From: Eugene Leitl [mailto:eugene.leitl@lrz.uni-muenchen.de]
>
> One should compare apples with apples. TI 'c6x MIPSen look good on
> paper, but do not translate very well in real-time performance.
I suspect as much (given that it's a DSP) but don't know any particulars of
the chip.
> I don't see anything doubling but the amount of transistors you can
> put on a die. This doesn't translate into anything too cool, as long
> as you get special cases piled on top of each other (pipeline setup,
> burst, keeping ALU/FPU filled with the right magic instruction
> mix) for it to happen. Bah, humbug.
Yes, this is why for the new Intel / AMD / IBM chips, clock speed seems to
be the only good indicator of actual performance in mainline tasks.* Since
they seem to have maxed out the number of functional units and improvements
are trending towards providing specialized multi-media instructions, we're
no longer getting general purpose clock-independent peformance boosts from
the new architectures.
Hopefully VLIW will start to address this in the next few years.
* = note that I'm talking about chip performance here, rather than system
performance. For system performance, these days a fast bus, big cache, and
lots of RAM seem to contribute more for most tasks than processor speed.
regards,
mez
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:03:40 MDT