Hyperspeed wrote:
As for slow Transmeta Crusoe, I saw a 1GHz model and Sony obviously approved of it for it to be in the flagship Vaio laptops!
IIRC the crusoe was only used in Sony's ultraportables... Anyway, the crusoe was never really about performance, it was about power consumption. It could never really compete with even the slowest x86 mobile chips from AMD or Intel in terms of performance. Now ARM have won the power consumption race.
And wasn't it well reported that the Dual-G5 Macs were outperforming even the highest rated Pentium machines?
Steve Jobs reality distortion field taking effect there... Perfromace per $ of the Dual-G5 Mac was much worse than x86 machines*. Though you do get multiprocessing, MacOS X and a beautiful case with it...
*Note: I still want a Dual G5 Mac with Tiger :-D
Apple ran an advert proclaiming the PowerPC G5 Mac to be the most powerful computer in the world!
Apple had to remove the claim in the UK, as after an investigation by the "Avertising Standards Agency", Apple were unable to prove their claim.
Isn't RISC producing much less heat than a comparable speed CISC chip?
Don't fall into that trap, the terms RISC and CISC haven't really meant much since the late 80's.
The PPC is
not a RISC chip (look at the MIPS chip if you want to see RISC), the PPC is better described as a Load-Store chip with a neat architecture. But the Modern x86 and the PPC both borrow as many features from both RISC and CISC designs as each other (though not necessary the same features ;-)).
It is true that the x86 ISA does incur penalties, like instruction decoding and lack of architectural registers, which does add to the transistor count when trying to correct (one of the reasons why a basic PPC core can be much smaller than an x86 core). But it turns out that the archtiectural improvements made to the x86 to overcome the limitations, actaully actually improved execution by several factors... notable out of order execution and branch prediction, which is a vital speed boots to current single core CPUs.