5900X drosselung sowie TDC/EDC-Messwerte
5900X drosselung sowie TDC/EDC-Messwerte
I adjusted the settings to -28 atm core and it worked well for a while under different loads. I realized the power cap has dropped from 140W to around 130W during full load. Setting TDC/EDC to stock values (90A/140A) seemed right, but after tweaking them up by about 2:3, I reached the 140W limit again. Adding roughly 10W helped boost performance slightly, though it pushed current to -2 on all cores—close to -26 or caused a crash. I’m currently at 100A/155A atm. Have anyone tried similar adjustments? Increasing the current limit and seeing extra stability?
I had a 5900x that performed well at -30 on slower cores, -20 on average ones and -15 on the best ones. That allowed me to reach 220W PPT, though the last 30W didn’t significantly boost performance, so I ran at 180W. I don’t recall the TDC/EDC values I used.
Consider adjusting the value per core. I didn’t realize I was using Linux and a recent kernel update caused a crash similar to Windows. It looks like moving four points higher in BIOS (all cores affected) would help align with Windows voltages. Raising current limits also seems to resolve the issue. I’ve been running smoothly for 24 hours with no crashes after testing.
You're pushing the CPU to its limits with very low power settings. They prefer higher extremes. My 5900X handled 260/170/200AC with +200MHz and delivered around 523GFlops in Linpack Extreme, covering the full 260W.
Your MSI B550-A Pro is indeed in the A-tier VRM range at 200A. The specs you mentioned—260W PPT, 170A TDC, and 200A EDC—seem correct. To add a 200MHz offset in BIOS, check the power management settings for your motherboard. For low power states, ensure your system is set to use those modes during general browsing or streaming. Your Dark Rock Pro 3 supports up to 250W, so it should handle the load comfortably.
I took the opposite route—130W/90A/110A. Under full load everything stayed near 100% and temperatures hovered around 67°C. With -28°C cooling, the performance dropped about 5-6% compared to higher loads like 180W/140A/190A. The temps were much higher, making it not worth the effort for me.
I receive about 20500 atm yet stay under 70°C. Conducted numerous tests but no issues. Used OCCT and Prime95 without errors. Still, the OS randomly restarts when temperatures drop to -30 and -29 across all cores. Unclear how to identify which cores are failing. Atm seems normal around -28.