Intel MacBook & iMac NOW available at Mac & Mac Sales!
Dan Osbourne
 |
|
 |
|
 |
|
 |
|
 |
|
0
MSCHAEF.COM
...
>MacBook Pro
...
>-Intel Core Duo processor up to 1.8GHz
>-Built-in iSight Camera
>-Front Row and Apple Remote
>-ATI Mobility Radeon
>-Mac OS Tiger and iLife '06
You missed the best feature: the magnetically attached power cord.
-Mike
--
http://www.mschaef.com
Mike Hore
> MacBook Pro
>
> New engine. Same soul!
>
> -Intel Core Duo processor up to 1.8GHz
^^^^^
I was wondering for a while what relevance this post had to afc,
but I think I just worked it out -- certainly a huge step backwards
to a ghastly CISC architecture designed in the '80s... an
absolute disaster compared to the nice clean PowerPC. It belongs
in a museum.
(All IMNSHO, of course)
Cheers, Mike.
----------------------------------------------------------------
Mike Hore mike.h...@OVE.invalid.bigpond.com
----------------------------------------------------------------
Richard E Maine
> You missed the best feature: the magnetically attached power cord.
AMEN! Do you have any idea (well, probably you do, since you mentioned
this) how much it costs to get a powerbook repaired if the pin has
broken off inside of that darned connector? :-(
Or how much it costs in sweat if it happens a second time and you decide
to try to self-repair it? :-(
I do. :-(
P.S. I'm amazed that the darn thing still worked after I put it back
together after the self-repair. I'm reasonably comfortable with diddling
around inside of computers, but that was a hard one.
P.P.S. Yes, my daughter is a bit hard on the poor machine. :-(
--
Richard Maine | Good judgment comes from experience;
email: my first.last at org.domain| experience comes from bad judgment.
org: nasa, domain: gov | -- Mark Twain
Bruce Hoult
Mike Hore <mike...@bigpond.com> wrote:
> I was wondering for a while what relevance this post had to afc,
> but I think I just worked it out -- certainly a huge step backwards
> to a ghastly CISC architecture designed in the '80s... an
> absolute disaster compared to the nice clean PowerPC. It belongs
> in a museum.
"designed in the 80's". Two lies in four words :-)
The 8086 was released in mid 1978. Whether it was designed is
debatable, but if it did occur it was certainly before then.
--
Bruce | 41.1670S | \ spoken | -+-
Hoult | 174.8263E | /\ here. | ----------O----------
Dan Osbourne
Mike Hore
> In article <BFED1112.6C8F1%mike...@bigpond.com>,
> Mike Hore <mike...@bigpond.com> wrote:
>
>> I was wondering for a while what relevance this post had to afc,
>> but I think I just worked it out -- certainly a huge step backwards
>> to a ghastly CISC architecture designed in the '80s... an
>> absolute disaster compared to the nice clean PowerPC. It belongs
>> in a museum.
>
> "designed in the 80's". Two lies in four words :-)
Mea culpa, IOW, guilty as charged, yer honor. Post in haste,
repent at leisure :-)
>
> The 8086 was released in mid 1978. Whether it was designed is
> debatable, but if it did occur it was certainly before then.
True. They wanted to keep as compatible with the 8080 as
possible. That particular albatross hangs around a lot of
necks... I hope Apple eventually see the error of their
ways, but I'm not holding my breath.
Bruce Hoult
Mike Hore <mike...@bigpond.com> wrote:
> > The 8086 was released in mid 1978. Whether it was designed is
> > debatable, but if it did occur it was certainly before then.
>
> True. They wanted to keep as compatible with the 8080 as
> possible. That particular albatross hangs around a lot of
> necks... I hope Apple eventually see the error of their
> ways, but I'm not holding my breath.
The problem is that the size of the market for running legacy x86
binaries has been such that both Intel and AMD have been able to afford
to thow sufficient extra money at making the cruft run fast to stay more
or less even with cleaner architectures.
Both Alpha and PowerPC have been slightly faster than x86, at reasonable
cost, at various times -- but not *enough* faster (2:1 or better) to
make one of those with emulator software the fastest x86 available, or
the fastest x86 at a given cost or power usage.
The reverse has also never been true in the past, but the new Pentium
M-style chips appear to be fast enough and low power enough to emulate a
G4 in a portable.
Andrew Swallow
[snip]
>
> True. They wanted to keep as compatible with the 8080 as
> possible. That particular albatross hangs around a lot of
> necks... I hope Apple eventually see the error of their
> ways, but I'm not holding my breath.
