I5 6600k 4.6ghz at 1.28v
I5 6600k 4.6ghz at 1.28v
I just achieved a stable overclock and tested it with Aida64 for two hours, Prime95 for fifteen minutes, and Asus realbench for fifteen minutes. Temperatures stayed around 60°C. Only Prime95 reached 70°C. Should I try pushing further? Can I rely on Aida and realbench enough? Because everyone suggests P95 but it causes high temperatures. I use a Kraken X52 with fans set to 55-60%.
Poisonarrow :
When should I claim I have a stable overclock? What tests do I run and how long?
I typically perform short 20-minute prime95 checks to determine the maximum. If a final round runs without errors overnight, I’d consider myself stable, then proceed to my RAM tests.
Your temperatures seem fine; you might be able to go higher. GL
Prime95 is often criticized online, but many people recommend it for stress testing and checking cooling efficiency. Running it at full speed can reveal maximum temperatures. A 4.6ghz run is a solid overclock with a 6600k at that setting, potentially keeping the CPU fan quiet. At around 1.35v, you might reach 4.8ghz or more. If your graphics card is a GTX 1070 or 1080, consider pushing it slightly higher, as small CPU gains can noticeably boost frame rates in certain games.
It seems you're dealing with some thermal paste issues under the IHS, but it's not ideal for everyone. My 3770k is running smoothly at 4.6ghz with 1.425v and my Corsair H60, reaching only around 75°C on the hottest part. I find it surprising how much less cooling these CPUs need compared to soldered components. My previous 2500k model could handle 5ghz at 1.42v, maxing out at 70°C with a cooler like the Arctic Freezer 13 92mm.
Poisonarrow :
Yes, I can clarify that.
A small FFT in prime95 version 26.6 is effective for checking core and cache stability without generating excessive heat, unlike larger FFTs that also test RAM. It’s mainly for assessing core stability rather than evaluating cooling solutions. Using the 26.6 version helps bypass issues with AVX instructions, which can limit higher clock speeds. With AVX, you’re aiming for a more stable overclock under less realistic conditions.
When can I claim I have a stable overclock? What tests should I perform and for how long?
Poisonarrow :
When should I claim I have a stable overclock? What tests do I run and how long?
I typically perform short 20-minute prime95 checks to determine the maximum. If a final round runs without errors overnight, I’d consider myself stable and proceed to my RAM tests.
Your temperatures seem fine; you might be able to go higher. GL