How far should I push this 8700k?

[testtube5]

So I started another thread about something else and it turned into questions about overclocking, so I'm posting this fresh thread.

If you care to read it, it's here: http://www.tomshardware.com/answers/id-3640097/safe-temperature-overclocking-8700k.html

I just overclocked my 8700k (With an h115i) to 5.0GHZ, Max temperature after 10min of AIDA64 = 77C

Radiator Fans running at 1500 RPM (Intake, Front Mounted Radiator), Case fan running at 2200 RPM (Intake), Exhaust Fan running at 1200 RPM.

(Going to replace the exhaust fan with the Noctua 3000RPM, it's just a stock 140mm at the moment).

Would it be safe to try for 5.1GHZ now (If I turn up my radiator fans to 2500-3000RPM (I don't care about noise at all)?

Also, another question. I set voltage to 1.350v. If I am stable at 5.0GHZ at that voltage, and I slightly lower that voltage, say to 1.325 or 1.3, would CPU temperatures go down while maintaining the overclock?

And one last question. After I overclock, how long should I run AIDA64 (CPU, FPU, Cache), before I know it's a stable clock and that I can attempt going higher?

My Build: https://www.pcpartpicker.com/list/FZXYcY

 

[volkgren]

Sounds like you got a good chip. Most try for 5.1GHz, but it honestly won't make any real world difference in gaming performance.

Lowering voltage will lower temp, it can also cause instability. Just find your lowest stable voltage for optimal thermals. Though 77C is pretty darn good, better than I expected. I tend to use Prime95 Small FFT's though (v26.6).

For how long to run test? Different opinions. My opinion is however long you feel comfortable with your system. If it's just my personal gaming system and I'm having fun 20-30mins. If it's more serious or long-term around 2 hours.

 

[testtube5]

Sounds like you got a good chip. Most try for 5.1GHz, but it honestly won't make any real world difference in gaming performance.

Lowering voltage will lower temp, it can also cause instability. Just find your lowest stable voltage for optimal thermals. Though 77C is pretty darn good, better than I expected. I tend to use Prime95 Small FFT's though (v26.6).

For how long to run test? Different opinions. My opinion is however long you feel comfortable with your system. If it's just my personal gaming system and I'm having fun 20-30mins. If it's more serious or long-term around 2 hours.

I AM gaming, but I need my CPU to run as fast as humanly possible for something else I do: I massive multibox in a 20 year old MMORPG called Everquest. It uses absolutely no GPU power whatsoever, and requires as much CPU power as possible, and as much RAM as possible.

Curious though, do you think I can even push to 5.2GHZ based on what you see with the temperatures I just posted? Or will you not know until you see what my temperatures are at 5.1GHZ?

 

[volkgren]

Anything I say would simply be speculation. We wouldn't know until you tried. All chips are different. I'm more familiar with the i7-7700K because that's what I own. And at 4.9GHz 1.35V mine didn't do well, it went past 90C. I delided and used liquid metal. But that's a different story. That's an option though, it dropped my temps by 20C. I could then do 5.0GHz 1.4V and still not go over 80C.

 

[testtube5]

Anything I say would simply be speculation. We wouldn't know until you tried. All chips are different. I'm more familiar with the i7-7700K because that's what I own. And at 4.9GHz 1.35V mine didn't do well, it went past 90C. I delided and used liquid metal. But that's a different story. That's an option though, it dropped my temps by 20C. I could then do 5.0GHz 1.4V and still not go over 80C.

Update: After taking all my intake fans to 3000RPM, I dropped to 74C (from 77C) Max in AIDA64 (Downloading Prime95 now). The only thing is that I think I have WAY too much positive pressure, 3x3000RPM intake fans vs a single 1200RPM Exhaust Fan, may be able to squeeze out another 0.5-1C drop with the 3000RPM Exhaust Fan.

Question for you: (I'm pretty sure you weren't the same guy I asked about delidding in the previous thread). Let's say I were to get a delidding tool, how dangerous would you say delidding would be if I did it very carefully? The idea intrigues and scares me at the same time.

 

[volkgren]

Well, I used a razor blade. No delidding tool. And it went well. Started at the corners, slow and steady, not much room for error EXCEPT going too deep.

If I had a delid tool I would have used it. But looking at them they seem like there's a lot of potential energy that builds up when applying pressure and I'm not sure what preventative measures are in place on the tools to keep that from slipping and damaging the CPU die. They may be more safe than they look, I simply don't know.