On 8 Aug 2006 at 21:13, Darcy James Argue wrote:

> On 08 Aug 2006, at 6:19 PM, David W. Fenton wrote:
> 
> > On 8 Aug 2006 at 18:01, Darcy James Argue wrote:
> >>
> >> Because looking at MHz alone will not tell you whether a Pentium D
> >> will outperform a Core Duo. . . .
> >
> > CPUs don't mean a damned thing if they're stuck in a motherboard
> > with slow components.
> 
> I don't *really* have to preface all of my comparative statements 
> with "everything else being equal," do I?

But that's my whole point -- "everything else" is *not* equal, so 
searching on CPU won't tell you anything useful.

Perhaps you're accustomed to product lines with far fewer options. 
Any Dell product line is going to have 3-8 CPU choices for that one 
motherboard (a product line = motherboard, generally speaking), so 
where you start is the type of features you want on the motherboard 
(high end, middle-of-the-road, entry-level) and choose a CPU from 
that.

> >> . . . The Core Duos are designed to deliver the
> >> same performance as the Pentium series, but at a much lower clock
> >> speed.  In other words, you might think that a 2.8 GHz Pentium D
> >> would be faster than a 2.16 GHz Intel Core Duo -- but you'd be
> >> very, very wrong.
> >
> > Or not. Depending on the machine it's installed in.
> 
> I guess I do.
> 
> [ahem]
> 
> "ALL OTHER THINGS BEING EQUAL, you might think that a 2.8 GHz Pentium 
> D would be faster than a 2.16 GHz Intel Core Duo -- but you'd be 
> very, very wrong."

But you WILL NOT FIND THE TWO CPUs INSTALLED IN THE SAME MOTHERBOARD.

So, all things are NEVER equal.

> > You have an irrational concentration on CPUs
> 
> It's hardly irrational to point out that the Core Microarchitecture 
> represents a major shift for Intel. They had previous concentrated on 
> clock speed above all else, and companies like Athlon had to try to 
> explain the "megahertz myth" -- higher clock speed is no guarantee of 
> superior real-world performance. With the Core chips, Intel was 
> finally willing to take a (psychological) step backwards in terms of 
> clock speed, in order to obtain better performance-per-cycle.

As as PC user, I DON'T CARE. CPU means *nothing* outside the context 
of a particular PC design. It may have at one time, say when you 
could buy a 32MHz or a 16MHz 386, for instance, but nowadays, the 
difference in terms of clock speed and mips between the fastest and 
slowest available CPUs for a motherboard is more on the order of 33% 
at tbe very most, and the contribution of that difference is vastly 
overshadowed by internal bus speed, memory speed, caching, disk speed 
and video speed. None of those other aspects were proportionally 
nearly as important in the days of 386 chips, where we did have the 
superfast components that we have today.

The balance of performance issues on a Mac may very well be 
different.

But if you're comparing Mac to PC, then all those things *do* matter. 
And CPU is simply not all that important *as a starting point*. 

That was your argument, that the website should be designed around 
the CPU speed (or allow you to navigate it around that concept), and 
that makes absolutely not sense for PC hardware.

-- 
David W. Fenton                    http://dfenton.com
David Fenton Associates       http://dfenton.com/DFA/

_______________________________________________
Finale mailing list
[email protected]
http://lists.shsu.edu/mailman/listinfo/finale

Reply via email to