The End of the '040 Macintoshes???
Doug Slansky
Macintosh? The reason I bring this up is twice now I have seen
"unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
INCLUDING monitor and keyboard. This figure seems quite low, given that
high end 040 systems are selling for quite a bit more than this (in similar
configurations). Even the least expensive 040 is around $1300 including
monitor and keyboard.
To further speculation, our Macintosh sys admin attended a product demo
from Apple reps for the new PowerPC models, and in his summary stated that
ONLY PowerPC desktop systems will be available from Apple by the end of
'94. Of course, we don't get much into the performa or LC lines here, so I
don't know if his comment related to those lines as well.
Even if this is untrue, and Apple continues to sell the 040 based systems,
I think there must be a MAJOR slashing of prices if the pricing on the
6100-60 (hate the names, btw) is even close (+/- a few hundred) to
accurate.
The way I see it, existing 040 stock will come down hard in price, either
through price cuts or simple stock elimination. Also, it would seem that
the resale prices will undergo similar changes.
Keep in mind, these are pure speculation - I just wanted to see what the
wisdom of the net has to say, since I'm hoping to pick up an AV Mac on the
cheap if I'm right....
cheers,
doug
--
Doug Slansky
Motorola Cellular Infrastructure Group
Advanced Products Division
AKA: slvrs...@aol.com
"Yeah, Right"
Mr_G._Landweber_student_tel_2-73550
Any netland speculation on the end of the 030/040 product line from
Macintosh? The reason I bring this up is twice now I have seen
"unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
INCLUDING monitor and keyboard. This figure seems quite low, given that
high end 040 systems are selling for quite a bit more than this (in similar
configurations). Even the least expensive 040 is around $1300 including
monitor and keyboard.
One of the biggest advantages of the PowerPC chips is that they are cheap,
so Apple will be able to produce the new Macs for less (and either drop the
prices or increase their margins...).
To further speculation, our Macintosh sys admin attended a product demo
from Apple reps for the new PowerPC models, and in his summary stated that
ONLY PowerPC desktop systems will be available from Apple by the end of
'94. Of course, we don't get much into the performa or LC lines here, so I
don't know if his comment related to those lines as well.
I don't believe this. In several cases, I have heard rumors that Apple
say they will still continue to produce 680x0 Macs as long as people want
to buy them. However, I expect that most of the price drops and technological
advances will come on the PowerPC Macs. Maybe Apple is so excited about the
PowePC that they think nobody will want a 680x0 Mac by the end of '94!
Even if this is untrue, and Apple continues to sell the 040 based systems,
I think there must be a MAJOR slashing of prices if the pricing on the
6100-60 (hate the names, btw) is even close (+/- a few hundred) to
accurate.
I don't think Apple is able to slash prices on the 680x0 series much as
their margins are already very slim.
The way I see it, existing 040 stock will come down hard in price, either
through price cuts or simple stock elimination. Also, it would seem that
the resale prices will undergo similar changes.
Yup. Resale prices will drop through the floor. Better sell you 680x0
Macs now!
Keep in mind, these are pure speculation - I just wanted to see what the
wisdom of the net has to say, since I'm hoping to pick up an AV Mac on the
cheap if I'm right....
The PowerPCs won't really be able to surpass the AV Macs until the end of
'94 at the earliest. Apple still positions the Q840av at the top of the line.
(These are all speculations on my part, too)
-- Greg
Erik A. Speckman
I expect the 040 machines to be fazed out before the third round of
PowerPC macs debuts. I don't know how long that will be but certianly
within a year. (I am not saying we will see the thirs round within a
year, just that we will see the last of the 68k macs within a year.0
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
I am not a Netter.
____________________________________________________________________________
Jim Woodgett
Speckman) writes:
> I expect the 040 machines to be fazed out before the third round of
> PowerPC macs debuts. I don't know how long that will be but certianly
> within a year. (I am not saying we will see the thirs round within a
> year, just that we will see the last of the 68k macs within a year.0
> --
Hmmmm. It took Apple just a little while longer to "phase" out the Apple II
after the introduction of the Mac (at least 8 years; maybe you can still get
them new?). Admittedly this was primarily due to the educational market. My
kids school has Macs and Apple IIs and I doubt that Apple will abandon that
market. In fact its rather a good way to slowly reduce their inventory
without endangering the PPC market by flooding the country with bargain Macs
(not that that is going to happen, except from current Mac owners wanting to
sell their 680X0 machines).
