Unable to reduce CPU or NB voltage levels.
Unable to reduce CPU or NB voltage levels.
What are you doing? Are you testing at 1600MHz for performance?
For this setup, the SuperPi achieved a 32M score of 9 minutes 50 seconds and 938 milliseconds. The FX-8350 didn't perform as expected, so I adjusted the settings accordingly. Check out the CPU-NB speed I was using. It seems accurate; some later AM3 boards supported AM3+, but I'm not sure if they reached beyond the first generation of Bulldozer chips. No board available to verify this here.
That's the North Bridge (uncore) frequency. That has it's own voltage. On some boards called chipset voltage on others NB voltage while the CPU/NB was the memory controller. Please verify I have that correct. It's been a couple of years. But in other words, I don't know why the CPU/NB voltage would need to be that high to push the NB core. I guess I never experimented with NB frequency, but I killed the IMC that Mac's got from the comp he won, FX9590. That one boinked out at 4.2ghz HT frequency on liquid. I should had done that on LN I suppose. Nice clocks btw. Edited April 23, 2021 by ShrimpBrime Additional Comment:
It's amazing! Usually I'd need to run water to achieve that high stable score, but this one is great!
The failure points were found at the core transistor level. The FX 8 core chips matched the FX II 1090T in transistor count. They were affordable, which made them appealing. Most of the FX chips I bought cost under $100, so I could afford 6 to 8 units alongside a high-end i7. I used them for benchmarking and cooling experiments. A $50 motherboard helped reach 7600mhz performance.
I didn't bother at all honestly. AMD could take my small change. They needed it and deserved it. No one wanted to support them since Intel was faster. Now the situation has shifted. The same individuals who opposed AMD are now using main systems with AMD. Two tough guys.
It might not be worth it. The recovery bios likely won't work with the FX chip. You may need an Athlon or similar just in case.