Question Overclocking Ryzen 3700x
Question Overclocking Ryzen 3700x
Hello everyone. I believe I obtained a really good chip, but I encountered a bit of a hurdle. When attempting to overclock my CPU even by .25 over the 43 ratio and then rebooting to run Cinebench, I frequently crash or force a restart back into the BIOS, indicating a CPU overheat warning. This seems unusual since 43 is considered very stable according to HWiNFO64, which reports me in the 80-85°C range during crashes.
I’m struggling to identify what’s causing the overheating—so I thought I’d come here. To be honest, I’ve been tweaking the voltage a bit to see if it can stabilize enough for a 43.25 or higher setting, sometimes reaching up to 1.4V and getting stable. However, around 1.36V it seems to work for power but still triggers that overheating crash. I suspect the chip itself is fine, but there might be an issue with the motherboard. I’m using an Asus ROG Crosshair VIII Hero, paired with a 1:1 1900 memory and fclock. Maybe this is typical for everyone? I just wanted to ask if this was normal or if there were other suggestions to try.
Additionally, I notice some unusual sounds from the start, even before overclocking. I wonder if this is common on Asus boards and whether a fix exists for it.
Here’s what I’ve managed to achieve at stable 43 with Cinebench runs.
From what I've understood, maintaining temperature below 75°C under load when pushing hard is essential. Therefore, 4.3G appears to be the optimal setting for a 3700x manual overclock. Some enthusiasts with excellent silicon and cooling solutions achieve higher numbers, but I believe 4.3G is sufficient for most users.
Running at 4.3G mine rate yields Cinebench 20 and 'desktop' stability throughout the day, including Realbench. However, increasing the frequency to 25Mhz raises the required voltage significantly—around 1.45 to 1.47V—and the temperature becomes difficult to manage with my 240mm AIO, causing it to crash quickly.
I don’t see much advantage in choosing 4.3 all-core over simply enabling PBO and setting scalar at 10X. The latter extends runtime slightly, improving ST scores by 5 to 10 points compared to 4.3G manual, while MT scores drop, making noticeable gains hard to detect, especially at 100 CB20.
Alright, this seems a bit odd to me. I thought a 0.25 to 1 would boost heat output, but still within spec and cooling. Even with a slight voltage increase, it still feels unrealistic to think overheating could be a problem based on the temperature readings. I was planning to get stable power around 44 at 1.36-1.4V, which would allow running low CPU tasks. But when I run benchmarks, it crashes by telling the BIOS the CPU is overheated. The monitor isn't very clear, but I guess something happens during the switch from 43 to 44 that triggers this issue.