[X4U] New Mac Mini's

[Zane H. Healy]

> Apple seems to have very carefully not included numbers for the Core  
> Solo on their web page,  but I'd say cut the Dual number in half  
> (since it's a SPECrate number) and take a little off and you should  
> have a rough estimate.
> 
> Note that SPEC numbers grossly exaggerate the performance improvement  
> of the Intel chips over the G4/G5 that you're going to see with real  
> life apps. Apple says that the Dual Core Mini is about twice as fast  
> as the 1.42 GHz G4 Minis, but the Solo core may be considerably  
> slower, depending on how multiprocessor sensitive the app is.

I'd noticed the lack of info on the Core Solo on their webpages.  The
inclusion of SPEC numbers was interesting, I believe this might actually be
a first for Apple.  Though I see that in their classic lack of hard 
information on performance they fail to give the non-rate numbers.  Apple
has *never* submitted performance information to SPEC as far as I know, I
just checked and they most certainly have never submitted CPU2000 numbers.

Going dual cpu doesn't actually mean a doubling of speed, but then the Core
Duo is 1.66Ghz, while the Core Solo is 1.5Ghz.  So guessing that it's about
half the speed for a multi-threaded, CPU intensive application might be
correct.

> Reports I've seen put the Intel MacBook Pro/iMac at only 10-20  
> something percent faster than G5s.

Hmmm, then the Core Duo should be more than enough for a little HTPC minus
the tuner.  Not sure about the Core Solo, without hard numbers on it.

> On the other hand the Intel base machines seem to be much better at  
> decoding HD content than the G5s. I haven't seen an explanation for  
> this yet, it may be due to the graphics chip they use.

I wonder if this is simular to how VNC sessions are painful when displayed
on a G5, but just as good as being on the actual system on a much slower x86
system.  Same with MS RDC sessions.

		Zane