Sam460 V Mac Mini G41.5

[LiveForIt]

It runs MorphOS natively, which is no more "emulation" than OS4.1, in fact they both work pretty much the same way, execute PPC code natively and 68K via a JIT emulation layer.  But as you can see above, the Mini G4 is pretty much on X1000-level performance.

You need a fast G4 to measure up, not just anyone, I don't think there is any G4 dual core MacMini's.

Anyway G5's are more simular to the PA6T.

Mac Mini runs MorphOS if you like that.
Sam460 / AmigaOne-X1000 runs AmigaOS4.1

So if you don't like MorphOS, it should not be too hard to chose, right :-).

[LiveForIt]

Powerbook 5.6 (1.667 GHz)

BENCHMARKs: VC: 7.819s VO: 3.808s A: 0.000s Sys: 0.255s = 11.882s
BENCHMARK%: VC: 65.8037% VO: 32.0497% A: 0.0000% Sys: 2.1466% = 100.0000%
VO: Enabling blanker

Pretty close to X1000.

Not really, because your comparing bananas with apples.

Your comparing ARGB bitmap graphics with YUV420 overlay, two different technology’s.

YUV420p is a 12bit format vs ARGB a 32bit format.

http://en.wikipedia.org/wiki/YUV

ARGB = 640x480*4 = 1 228 800 bytes
YUV420p = 640x480*6/4 = 460 800 bytes

ARGB takes 266% more space then YUV420p, conversion & bandwidth is also costly when you most use ARGB.

So basically the AmigaONE-X1000 has to work harder to display the same graphics.

[LiveForIt]

Sorry, but that's just a (lame) attempt of a academic excuse.
I compare exactly one thing:

How fost does my powerbook replay a certain video and how fast does it an X1000 or Sam460.

Today, reproducible and reliable. Why the result is like it is is not of my interest. I wanna watch that video, not think about the technical internals. And I give a certain brown matter about what could be in future. It matters what's actually there. And for that i provided numbers. They speak for themselves.

But hey, yeah - two more weeks and everything's brilliant of course.

Well, I'm just saying if you like to know how fast PA5T is. FFMEG might be nice benchmark tool, or maybe you can benchmark MP3 Decoding.
Try not to compare things that use overlay with things that do not.