🚀 go-pugleaf

In article <4dmcaiINN459@keats.ugrad.cs.ubc.ca>, c2a192@ugrad.cs.ubc.ca
(Kazimir Kylheku) wrote:

>In article <jragosta-1601961045420001@ppp-1001.dca.net>,
>Joe Ragosta <jragosta@dca.net> wrote:
>>In article <4de92u$ded@baygull.rtd.com>, Lawson English <kensal@rtd.com>
wrote:
>>
>>> However, the debate that has been raging is:
>>> 
>>> WHY are the Pentium Pro SPECint/fpXX scores so high in comparison to the 
>>> PowerPC and other CPU's, when the BYTEMark scores are so low comparing 
>>> the same systems?
>>> 
>>> 
>>> Simply referring us to the half of the figures that happen to favor 
>>> Intel doesn't resolve the issue. We KNOW that Intel beats the pants off 
>>> of PowerPC's and many other CPU's on SPEC. 
>>> 
>>> But why is there such a HUGE difference between the SPEC and the BYTE?
>>
>>Or on applications performance. WinNT on a 604 is much faster than WinNT
>>on a P6 (according to Byte Magazine).
>
>Is that a 604-based Mac, or some other box?

No. WinNT does not run of PowerMacs. The above poster was referring to a
PREP-compliant computer that uses a PowerPC 604 processor. IBM, Motorola,
Firepower, and several other manufacturers make them. They run WinNT for
PPC or AIX (IBM's Unix variant).

>In any case, it's the operating sytsem of the PowerMacs that makes them so
>slow, not the CPU architecture. Another reason is that to keep prices down,
>Apple is putting some pretty lame video hardware into these computers. 
>It's the same thing with PeeCee clones, except there is more of an awareness
>about the internals. When I see an advertisement for some PeeCee, I can tell
>right away whether it has a decent video card or whether it's been emasculated
>to keep the cost down. A Mac is just a seamless, integrated Mac.

I assume this and the rest of your post is a troll for flames; you seem to
have no clue about Power Macs or the MacOS. Certainly, MacOS could be
better in many ways (all native, PMT, protected memory), but I would
hardly call PowerMac's 'slow".   

>The poor performance of the PowerMacs is damaging the reputation of the RISC
>architecture on which they are based. 

Yeah, and Windows damages the reputation of the x86 architecture. Yet
Windows currently runs on the vast majority of x86 boxes. Same for MacOS
on PPC.

Just last week a CS grad student was
>complaining how slow the PowerMac is that he has in his office. He uses it for
>image editing (Photoshop). Everything else is done on some run-of-the-mill PC
>running Linux, whose performance is crisp. It's only because the latter has
>system software that manages the CPU, memory and I/O better. Just things like
>demand-paged executables and virtual memory, shared libraries and the dynamic
>buffer cache are what makes the difference, not to mention pre-emptive MT.
>On the PowerMac, it is excessive disk activity and programs that hog the CPU
>more than they should that make it behave sluggishly.

Your ignorance is showing (again, see first question, above).
MacOS has virtual memory and shared libraries. Photoshop uses its own
implementation of virtual memory. Just how many other programs is your
alleged CS grad student trying to multitask while doing image editing
anyway? 

You are aware that Photoshop needs lots of RAM, right?  

PMT would be nice - given that it isn't the case, use your tools properly.  

>
>I'm sure that the actual image processing algorithms of Photoshop would run
>more slowly on that system, but you wouldn't be able to tell, or it just
>wouldn't bother you because the machine would remain reponsive. Unfortunately,
>you can't get Photoshop for Linux just yet, which is why the Mac is there.

Linux seems to be a wonderful alternative OS for x86 boxes. But don't you
hold your breath until Photoshop comes out for it. People who don't pay
for an OS are unlikely to want to pay $500 for a program, either.

>
>The Powermacs are creating, in the minds of some consumers,  an unbalanced
>perception of the capabilities of the POWER PC architecture.
>-- 
In the minds of other consumers, unbalanced perceptions of Powermacs seem
to have already been created.

-- 
ted.oliver@asu.edu