Ryzen temperature and voltage data
Ryzen temperature and voltage data
I've reviewed many discussions about ideal temperatures and voltages for Ryzen processors, so I'm not really seeking that information. When I began experimenting with overclocking my FX-6300, others were quite cautious about exceeding 1.4V and maintaining a temperature near 55°C. I followed their advice. Now, some people claim 1.55V / 80°C is stable—something that seems unusual to me! I'm just curious about Ryzen's performance when pushed to the very high end. From what I understand, Ryzen tends to be quite sensitive to heat. If it gets too close to 70°C or higher, it’s likely to fail, no matter the frequency, workload, or voltage. I don’t think it will ever reach the 80s or the 90s where protection throttling is seen as unrealistic. Regarding voltages, I’m aware AMD suggests a limit around 1.45V, but that’s mostly theoretical and based on their models. How does this actually play out in real life?
We still have uncertainty about this. Long-term testing would be necessary to understand which voltages and temperatures affect a Ryzen CPU over time. A voltage of 1.55V might not immediately damage a Ryzen CPU if cooling is maintained, but the long-term effects are unclear. I wouldn't suggest using 1.55V for a gaming PC; it's better to stick with 1.45V or even lower.
We're still figuring things out. Long-term testing is necessary to understand how voltages and temperatures affect a Ryzen CPU. A voltage of 1.55V likely won't completely destroy a Ryzen CPU if we maintain proper cooling; however, I wouldn't suggest it for high-performance builds like gaming PCs. It's better to stick with 1.45V or lower.
My board's voltage becomes an issue mainly at low loads, as the LLC output is quite limited. At around 3.9GHz (around 1700), it drops to about 1.31V. It remains stable at 3.95V but falls between 1.44-1.45V during light usage and 1.36V under full load. The temperatures are acceptable, ranging from the high 60s to low 70s on a 240mm AIO. The main concern is the low-voltage drop in Prime95, which can cause significant temperature spikes.
Playing games only occasionally brings one core close to full load, and even then, performance isn't as strong as at P95. Temperatures rarely exceed 50°C, especially with well-optimized code. Encoding also stays within safe limits, rarely reaching the 60s.
The issue is primarily the low-voltage behavior, so I'm checking if it's safe to allow voltages above 1.45V in the long term. Also, I'm using STI2 - TFN core voltage settings, and the actual core voltage reported by telemetry matches what the VRM output shows. This suggests the LLC solution isn't very effective, which is why it's less reliable.
If I reach 4.0GHz, I might need to increase the voltage to 1.5V, but that would be a difficult decision. Additionally, VRM temperatures during P95 would likely be too high, causing throttling even with a fan on the heatsink. For regular use and gaming, it seems manageable.