Is that difference because you're testing it in 64 bit mode but looking at the 32 bit mode scores on geekbench?
No, I tested my Mini in 64-bit mode.
I assume the difference is because I have the Fusion Drive. The main GeekBench score is a composite of CPU, graphics, and disk performance. I suppose we really ought to be looking at CPU scores here.
It did test higher once I added the CL9 RAM. It didn't affect the main score that much, but some of the memory scores bumped-up by as much as 30%
Also, I'd love to see a mini working at 80% and up cpu usage for more than 10 minutes and *not* go on fire or slow down.
I'd been skeptical of that myself, but was reassured by posters here that it's not the case.
That may be true for older Minis, but not for the 2012 model. One caveat is that I have 1.35V RAM installed, so should run cooler than the standard RAM. (But I also have 16GB.)
I ran the GeekBench Stress Test for 8 hours, and confirmed that it pegs the CPU at 99+%. GeekBench does a simple score while it is doing the Stress test, and this remained in a very tight range, with insignificant variance. And then I re-ran the full scoring, with no change in values. After 8 hours, it was warm to the touch, with the hottest part the back left corner at the top. It wasn't near as warm as my iPad 3 gets.
Were this the case, I don't think it would get the GeekBench scores that it gets. GeekBench takes long enough to heat the Mini up to it's maximum temperature by the end of the test. I assume that part of the strategy of GeekBench is to run long enough to heat up the CPU to max. I don't think there's much chance manufacturers can game this, because if you can keep it cool for 5 minutes straight-out, you can keep it cool for 24 hours straight-out. 5 mintues is about the time to heat it up to the max, and then takes about 1 minute to cool down.
At idle, (not that useful, I know) the Mini runs the CPU at 117F. My 2008 Aluminum Macbook is running 147F. Think it needs a good cleaning.
I don't think TurboBoost comes into play, since as I understand it, that only kicks in for very brief bursts when the CPU has been otherwise idle.
Might the Mini throttle the CPU under heavy load? Perhaps, but if so, it still runs at quite a nice clip, enough to give those scores. One could experiment by providing more cooling or putting it in a constrained environment to let it heat up more. But I assume Apple knows how much heat it generates at full load in a typical environment, and designed just enough heat removal to be able to run at full load. I'm satisfied that it will handle my 5-minute compiles just fine.
Might be a different story in the summer.
This isn't a notebook, and they could have well-afforded to make the case as big as necessary. Twice as tall, and they could have put in a monstor fan. I'm sure, though, that there were some "interesting" discussions between Jonathan Ive and some mechanical engineers.
