Re: [Sheflug] Re: Open Office

["Alex Hudson" <nospam [at] alexhudson.com>]

>  Still, as Alex says, the Mhz/GHz rating of a processor isn't a great
> measurement - IIRC, we've got Intel to blame for that one - iMarks or
> whatever they're called and rating everything by Mhz.
>
>  Good old MIPS (I guess it's more like Billions of IPS now)..that's what
> you need!
>
>  I mean, just how do you compare an UltraSparc III to an Itanium, or
> Athlon, or a ZX80?

Mmmm. MHz is a good measurement of the core clock speed, that's about all..
for example, it's only really going to tell you the speed of compute-bound
processes. And, that assumes your L1 cache is working at the same core speed
(not necessarily true). And, with a register poor arch (x86 anyone?) you
have more stalls waiting for memory, more potential cache misses, etc. That
shows the processor down. As will a poor compiler (MMX at same time as
flops? Nope...).. the difference between a good compiler and a bad one, or
bad switches, or, or, or, or, can be as much as 20%, maybe even more for
something really dunder-headed / modern (C++ STL).

I could go one, but I'm bored now ;)

Alex.

---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word 
 "unsubscribe" in the body of the message. 

  GNU the choice of a complete generation.