On 2 Nov 2004, at 13:21, chthonic streams wrote:
quoted 19 lines The Apple eMac is $799, including monitor (well, it's all built-in>> > The Apple eMac is $799, including monitor (well, it's all built-in
>> to
>>> the monitor). The Apple iBook starts at only $999. What was this
>>> premium you were referring to again? :)
>>>
>>
>> That's still expensive considering the performance you would get from
>> similiarly priced PC commodity hardware. I consider notebook computers
>> still a premium product in general due to the level of integration.
>
> what do you mean by "performance"? i ask because i had heard that the
> whole question of speed, touted in specs as "1.5Ghz" and so forth, is
> different on each platform. i.e. a 1.5Ghz processor on an
> IBM-compatible is not actually as fast or efficient as the same
> "speed" on a mac. thus PCs (if we must call them that, because apples
> are PCs too you know - wink wink) are always playing the one-upmanship
> game with apple - "our computer is faster, look at the specs!" but as
> any gearheads here who make music should know, specs often blur the
> truth.
Apple have talked about this "The Mhz Myth" on their website, but this
is also what I recall from my CS degree: Mhz isn't a straight
measurement as different processor architecture work in different ways.
Intel Processors (and I assumed AMD?) are CISC : Complex Instruction
Set Computer. These have a wide range of in built processor
instructions which can do a lot of complex things, but they also run
instructions through a pipeline to carry out instructions, which is
comparatively quite long to that used on RISC (Reduced Instruction Set
Computers) such as the Motorola PowerPC chips used in Macs.
In layman terms this difference means that whilst instructions are
piped through quicker on a faster (MHZ) machine when you come to doing
some things the shorted pipeline is more efficient, as the longer
pipeline means that sometimes code which is partially run has to be
abandoned because of logic changes further down the pipeline. The
longer your pipeline becomes the more that increasing the processor
speed will have little effect on actual performance.
However, that isn't always the case. So it turns out that RISC
processors are better for some things, and CISC are better for others.
RISC are better for multi-processor set ups IIRC - which is why a lot
of grids and super computers seem to be RISC orientated - but it's been
a while and I'm welcome for someone to tell me I'm wrong.
Whatever - this does mean that straight Mhz performance comparisons
between different processor types are bunk. You are better to use
MegaFlops or GigaFlops (a measurement of processor instructions
executed per seconds). Across processors of the same chip architecure
(ie i686 vs i686 or Celeron Vs Celeron or PowerPC 'G5' vs PowerPC 'G5')
it can be a useful measurement.
Though this is a bit academic, because no matter what the speed of your
processor, if the Operating System doesn't allow you to take advantage
of that it's not really any use, and if your operating system is
bloated in anyway (like, say, I feel Aqua is on OS X compared to OS 9s
windowing interface, but also so is Windows XP).
---------------------------------------------------------------------
To unsubscribe, e-mail: idm-unsubscribe@hyperreal.org
For additional commands, e-mail: idm-help@hyperreal.org