Initial Inquiry About Overclocking
Initial Inquiry About Overclocking
Hello, I assembled my initial PC in March using these parts:
Processor: I5 6600K
Cooler: Noctua NH-D15
Motherboard: Asus Maximus VIII Hero
RAM: Corsair Vengeance LPX 2x8gb DDR4 3000 MHZ
Storage: SSD Samsung 850 EVO 500 GB
Hard Drive: Seagate 2 TB Hybrid Drive 7200RPM
Graphics: MSI GTX 980Ti
Case: Fractal Design Define R5
Power Supply: EVGA Supernova 750w Platinum Efficiency
I haven’t overclocked before, but I’m planning to try now and have a few queries. I’ve already set up my XMP profile for the right memory speed; I’ll adjust the multiplier and raise the voltage accordingly.
Regarding stability tests, what suggestions do you have? I’ve come across mixed advice about Prime95’s AVX usage—does that really matter? Should I stick to one tool or combine them (Prime95, Linx, X264, XTU, etc.)?
Thank you in advance for your assistance!
x264 works well. Prime95 26.6 and Aida64 are also fine. You can use just one or experiment with a few.
The main reason no one suggests Prime 28.9 is that it demands an impractical amount of resources on your system. It's not harmful, but I consistently run 28.9 on my machine (stock clocks) for GIMPS.
I enjoy using OCCT for stress testing and monitoring system metrics.
I opt for the simple approach, keeping all settings in auto mode and adjusting only the multiplier.
Just keep an eye on the vCore to ensure it stays within safe limits.
I use OCCT which seems to offer a better collection of instructions.
It will halt the test at 85c.
First, adjust the RAM to 1.2v, which corresponds to 2400 speed.
Increasing XMP speeds can raise the RAM voltage and affect your OC.
Higher RAM speeds don’t necessarily improve actual performance.
http://www.silentpcreview.com/article1478-page1.html
Keep the voltage on auto, then slowly increase the maximum multiplier.
You’ll hit a Vcore limit (around 1.4v) before reaching the maximum temperature.
Track Vcore using cpu-z.
Here’s what you can anticipate:
As of 5/2016
What percentage might succeed with an overclock at a reasonable 1.40v Vcore?
I5-6600K
5.0 2%
4.9 11%
4.8 36%
4.7 64%
4.6 88%
Also, use speedstep with adaptive voltage.
This will lower the multiplier and Vcore when the CPU is idle.
If you’re seeking a safe and practical approach, it’s best to stop here.
You might achieve more by studying all the details of Skylake overclocking.
I use OCCT, which I believe offers a more accurate set of instructions. It will halt the test at 85c. First, adjust the RAM to 1.2v, which corresponds to a speed of 2400. Higher XMP speeds can raise the RAM voltage and affect your OC. Faster RAM provides minimal real performance gains.
[Link](http://www.silentpcreview.com/article1478-page1.html)
Keep the voltage on auto, slowly increase the maximum multiplier. You’ll hit a Vcore limit (around 1.4v) before reaching the maximum temperature. Track Vcore with cpu-z.
What you can anticipate:
As of May 2016
The percentage that might succeed with an overclock around a reasonable 1.40v Vcore:
I5-6600K
5.0 – 2%
4.9 – 11%
4.8 – 36%
4.7 – 64%
4.6 – 88%
Consider using speedstep with adaptive voltage, which will lower the multiplier and Vcore when the CPU is idle.
If you’re seeking a safe approach, stop here. You might achieve more by studying all the nuances of Skylake overclocking.
Do OCCT alone suffice? What settings should you apply in OCCT for testing your overclock?
I experimented with OCCT, achieving an OC of 4.5 GHz and 1.28 V. Running large data sets caused peak temperatures near 58°C. Could I push higher? What maximum temperatures should I aim for?
I also adjusted the CPU core cache limit to 255.5 to prevent throttling, and set the load line calibration to level 5 (level 8 is the upper limit).
Should you keep voltage on auto or manually tune it and test different voltages? Also, does disabling speedstep pose any long-term risks for chip health?