AMD FX-8150 Prime95 Fatal Error

[firefiber]

So I was trying to run Prime95 on my FX-8150 after OC'ing it to 4.23GHz. It ran for about an hour, before one of the cores gave a fatal error (round off 0.5, expected 0.4 or lower, something like that). The rest of the cores kept on going. I had run two hours of OCCT (Linpack) before this, and it was stable. No errors.

My vCore is set to 1.3V, dram (no OC) is 1866 set to 1.5V. Could this be an issue with the multiplier set too high, or vCore set too low (or too high )? Any ideas on what I should do next? The temps are pretty stable. Never goes above 50°C (CPU temp) and the core temps hover between 40°C and 45°C.

 

[steddora]

On IBT were the results the same? Basically that's saying there was a "math" error from one of the cores. That's bad! That's instability! If your processor isn't doing math correctly; there's a problem. I'd either start looking into tweaking the voltage or stepping back the CPU a little. Just because you can pass linpack doesn't mean it's stable. Prime95 on blend with all the cores churning gives you stability testing at probably the best percentage you can find. Linpack is more of a test of thermal conditions in reality.