Re: How many times must Apple make such DRASTIC architecture changes?

[Tim Cutts <timc@chiark.greenend.org.uk>]

In article <d8d64h$j41$1@panix3.panix.com>,
Brian Reynolds <reynolds@panix.com> wrote:
>I had forgotten about the Alphas.  For all the talk about how good a
>processor Alpha was, I never saw any of them out in the wild, only at
>trade shows.

They are fabulous.  It's a great shame they've been shot in the head.
The Sanger Institute (http://www.sanger.ac.uk/) where I work got into
Alpha very early on; they realised that the human genome is too big to
handle as a single lump on 32-bit CPUs; they would have to go 64-bit as
early as possible, so they did.

The Alphas are very fast indeed, particularly at some codes.  We use a
program called genewise (http://www.ebi.ac.uk/Wise2/) for gene structure
determination.  It still runs faster on a three-year-old 1GHz Alpha EV67
processor than on anything currently on the market - and I've tried it
on AMD Opteron, Intel Xeon EM64T, Intel Itanium2, PowerPC G5 and IBM
POWER4+.  Itanium2 and POWER4+ both came close, but didn't beat it, for
that particular program.

Incidentally, the program's author develops it on Mac OS X
these days.  :-)

Tim