The strange case of the AMD FX

Hi. This is a strange situation. Excuse the english please. I noticed a bug in my cpu (amd fx 9370) where the C1 idle state was unused, and it would skip to C2 (lose cache data at C2). So.. after much tweaking my 9370 @4.35 no turbo, beats a 9590 in cpu-z and corona. Cpu-z result https://valid.x86.fr/3vjzqn .Maybe other people should use perfmon to check C state transitions??.

Take care.

1 Like

How in the world did you get a bus speed of 280! Did you run into any flat spots between 200 and 280, like around 230?

Also gotta download the newer version, guess they changed the benchmark. Now they do multi first then single

Hi Trucker. The bus speed seems fine up to 300, anymore than that the ram has a fit.
NB and HT voltages need a small bump, CPU/NB needs at least 1.35v if you push it near 2500-2600 at high bclk.
vrm switching frequency @400, HPC enabled in bios.

CPU-Z 1.79.1

New test results. [email protected] 200x22 1.35v no turbo. 8gb 1866 8-9-9-24-32 Sabertooth 990fx r2.0.

1 Like

I have the 8320e on an Asrock 970m pro3 which was dirt cheap with the cpu.
Weak power delivery normal with a budget board.
I turned C6 off, turned up the base clock slowly till my games failed then turned C6 back on and ended up with:

Highest vcore was 1.488 and the limit is 1.515 but if I tried that on all 8 cores with 4 "no name" VRM's and a tiny heatsink my mb would not survive :slight_smile:

Drives me nut's that cpu-z changed the benchmark numbers