Ryzen voltages are quite a mystery! What do you mean by "@#$%?"
Ryzen voltages are quite a mystery! What do you mean by "@#$%?"
He clarified his point earlier, and what he shared was accurate. Update the BIOS, apply the newest chipset driver, and refrain from manual adjustments. Trying to force AMD settings will actually reduce performance (a mistake I made before). Note: Also steer clear of NZXTs CAM tools.
Manufacturers often misconfigure the standard voltage parameters. Look at early ASUS X99 boards pushing 2 volts onto CPUs and damaging them under stock conditions. Consider all Z390 boards with varying Tjmax, voltage limits, and power capabilities. Or think about how each maker once experimented with different voltage controls and auto-boost features. Reducing the vcore won’t harm the CPU—extreme cases might cause black screens, BIOS resets, and a slight performance boost to compensate.
Intel differs from AMD by offering clear baseline specifications for manufacturers such as Max Safe Voltage, TJMax, and TDP. I purchased my 3800X on launch day and followed various advice about optimizing the Zen 2. Despite these efforts, some guides failed to improve performance, others caused issues, and a few even reduced speed. Experimenting with manual overclocking revealed it often damaged single-core performance, slowing down tasks that rely on one thread. Adjusting settings manually would prevent fast cores from scaling, leading to overall efficiency losses. After AMD's AGESA 1.0.0.4B release along with the chipset driver update, noticeable gains were seen—scene performance jumped significantly, reaching up to 4.5Ghz, and both speed and thermal management improved.
Haha, I actually recalled this and replayed it when @Jurrunio talked about clock stretching. As seen in GN's video, performance fell sharply from the 1.0V vcore while GHz stayed normal. That confused me because my numbers went the opposite way—lower vcore boosted all core speeds and lifted my cinebench scores by 250 points. Yeah, I'm adjusting settings to auto now. I was worried those vcore spikes were too high, especially since Zen 2 is still new and everyone seems to have mixed advice.
The readings in Windows HWinfo show stable idle power at 1.47V, while under full core load it drops to 1.364V. Adding PBO and AutoOC hasn’t changed much. It’s puzzling why some sources say 1.35V is the safe limit, yet others claim higher values are acceptable.
Thanks for your feedback (another 3600 owner). Confirming the ~1.36V under all-core load is accurate. Like you, I'm also dealing with mixed information about Ryzen voltages—so I went ahead and reduced the voltage to be more stable.