The Logic Off Topic list

At 03:37 AM 3/07/02 -0400, you wrote:
>>Intel/AMD are faster on FPU and at least comparable on Integer. The
>
>Who is fastest is constantly changing, but its often Apple.
>
>>Photoshop test is not relevant -- you are looking at specific routines
>>coded for Altivec. 

MMX is old news -- we are talking 5 years out of date -- eg Pentium 166 MMX
circa 1996.

Since then Intel has SSE and SSE2 in the Pentium 4. The SSE2 units are
analogous to Altivec and when things are coded to exploit them they are
accelerated to a great degree -- MPEG encoding, audio etc. Hammer will have
SSE2 also.

Mac users are often carrying around a lot of out of date info about PCs --
the hardware advances rapidly with PCs and the OS is now robust Win2K just
doesn't crash for me. I am running it on a Tosh Laptop and it has NEVER
crashed. Imagine that. What was true about PPC vs x86 years ago is
irrelevant now. You may be informed about these things -- I don't know.

Take all the talk about RISC --reduced instruction set computing. People
still crap on about it. In the next breath they go on to enthuse about
Altivec. Know what that is? -- a shitload of new instructions that run on
execution units that have NOTHING to do with RISC. The instruction set for
the G4 is now enormous -- same for P4. 

>You can inform yourself, that static theoretical tests are not real world 
>tests. The speed with which a computer actually works depends on many factors 
>- thats why they do real world tests - same programs doing the same tasks on 
>the different platforms. In theory, the Brazilian team was not supposed to 
>win the world cup - but they did.

Didn't surprise me at all.