? I'm thinking 1.35v should be max but I've been pumping 1.5v through the chip trying to get 4.5GHz. It's not stable at all at 4.5GHz or 4.4GHz. Doesn't matter what settings I use either I just can not get it stable. There was a time I got 4.497GHz stable but I can't seem to get it now. I don't remember those magic settings.

 

[Philballer17]

That's really unethical man.

4.5ghz is fine, any higher really won't yield that much more performance.

This is true, dont take my advice. lol

 

[theonerm2]

LOL. You aren't alone. My 1700X (last I checked) only hit 3.8ghz.

It doesn't actually fail. It just gets too hot under occt. The stress test I'm running. With my RAM at 2400MHz I get higher temps. At stock speed I get 80c with my RAM at 2400MHz. This is under OCCT. And at stock speed with the RAM at 1333MHz I get 59C under load with OCCT. Huge difference. The problem is though that the auto voltage takes the 4.5GHz OC to 1.337V on the Vcore. Even when I manually set Vcore to 1.25V it still shoots up to 100c. I might try realbench instead. It might not run it as hot as OCCT does. I know silicon lottery tests with realbench for an hour and says they're stable. So I'll try it too. Also my cooler used to run 1.35v on the vcore just fine. Something is wrong now it seems. I get 40c idle when I overclock. That can't be right. I don't know if it's the cheap thermal compound I'm using or if my cooler is messing up on me. You wouldn't think thermal compound would make that huge of a difference. I'm using Protronix Series 7 Silver Thermal Paste. I used liquid metal in the past. Boy was it a mess to deal with when trying to clean it off. It got on the motherboard and everything. There are still discolored spots on my AIO from it. But it seemed to run a lot cooler when I had liquid metal on it. It was last year I believe when I changed the thermal compound from liquid metal to Protronix Series 7 Silver Thermal Paste. It was cool labs liquid pro. I had previously only used Arctic silver 5. I'm not sure if it's the thermal compound or what but something is wrong with my cooling.

https://www.amazon.com/Protronix-Thermal-Performance-Heatsink-Compound/dp/B01MU2FLVJ

 

[theonerm2]

I just noticed Core #3 only gets up to 72C while Core #0 reaches 100c. I don't want to delid again. Even at 4.4GHz the thing gets too hot.

 

[Philballer17]

I just noticed Core #3 only gets up to 72C while Core #0 reaches 100c. I don't want to delid again. Even at 4.4GHz the thing gets too hot.

That means you have a bad chip. Core 1-4 temps should all be around the same number. Core 0, and 1 will be a bit higher, but should only be 4-5 degrees higher than others. That chip is definitely bad trust me.

 

[Georg Rauh]

I have an Asus Z97-A. I got the 4790K today. Tried overclocking. 4.5GHz fails. Runs fine at stock speed though. Haven't tried 4.4GHz yet. My luck with cpu's and overclocking is so terrible. I think I'll quit overclocking for good now. I've been burned too many times. No more Z boards. No more AIO coolers. No more K chips. I'm done overclocking CPU's. I'm just going to retire my H105. Lord knows it's going to fail soon anyway. I've had it running 24/7 for 2 years now. I'm going to get a Hyper 212 evo. Overclocking costs so much extra. I paid $60 extra for my motherboard, $50 extra for the CPU back when I got it, I paid $80 extra for the cooler. That's almost $200. All for overclocking and I'm just keeping it at stock speed in the end. I wish I just got a cheap board and a Xeon E3 1231 V3 back in 2014. That way I'd never have been worried about overclocking to begin with.

As mentioned, I have a very old 4770k (non delid), and a Asus Z87 Pro (very good board). Mine also does "only" 4.4 at 1.250 and it's getting already crazy hot in synthetic tests (high 80s up to 90). With tests that do AVX commands, like latest Prime even hotter.

I also could in the years I have that CPU not reach 4.5, I'd need at least 0.100V more (would have to delid), there is a HUGE step for me between 4.4 and then anything higher. Saying some chips simply don't go higher. I am totally fine with my 4.4 tho. And stop testing with OCCT etc....you wont ever get happy since Haswell just DOES get insanely hot, you just have to accept it and stop synthetic tests. Run Real Bench instead.

 

[Stian van der Krug]

Wow, you guys are missing out. My max OC is 4.9 GHz!! And my daily OC is 4.7 GHz at 1.325v. Delid and not over 70°C under stress. Custom water loop.

 

[zebarjadi.raouf]

Geez, first of all, Intel was really rusting when applying glue on the Haswell as it was one of the first attempts when transitioning from solder, and unfortunately by now, the thermal paste on all chips should have turned to cement too. You're better of delidding it, scrapping the glue and TIM, applying something thick like AS5 or Z9 to resist pump out effect or go Liquid metal. I just delidded mine with a razor. temps difference between first and last core is 5c.

Second of all, I refuse to believe Haswell can't do at least 4.2GHz. I got mine running at 4.5GHz with my first attempt. Haven't tried to find the max as I never needed to. Might do it for fun later on. Should be around 4.8-5.0 with the voltage I'm running on.

Did you do your overclock correctly? Everything set manually (not auto), a modest increase of Agent voltages, LLC enabled, power saving disabled, stability options enabled, ...

 

[Gam3r01]

While it technically does not meet the requirements to be closed, keep in mind this thread is over a year old.

 

[Philballer17]

Should have never de-lidded to be honest. You probably didn't do it the right way, which is why you have a hot spot in core #0 reaching 100C temps. You shouldn't do what testers and high performance overclocking techs do. They're just trying to achieve a higher benchmark score. They wouldn't ruin their personal chips by de-lidding either. Good luck. The 4770k is a sad excuse of a 4790k anyways.

 

[zebarjadi.raouf]

You probably didn't do it the right way, which is why you have a hot spot in core #0 reaching 100C temps.

He probably didn't position the IHS properly. Happened for me on my first try.

You shouldn't do what testers and high performance overclocking techs do

I'm not a tester, nor am I an enthusiast. My 4770k had gotten old, getting relatively high temperatures, thus I decided to delidd to give it a boost. Something like cleaning your GPU or heatsink.

The 4770k is a sad excuse of a 4790k anyways

4790K is 4770K, just optimized and rebranded. i7 Haswell can go toe to toe with i7 Skylake if you do your OC and delid properly. 20-40c drop on temps with 1.5GHz+ OC is no laughing matter. That means you don't have to upgrade for another couple of years.

My temp drop was 37c, do note it was 5 years old and a Haswell at that, not brand new like the ones you see on Youtube which give 3-10c boost. Right now I'm running 4.5Ghz with 62c max FPU stress test,