Cooling requirements of i5-8600K vs. i3-8350K

[themanuel2]

Does anybody know frome experience or data available in the web whether i5-8600K requires more cooling than i3-8350K to achieve the same overclocked speeds, due to having six cores instead of four?

Maybe this is a dumb question as in theory it sounds like it should be this way but perhaps there are other factors that make them both comparable in practice.

I'm referring to an average basis here. I realize that every individual chip is slightly different.

 

[Phaaze88]

More cores = more heat
The 8600k does indeed require more effective cooling to overclock than the 8350k.
Now, if you wanted to compare the 7600k to the 8350k instead, the difference would've been negligible.

 

[SkyNetRising]

i5-8600K is 95W TDP chip, i3-8350K is 91W TDP chip.
Cooling requirements for them are pretty much the same.

 

[themanuel2]

Thanks folks, that's why I am hoping for an answer based perhaps on actual experiments as we can deduct something from information in paper but the actual results may prove different.

I have an i3-8350K now and I'm having regrets for not getting the i5-8600K instead so I might still exchange it before the return period ends. My concern is whether the cooling system that I have, CoolerMaster ML120L RGB (120mm) will suffice for the i5-8600K. It's not stellar but does the job for my i3-8350K.

 

[Phaaze88]

Thanks folks, that's why I am hoping for an answer based perhaps on actual experiments as we can deduct something from information in paper but the actual results may prove different.

I have an i3-8350K now and I'm having regrets for not getting the i5-8600K instead so I might still exchange it before the return period ends. My concern is whether the cooling system that I have, CoolerMaster ML120L RGB (120mm) will suffice for the i5-8600K. It's not stellar but does the job for my i3-8350K.

http://www.coolermaster.com/tdp-and-socket-compatibility/
It's rated for a 180w tdp cpu. It's more than enough for an 8600k. Switch if you want to.

 

[SkyNetRising]

CM ML120L RGB is for cpus up to 180W TDP. It's plenty for both of those cpus.

 

[7664stefan]

The ML120L will suffice for the i5 most probably for clock rates up to 4,5Ghz. Could not find a review to prove this. Above, a more powerful cooler with a 240mm radiator would be my recommendation.

 

[themanuel2]

Thank you both. I suppose I could live with a couple hundred fewer MHz in order to have the extra cores for future proofing and for a few applications I have now that would benefit from it right away.

Perhaps I should approach it that way.

 

[7664stefan]

Currently the i5 would also be my choice for a gaming rig where actually the difference to the OC i3 is rather minor.
What I feel the gaming industrie is still lacking is a proper multi core or hyper threading support. Advantage is still too low.

As a reference: https://www.tomshardware.com/reviews/intel-coffee-lake-core-i5-8600k-cpu,5264-11.html

 

[theonerm2]

Check this out. https://www.hardocp.com/article/2018/01/17/cooler_master_ml120l_rgb_allinone_cpu_review/3

They used a 3.9GHz Ryzen 7 1700 CPU. It drew slightly over 150watts. So 152 watts = 83c roughly on that cooler. Now a 5GHz 8600K will draw anywhere from 160 watts to 170 watts in prime95. It depends on the CPU. 170 watts = 93c roughly on that cooler. With the GPU also running a stress test your CPU would hit a toasty 96c. That's just 4 degrees from boiling point of water. Keep in mind these are stress test numbers. Your numbers while gaming will be 15c to 20c lower. So I'd expect to see average temps in the upper 70's to lower 80's. That's within the acceptable range. If you only OC to 4.7GHz or so temps will be reduced by 5c to 10c under prime95 load.

Also keep in mind these numbers are only estimations as I do not have the actual hardware to test with. I used simple algebra to do most of my calculations.

Even if I couldn't overclock very far with the 8600K I'd still get it. You get 50% more cores.

 

[themanuel2]

Thanks for the link to the poster above and to everyone else for your comments. I think it's safe to say that, at worst, I'll be trading off a small amount of speed for tons more potential with 50% more cores. My only loss would be a slight degradation on one or two emulators that are dual core bound and rely on raw per/core performance so it's not a big deal.