Is this overclocking issue due to bad luck or my setup?
Is this overclocking issue due to bad luck or my setup?
Hi, I recently installed a Ryzen 5 1600 processor paired with 16GB Corsair LPX 3200MHz RAM and an ASUS PRIME B350-PLUS motherboard.
My observations are:
- D.O.C.P 2933 MHz 16-18-18-36 RAM
- 3.6 GHz offset plus 0.04250 and... 85C (or more, too afraid to test) under AIDA64...
The RAM performance is solid; I didn’t expect it to reach nearly 3000 MHz from the Hynix chip, but overclocking the CPU turned out to be a complete failure.
On Auto parameters in AIDA64, the maximum stable speed is 70 C under load—which is quite low.
Maybe I made a mistake because achieving such high speeds isn’t feasible?
Everyone I spoke to reported around 67C under load (with stock cooler) at 3.7 GHz and 1.3+ V... how is that possible?
PS: I also have a Phanteks P400S with the original fans.
Is that included in the most recent bios? The new versions seem to offer better stability and might support a slightly lower voltage. That's a solid question. I plan to check tomorrow and see if an update is possible.
Also consider a better heatsink. After using my old EVO 212 for about eight years, I've reached a stable max OC of 3.95GHz @ 1.46V on my Ryzen 1400. I rarely exceed 170°F under hard settings, even with the stock thermal paste. (I'm planning to switch to a 105CFM model and use Arctic MX4 thermal paste next.)
Elbert is checking if the latest BIOS includes improvements.
JWoody shares his experience with a stable max frequency of 3.95GHz at 1.46V on his Ryzen 1400, noting he stays under 170°F even with high settings and using stock thermal paste. He mentions switching to a new fan and Arctic MX4 paste.
He updated the BIOS via EZ Flash 3 and reports better performance.
He set the CPU frequency to 3.6 GHz at +0.62500 V, expecting up to 3.7 GHz, but observed readings of 83°C (HWMonitor), 76°C (AI Suite) and 85°C (HWMonitor).
He questions HWMonitor's accuracy, pointing out discrepancies with AIDA64 stress test results.
Consider trying a different speed fan, though those two might just be minor variations. The second attempt likely benefited from cached data to improve accuracy. Essentially, it optimized its approach for the second run.