Minimizing PPT without affecting stock EDU and TDU
Minimizing PPT without affecting stock EDU and TDU
I purchased the Ryzen 5900x CPU and noticed its temperatures were too high. To address this, I performed undervolting to lower the PPT, which helped reduce power usage and cooling needs. The results were much better.
The standard PPT was 142, and after adjustment it dropped to 110—this brought the temperature down significantly while keeping the Cienabench 23 score nearly identical to the original.
I conduct stress tests for both single-core and multicore performance, gaming, and heavy workloads over extended periods, and everything remains stable.
I’m not seeking further enhancements; my goal is simply to ensure the PPT reduction doesn’t negatively affect the CPU while maintaining EDC and TDC values as they are.
It's acceptable. The displayed values represent the boundaries; if you stay within them, the boost will keep rising. PBO ECO operates similarly by intentionally lowering the limits to minimize power use. In fact, boosting clocks could negatively impact performance and some stock results if limits are cut too much.
Longevity remains unaffected because excessive current, heat, or voltage are the main issues that damage the chip. Overcurrent leads to dielectric breakdown, overvoltage causes electromigration (possibly misremembered), and heat is another well-known constraint. Your setup indirectly reduces both current and voltage via power restrictions, so you're safe.
Stability isn't expected to be impacted based on what I know.
It's acceptable. The displayed values represent the boundaries; if you stay within them, the boost will keep rising. PBO ECO operates similarly by intentionally lowering these limits to minimize power use. In fact, pushing boost clocks too high may reduce performance and some stock results, depending on how much limits are cut.
Longevity isn't at risk because excessive current, heat, or voltage are the main issues—overcurrent leads to dielectric failure, overvoltage causes electromigration (possibly miscalculated), and heat is a well-known constraint. Your setup indirectly reduces both current and voltage via power caps, so you're safe.
Stability remains unaffected as far as I know; it generally depends on voltage and temperature. The motherboard handles voltage automatically in PBO settings, while temperatures are managed through your chosen cooling method.
TDC refers to the maximum current a board can sustain under thermal limits, while EDC is the peak current capacity.
A higher PPT limit is often adopted to ease power restrictions for demanding multithreaded tasks, especially in processors with many cores like yours. Lower or fixed limits can restrict performance, but results may vary by user, so custom PPT settings could be advantageous if you don’t notice a drop in speed.
TL;DR: no cause for concern. If performance and stability meet expectations, everything should be fine.
Thank you for your feedback. The reason I shared this is precisely about the PBO ECO setting. With PBO ECO they managed to bring the PPT down to 88W (rough estimate), while also lowering EDC and TDC to more favorable levels. I aim not to restrict the maximum current because I want to preserve performance and avoid excessive adjustments to stock parameters (since I trust AMD's approach). However, my goal is to curb power consumption limits.
The undervolting with the PBO Curve Optimizer successfully balanced voltage relative to clock speed, resulting in a stable CPU with reduced temperatures. Nevertheless, it led to the CPU overclocking itself until it hit its power usage cap, and the thermal benefits diminished somewhat.
I also wanted to prevent the CPU from self-overclocking further by tightening the overall power consumption, considering it already runs at higher clock speeds due to the undervolting. I just need to manage its thermal output without adding extra power needs. But when I noticed that ECO mode also reduces EDC and TDC alongside PPT, I felt a bit concerned and wanted to verify whether sticking to their default values (140A and 95A) remains acceptable if I lower the PPT.
Fair question
😀
I already use the Dark Rock Pro 4 as a CPU cooler. While gaming the CPU can hit 85°C (stable) and even 90°C (spikes). This is fine for this processor, but I don’t like it
😀
With undervolting and lowering PPT, I maintain performance in benchmarks while keeping it cooler by 10°C-15°C – that’s a big improvement. And if that CPU runs smoothly at lower voltage, why not cut power usage?
The DRP4 is a great cooler, and I’m skeptical that swapping it for another model could bring such a difference...
I understand how restricting PPT would assist. If your DRP4 can't reach the low 80s, there aren't many alternatives besides a custom loop. It seems to be near the upper limit of air coolers, according to what I know.
True, but I'm only seeing these temperatures during gaming. For instance, in Cinebench I'm stable around 70°C.
I think the GPU heat increases while playing, which might be affecting the DRP4 air cooler's performance.
Perhaps installing more or better case fans could help lower it even more, or switching to an AIO/custom loop, but I doubt it will bring the temps down significantly with the original parts.
If you want to lower GPU temperatures, consider using an AIO. As shown on my profile, I have one and it maintains a stable 47 under Kombustor, even though it's only 120mm. In games, it stays at 41 at maximum. It's great for removing heat close to the CPU too.