Associate
- Joined
- 3 Feb 2009
- Posts
- 896
So, having recently bought and built a 9600 based system I'm slightly bemused by having a play with the OC settings.
I'm on an Asus Strix-E board and having failed to get the automated utilities to work (constant crashing/failing to get through the process) I've gone back to good old manual OC'ing.
The strangeness occurs when I manually set the core multiplier (sync'd cores - 50x). It will happily boot to windows with 1.35V adaptive, but when I run a stress test like p95 the clock speed immediately drops to somewhere in the 3.5Ghz region and then bounces around 4.2-4.4.
Temps are all fine, with nothing getting above 50C. The voltage drops quite a bit, down to just over 1V but that's usually when the multiplier drops down to 8-10 so I don't think that's related.
I've watched a few videos and had a bit of a google but I'm out of ideas. Has anyone seen this kind of behaviour before? It feels like there's some sort of artificial thermal throttling going on except it's not taking notice of the actual temps on the chip...
Edit: as soon as the stress test stops, the clock bounces back up to 5.0Ghz before speedstep takes it back down for idling.
I'm on an Asus Strix-E board and having failed to get the automated utilities to work (constant crashing/failing to get through the process) I've gone back to good old manual OC'ing.
The strangeness occurs when I manually set the core multiplier (sync'd cores - 50x). It will happily boot to windows with 1.35V adaptive, but when I run a stress test like p95 the clock speed immediately drops to somewhere in the 3.5Ghz region and then bounces around 4.2-4.4.
Temps are all fine, with nothing getting above 50C. The voltage drops quite a bit, down to just over 1V but that's usually when the multiplier drops down to 8-10 so I don't think that's related.
I've watched a few videos and had a bit of a google but I'm out of ideas. Has anyone seen this kind of behaviour before? It feels like there's some sort of artificial thermal throttling going on except it's not taking notice of the actual temps on the chip...
Edit: as soon as the stress test stops, the clock bounces back up to 5.0Ghz before speedstep takes it back down for idling.
Last edited: