[BUGS] hardware changing time .. moving forward

[Peter Jeremy]

On Mon, Jan 28, 2008 at 03:27:10PM +1100, Andrew Reilly wrote:
>Naah, the 68k (post '020, anyway) had some fundamental errors
>(just like the VAX) (admittedly, they seemed like good ideas at
>the time) that made progressing to out-of-order super-scalar
>issue close to impossible.

AFAIK, the only architecture where the designers actually considered
what was likely to happen in the future was the DEC Alpha.  Other
architectures mostly just looked at what was possible using the
technology and fads of the day.  (Or just grew by adding warts on
top of warts like the iA32).

>  For all it's warts, the ia32 didn't
>have those same problems, and so can go faster.

The iA32 has the advantage of massive profits which mean they could
afford to make the architecture do things that no sane designer would
consider feasible.

>the result doesn't suck much at all, IMO.  ISA matters almost
>not at all, these days.  Code in C, Java or (better) some
>variety of lisp, scheme or ML and be happy.

HLLs are easier to implement using some than others.  A "better"
ISA will provide better throughput for a given technology level
(clock speed, gate count, dissipation etc).

>together for it really beautifully (if a bit slowly).  I don't
>know that I'd trust FreeBSD to run a laptop properly for me yet,
>though: too much weird hardware in them.

It works OK on my laptop though less well on my son's.  And the
situation with laptops seem to be getting worse over time - vendors
are implementing ever weirder hardware faster than FOSS can keep up.

-- 
Peter Jeremy
Please excuse any delays as the result of my ISP's inability to implement
an MTA that is either RFC2821-compliant or matches their claimed behaviour.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://mailman.barnet.com.au/pipermail/bugs/attachments/20080128/7748aa6f/attachment.bin