Is this your first time increasing the clock speed?
Is this your first time increasing the clock speed?
This is my initial attempt at overclocking. I own an i5 4690k with a 3.5ghz clock speed that functions adequately for all my tasks, but I wanted to try overclocking since it's the main motivation behind getting the K version.
I'm cautious about risking damage to my CPU, so I'm carefully managing voltage and temperature, which makes me think the voltage might be insufficient for the frequency I've set. I've read that Intel only lists the i5 4690k up to 72.72°C, so I've aimed to keep it below that during stress tests, even though some sources suggest it's probably okay up to 80°C.
My CPU's default voltage is 1.118V, and initially I lowered it to 4.0GHz at 1.1V (below stock). This performed well in Intel Burn Test, reaching a maximum of 69°C with my Hyper 212 EVO cooler. Then I increased the frequency to 4.2GHz at 1.1V, still operating smoothly at around 70°C. Next, I jumped the voltage without changing frequency and set it to 4.5GHz at 1.1V, but the system wouldn't boot, so I reset the BIOS and tried again.
After that, I adjusted the frequency to 4.4GHz at 1.15V, which allowed the computer to boot and complete Intel Burn Test at 71°C, though it flagged instability. I then raised the voltage to 1.16V and rerun Intel Burn Test, achieving a peak of 73°C. I was somewhat pleased with these results. For real-world testing, I played Arma3 King of the Hill for about an hour, achieving approximately 60 FPS in-game with occasional dips, compared to around 45 FPS before. Afterward, I checked HWMonitor and found the highest temperature reached was 56°C.
I also tried Prime95 (though I accidentally used the latest version with AVX, which I've since learned shouldn't be used for stress testing due to extreme temperatures). This caused the system to crash. Eventually, I scaled back to 4.3GHz at 1.15V, passing Intel Burn Test with a maximum of 72°C. I further reduced voltage to 1.13V for an IBT max of 71°C and then to 1.12V for 70°C, achieving full stability.
In summary, my findings are:
cpu: i5 4690k, stock voltage 1.118V
cooler: Hyper 212 EVO
ambient temperature: 20.5°C
temperatures recorded: max from Intel Burn Test (20 trials, standard stress)
ghz | volt | temp | boot | stable | fan speed
Burn temperatures don't accurately represent real-world performance unless you push your PC to its limits. Gaming temperatures won't match what you see during a stress test, and some tests like Prime95 are the worst-case scenarios. I rely on PovRay (10 minutes) and OCCT (10 minutes) as stress tests that are sufficient to detect memory or heat problems; going beyond is unnecessary. This information comes from Anatech's previous test, which hints at the need for higher voltage. Using an Evo with a stock fan adds roughly 8°C to each result. The top cooling performance you get depends on pairing it with an NF-A14 iPPC-3000 PWM cooler, priced at $27—it's expensive but effective.
Is your Evo running in Push Pull? If not, add another fan—this makes a big difference because you can run slower fan speeds. You can safely go up to 1.3V, so don’t hesitate to experiment with the voltage. The "Torture" Test is much more demanding than regular gaming, making it a worst-case scenario.
1.25 - 1.3 vcore should provide stable performance at 4.5-4.6 GHz.
Check out this helpful guide: http://www.overclock.net/forum/5-intel-c...stics.html
Delalo :
Your Evo is not using Push Pull mode. Adding another fan would really help because it lets you run slower speeds without losing performance. You’re safe to go up to 1.3V, so don’t hesitate to experiment with the voltage. The "Torture" Test will put more strain on the system than regular gaming, making it a tough scenario to handle.
A 1.25 - 1.3 vcore should give you stable speeds around 4.5-4.6 GHz. Check out this guide for useful tips: http://www.overclock.net/forum/5-intel-c...stics.html
I’m pretty sure I only have one fan on my Hyper 212 Evo, but I can’t confirm right now. From what I saw, even at 1.2V, just 4.0 GHz would cause temperatures to reach 76°C in the Intel Burn Test. Is it okay to run hotter than the recommended levels for long stress tests? Also, if you hit 4.5 GHz, it would start at 1.17V and drop to 75°C in the Intel Burn Test, but it would crash right away when running Prime95.
Burn temperatures don't accurately represent real-world performance unless you push your PC to its limits. Gaming temperatures won't match what you see during a stress test, and some tests like Prime95 represent the worst-case scenarios. I rely on PovRay (10 minutes) and OCCT (10 minutes) as stress tests that are sufficient to detect memory or heat problems; going beyond them is unnecessary. This advice comes from Anatech's previous testing, which highlights the need for higher voltage. With an Evo and a standard fan, adding about 8°C to each result would improve cooling significantly. The top choice is an NF-A14 iPPC-3000 PWM cooler priced at $27—it’s expensive but highly effective. NF-12s offer only marginal improvement over the stock fan on this setup and are nearly as costly as the unit itself before someone suggests otherwise. Alternatively, the CM Sickleflow 120 provides a slight advantage and tends to perform better overall.