Tips for beginners on CPU overclocking
Tips for beginners on CPU overclocking
So there you go. I just bought the AOC C24G1 monitor, which is a solid 1080p at 144hz VA. I used to have a 1366 x 768 60hz screen and didn’t really focus on frame rates back then. Now, I definitely want to hit 144fps in most games. Occasionally, my games stutter when using higher resolutions, and frames keep dropping—maybe it’s the GPU or the CPU, but I’m not sure. Here’s what my setup looks like: MOBO:B350 Mortar, RAM (8gb) DDR4 2 x 4 with ballistix sport at 2400mhz, a Zotac GeForce GTX 1060 AMP! Edition 3gb, Ryzen 3 1200, and the monitor is the AOC C24G1. The HDD is a Seagate BarraCuda 2TB SATA 6Gb/s drive with 64MB cache, 3.5-inch size. My power supply is a Seasonic s12ii 520w, certified for 80+ bronze.
I’ve usually managed to keep stable 144hz in games like Fortnite and League of Legends without any problems. But in Apex Legends and other titles, frame rates can drop to 120-100 and even lower.
Now I have a few more questions:
Would overclocking the CPU make a difference at all? If so, what improvements would really help?
Should I upgrade my RAM or aim for a more stable build? Maybe something with higher MHz?
Is it safe to overclock this PSU?
Thanks a ton!
Other users report achieving 4GHz speeds, which is considered a fairly common overclock for that processor. First, install hwmonitor and keep temperatures between 75-80°C. Proceed gradually, using the 95 version 26.6 exclusively—avoid newer versions with AVX instructions that can cause overheating. If your CPU temperature goes above 75°C, halt the testing immediately. When increasing the clock by 1000MHz or 0.1GHz each time, perform a stress test lasting about 30 minutes to detect issues, but allow up to two hours for stability confirmation. There should be no errors at the default speed; if you reach 3.7GHz, you might avoid errors and skip further vCore adjustments. The guidelines suggest a maximum vCore of around 1.4V for Ryzen chips, though some recommend lower values. Numerous guides on overclocking Ryzen exist—proceed cautiously and don’t begin with the highest settings. At certain points (3.8-3.9GHz), you may encounter Prime 95 errors and need to increment vCore slightly until a stable setting is found. Variations exist between chips due to manufacturing differences. Additionally, using slower DDR4 2400 modules could have improved results if paired with higher memory overclocking. For reference, here’s a brief overview of Fortnite benchmark data that might offer useful insights: In summary, avoid rushing to the lowest stable core and focus on careful progression.
Thank you so much, almost everything I required except for everything <3.
You might avoid knowing thermal stability because AVX instructions can expose temperature-related performance variations that aren't visible under standard testing conditions.
Since 99.9% of games rely on minimal or no AVX, AVX-2 or AVX-512, checking for 130+% heat isn't useful. P95 26.6 small fft results are nearly identical to a full continuous load, allowing any cooler—air or AIO—to produce the highest possible CPU heat from gaming tasks. P95 isn’t meant for stability testing; it’s solely for evaluating CPU heat output. Instructions like AVX only increase heat generation, not stability issues. For stability assessments, use Asus RealBench, which stresses the entire system at once across CPU, RAM, USB, PCIe, GPU, etc., to detect problems.
Since AVX-2 and especially AVX-512 are typically found only in professional production and content creation software, average gamers don't need to worry about them on their regular machines. Most new motherboard BIOSes include an AVX offset that can reduce clock speeds by around 100MHz if the CPU decides to use those instructions. Therefore, achieving 5.0GHz during normal gaming or 4.8GHz under high AVX loads is essentially the same outcome. Running a 4.8GHz overclock with an oversized cooler just to test temperatures with AVX enabled often leaves you a few hundred MHz short of the target, rather than achieving noticeable AVX performance in games.
It might seem unexpected, but some individuals go beyond simply playing games with a system. It becomes clear when I can spend an entire day stress testing on Realbench—and it won’t pass Cinebench20, nor will a single run through my own Handbrake encoding test (which involves encoding two movies in about five minutes). I need to understand the operating margin so that as components age and dust builds up, performance remains consistent. Realbench doesn’t provide this information.
Asus RealBench incorporates handbrake into its evaluation process.