Overclocking i5 6600k
Overclocking i5 6600k
Hello, I wanted to know the limits of pushing the vcore during overclocking. The chip has a maximum of 1.52 according to Intel, but I’m curious if 1.45v is too high for 4.8Ghz on four cores. When it shows a blue screen, is that due to overvoltage or undervoltage? And when it freezes, is it because the clock speed is too high or the voltage is too low? I own a Cooler Master EVO v2 and it’s running at 50 degrees under load at 1.4v with 4.6Ghz (I’m still learning overclocking).
Freezes and blue screens might stem from either a high clock speed or insufficient voltage. Both factors can lead to instability, making it possible for either one to be responsible. In some instances, excessive voltage can also trigger instability, though thermal problems are more common with too high voltage before stability issues appear—unless you're using an exceptional custom cooling solution, which is rare.
Here’s a helpful resource for overclocking on Skylake:
https://www.tweaktown.com/guides/74...e-...index.html
For those new to overclocking in general, my introductory guide is worth reviewing before attempting more advanced settings. Please refer to both resources while adjusting your configuration.
CPU overclocking guide and tutorial for beginners
Explaining the fundamentals. If you’re here, chances are you’re seeking assistance with basic overclocking. Or perhaps you’re an experienced overclocker curious about fresh perspectives. Be confident—I haven’t missed anything. This is just meant to...
50C under heavy load at all-core 4.6 GHz? Those readings look reasonable, but I’m still doubtful... (I should check actual temperatures using CPU-Z/bench/stress while running for 10 minutes).
Unfortunately, with four cores and four threads, reaching even the final 100 MHz might not really boost gaming performance much.
Eventually, around 4.6 to 4.7 GHz, temperatures will climb rapidly with each 100 MHz increase, particularly if a small voltage adjustment is required... (My 7700K went from 72°C at load to 85°C when switching to 4.8 GHz, so I opted for the lower speed.)
This idea works well for any overclock attempt. If you need to keep the hardware for a while, identify the limiting factor, adjust your settings, and reduce the multiplier by 100mhz. This approach helps because it minimizes electromigration and VT shift at lower frequencies, allowing you to slightly lower voltage while maintaining stability. It also improves thermal response and reduces electromigration.
Skylake
All Core SSE Frequency
All Core AVX2 Frequency
BIOS Vcore
Capable
6600K
4.50GHz
Not Tested
1.376V
100%
6600K
4.60GHz
Not Tested
1.392V
Top 91%
6600K
4.70GHz
Not Tested
1.408V
Top 70%
6600K
4.80GHz
Not Tested
1.424V
Top 38%
6600K
4.90GHz
Not Tested
1.440V
Top 14%
6600K
5.00GHz
Not Tested
1.456V
Top 2%
https://siliconlottery.com/pages/statistics
During testing, execute a stress test with full CPU utilization.
Aida64 offers little value. It merely reinforces the notion that everything is under control, which is only partially true. Compared to tools like Prime95, Heavyload, and others, it falls short in providing a reliable indication of performance. As explained by Computronix, our expert on Intel CPUs, this approach can create a misleading impression due to the unusually low temperatures observed.
To obtain an accurate assessment, refer to the guidelines I previously shared in my overclocking tutorial. As Comp mentioned, it's essential to use software designed for steady-state workloads when evaluating CPU thermal behavior.
I'm familiar with five such tools. Ranked by load intensity:
(1) Prime95 - Utilizes small FFTs (AVX disabled or AVX offset set in BIOS)
(2) HeavyLoad - Simulates stress on the CPU
(3) FurMark - CPU Burner utility
(4) Intel Processor Diagnostic Tool - CPU Load utility
(5) AIDA64 - Contains a System Stability Test with Stress CPU feature
AIDA64's Stress CPU test cannot load overclocked or overvolted CPUs, making it ineffective beyond giving users a false sense of security. HeavyLoad is the most viable alternative, with temperatures and power consumption staying within 3% of Prime95's results.
Here’s a concise summary of my testing method: configure BIOS settings carefully, then boot into Windows and install Prime95 alongside HWinfo or CoreTemp. Run the Small FFT test with AVX/AVX2 disabled, monitoring core temperatures closely. Aim for 80°C max on most Intel CPUs (85°C limit), and under 80°C for Ryzen models.
If the thermal limits are met, proceed to stability checks using Realbench or a similar tool. For extended reliability, consider running Prime95 in Blend mode or Small FFT for 12–24 hours. Avoid running HWinfo or CoreTemp during this process.
Any instability detected—whether from stress tests or real-time monitoring—should prompt adjustments to BIOS settings or voltage levels before proceeding.
I worked with my 3770k to reach 5.0GHz at 1.42v, then lowered it to 4.9GHz at 1.328v. The test temperatures decreased from 88°C to 72°C. I didn’t notice any noticeable changes in frame rates, and the device remained unchanged for six years straight. Eventually, costs become more than worth it, especially when the advantages are minimal.
There are two main types of tests: stress and temperature. Choose the one that fits your goal and use the appropriate tool. All-in-one tests aren’t ideal since they don’t target specific issues. You can’t reliably measure temperatures with a stress test that forces the CPU under varied loads, nor can you guarantee max temps from a steady-state test.
A solid OC should provide stable voltages, manageable temperatures, and consistent performance, no matter the clock speed. A poor OC will fail in any one category—voltages, temperatures, or stability—based solely on desired speeds.