8600k stable overclock
8600k stable overclock
I have been testing the stability of my 8600k at 4.8 ghz and noticed consistent performance in the 1.19-1.21v range. Cinebench runs smoothly, but I've seen others requiring higher voltages, which worries me about reliability. What additional tests should I perform to confirm stability?
Be honest with me: it seems unlikely you can achieve 4,8Ghz at a reasonable temperature.
Increasing from 4,5Ghz to 4,8Ghz will consume even more power, possibly requiring around 1,26V to 1,3V for Prime95 stability.
With such high voltage, you might encounter temperatures in the 90s or higher.
You should be grateful to Intel for this limitation. Their TIM between IHS and DIE is quite poor...
Execute Prime95, if your computer doesn't crash after several hours of stress testing, then you possess an excellent CPU.
I'll try it out. With prime 95, should I be concerned about high temperatures? My last use brought me up to the high 70s after around 30-40 minutes.
The CPU won't immediately shut down when it hits 80 to 90 degrees. Avoid keeping the CPU running continuously at this temperature; Prime95 is an extremely demanding stress test. For typical tasks like gaming or video editing, you won't reach such high temperatures compared to Prime95. If you manage to stay under 70-80 degrees during several hours of Prime95 use, your CPU will be fine for most situations.
To second Cookie's reply, numerous stress tests resemble running a 100m race continuously for hours... In reality, such intense scenarios are uncommon, just as demanding tasks rarely appear in everyday life. The same applies to your computer; you're prompting it to perform actions it wouldn't normally handle.
To second Cookie's reply, numerous stress tests resemble running a 100m dash over and over for hours... In reality, you'd seldom face such demanding scenarios. The same applies to your computer; it's being asked to perform tasks it wouldn't normally handle in real life.
Prime 95 version 26.6 is being used. A newer release would also evaluate AVX instructions, which could significantly increase CPU temperature. Unless you're performing intricate scientific computations, testing AVX isn't necessary. Focus on simple FFTs rather than the blend test—blend tests mainly affect memory and don't strain the CPU.
The CPU will reduce performance to safeguard itself at 100°C. Unless your cooling setup is very strong, you likely won't need to worry about this. Also, with only a modest overclock of 4.8GHz and a low Vcore, you might need to adjust it slightly. You should notice the change.
Don't stress if you encounter issues—this is just the system protecting itself.
Good luck with your overclocking journey!
Thanks for the answers you received about the voltage being too low for the prime tests. It didn't cause a crash, just halted the test runs. I wasn't aware last time I used prime that it would run different tests and I thought the temperature was increasing without realizing it was changing. I lowered the multiplier to 45 and set the voltage to 1.2v, which successfully passed all tests. This gives me a stable profile at 4.5ghz. Temperatures peaked at 83°C Max, but never reached 80°C. I plan to try again at 4.8ghz now that I better understand prime. I'll also check the version I'm using to see if it affects temperatures, assuming the latest one includes AVX support.
Be honest with me: I don’t think it’s possible to achieve 4,8Ghz at a reasonable temperature.
Increasing from 4,5Ghz to 4,8Ghz will consume even more power, and you might need about 1.26V to 1.3V for Prime95 to run smoothly.
With that level of voltage, I expect you’ll hit the 90s or even higher.
You should be grateful to Intel for this issue. Their TIM between IHS and DIE is really poor...