Can a version of the ARM microprocessor be produced that can match the
X86 family for speed? And not just in Intel's foundries.
Andrew Swallow
MSCHAEF.COM
Mike Hore <mike...@bigpond.com> wrote:
...
>Dan Osbourne wrote:
...
>> -Intel Core Duo processor up to 1.8GHz
> ^^^^^
>I was wondering for a while what relevance this post had to afc,
>but I think I just worked it out -- certainly a huge step backwards
>to a ghastly CISC architecture
Isn't it more of a CISC ISA in front of a RISC architecture? From P6
onward, Intel chips have been translating from x86 instructions into
micro-ops which are apparantly pretty RISC like. Worrying that the
final translation to RISC ocurrs in hardware instead of the last stage of
software seems a little pointless...
-Mike
--
http://www.mschaef.com
MSCHAEF.COM
Bruce Hoult <br...@hoult.org> wrote:
...
>The problem is that the size of the market for running legacy x86
>binaries has been such
Why is that a problem, exactly? A large base of legacy software that
people want to run seems to indicate that there are a bunch of people
happily and productively running x86 boxes... isn't that the whole point?
-Mike
--
http://www.mschaef.com
Mike Hore
Hi Mike,
Yes, that's true, but it all depends on your point of view. Mine is
as a sometime compiler writer, and the compilers have to generate
that ghastly x86 code. I guess it doesn't matter a huge deal to the
end user, but as Bruce said, a lot has to go on under the hood
that Intel & AMD have been able to afford because of the huge
market. It's the market reality, but it's inelegant and messy.
I can't help feeling that these inelegant chickens are going to
come home to roost one day...
Andrew Swallow
If the design is genuinely elegant then the computer should be able to
run more lines of C or FORTRAN per second for the same clock speed.
This is becoming more important now that clock rates are only increasing
slowly.
Andrew Swallow
Bruce Hoult
Andrew Swallow <am.sw...@btopenworld.com> wrote:
> > Yes, that's true, but it all depends on your point of view. Mine is
> > as a sometime compiler writer, and the compilers have to generate
> > that ghastly x86 code. I guess it doesn't matter a huge deal to the
> > end user, but as Bruce said, a lot has to go on under the hood
> > that Intel & AMD have been able to afford because of the huge
> > market. It's the market reality, but it's inelegant and messy.
> > I can't help feeling that these inelegant chickens are going to
> > come home to roost one day...
>
> If the design is genuinely elegant then the computer should be able to
> run more lines of C or FORTRAN per second for the same clock speed.
> This is becoming more important now that clock rates are only increasing
> slowly.
Clock speed is largely irrelevant over a fairly wide range. For a given
gate delay, you can design a CPU that does a lot in each clock cycle and
has a low clock speed, or a CPU that does very little each clock cycle
and has a high clock speed, and both will run programs at about the same
rate. Athlons and Centrinos have much lower clock rates than P4's, but
run just as fast.
For the same amount of money to pay CPU designers, and the same amount
of money to spend on the latest technology factories, you can get a
faster PowerPC or MIPS or Alpha than x86. The problem is that not
everyone has the same amount of money to spend.
Even so, the fastest PowerPCs (2.7 GHz G5) do in fact run code as fast
as or faster than the fastest x86's from Intel (AMD64's may have gotten
ahead a little now). So the high end tower machines will be the last
ones that Apple converts to Intel.
Peter Corlett
[...]
> I was wondering for a while what relevance this post had to afc, but
> I think I just worked it out -- certainly a huge step backwards to a
> ghastly CISC architecture designed in the '80s... an absolute
> disaster compared to the nice clean PowerPC. It belongs in a museum.
x86-64 sucks slightly less hard if you ignore the legacy cruft and
stick to just the new 64 bit cruft.
--
Everyone must believe in something. I believe I'll have another drink.
- W.C. Fields
Mike Hore
> Mike Hore <mike...@bigpond.com> wrote:
> [...]
>> I was wondering for a while what relevance this post had to afc, but
>> I think I just worked it out -- certainly a huge step backwards to a
>> ghastly CISC architecture designed in the '80s... an absolute
>> disaster compared to the nice clean PowerPC. It belongs in a museum.
>
> x86-64 sucks slightly less hard if you ignore the legacy cruft and
> stick to just the new 64 bit cruft.
Hi Peter,
Yes this is probably true -- but "cruft" is still the word.
Looking at the 64-bit extensions manual, I see that the ADD
instruction has 22 variants. I think. Very CISC-y.