I agree that the "commercial" market may adapt very quickly due to
competition with the other clone makers and the fact that the 680X0 Macs cost
more to make for less performance.
Given this, its remarkable that Apple have sold 2 million UnderPowerMacs this
past quarter!
Jim
David Messina
>Speckman) writes:
>
>> I expect the 040 machines to be fazed out before the third round of
>> PowerPC macs debuts. I don't know how long that will be but certianly
>> within a year. (I am not saying we will see the thirs round within a
>> year, just that we will see the last of the 68k macs within a year.0
>> --
I would agree with that PowerPCs will probably constitute a large part
of Apple's market, barring any major design/availability problems.
But keep in mind that Apple will still appeal to home and educational
buyers at the lower end of the market. Just because the 68k machines
will become very cheap as technology marches on doesn't mean people
won't buy them. Rather, people who have never bought computers before
will probably start buying.
Think of it this way: Calculators used to cost a fortune. Just
because you can buy a loaded HP for ~$200 doesn't mean nobody sells
simple +-*/ calculators -- it means calculators are in *everybody's*
price range, and everybody has one.
So expect the 040s to stay around for a while. 000s, 020s, and 030s,
on the other hand, will disappear much more quickly.
Regards,
David
dn...@columbia.edu
Kevin Graeme
jwoo...@ocicl.oci.utoronto.ca (Jim Woodgett) wrote:
> In article <2i3j6p$8...@scratchy.reed.edu>, espe...@reed.edu (Erik A.
> Speckman) writes:
>
> > I expect the 040 machines to be fazed out before the third round of
> > PowerPC macs debuts. I don't know how long that will be but certianly
> > within a year. (I am not saying we will see the thirs round within a
> > year, just that we will see the last of the 68k macs within a year.0
> > --
>
> Hmmmm. It took Apple just a little while longer to "phase" out the Apple II
> after the introduction of the Mac (at least 8 years; maybe you can still get
> them new?).
Just for the record, Apple was still producing the AppleII computer until
just this last December when they "quietly" quit making them. They _are_
still apparently making the Apple IIe card for the Macintosh though.
It's hard to kill some things.
_______________________________________________________________________
Kevin Graeme | Macintosh Networking Dude
Computing and Networking Services at Univerity of Wisconsin-Eau Claire.
_______________________________________________________________________
Dave Wiley
>Any netland speculation on the end of the 030/040 product line from
>Macintosh? The reason I bring this up is twice now I have seen
>"unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
>INCLUDING monitor and keyboard. This figure seems quite low, given that
>high end 040 systems are selling for quite a bit more than this (in similar
>configurations). Even the least expensive 040 is around $1300 including
>monitor and keyboard.
[snip]
>The way I see it, existing 040 stock will come down hard in price, either
>through price cuts or simple stock elimination. Also, it would seem that
>the resale prices will undergo similar changes.
>Keep in mind, these are pure speculation - I just wanted to see what the
>wisdom of the net has to say, since I'm hoping to pick up an AV Mac on the
>cheap if I'm right....
No help here, sorry. I just wanted to congratulate you on a damn fine
question.
Now you've got me wondering. Suppose that Apple was so eager to ditch
the '040s that the price would be worth it not even considering the
price of the mother-board. One might be able get one of these models
then upgrade the guts (CPU? mother-board? every board?) to PPC when they
came out and still wind up money ahead. Plus you'd have the use of a
very fine machine in the mean time (upgrades are usually the last thing
to come available). The question is, of course, are any of the current
models upgradable to PPC? Would Apple say if they were?
--
david wiley "God is dead." - Nietzsche
Intergraph Corporation "Nietzsche is dead." - God
(205) 730-6390
wi...@wiley.b11.ingr.com (Shamelessly stolen from Signals Catalog)
Ian Hochman
>In article <940125134...@bono.oci.utoronto.ca>,
>jwoo...@ocicl.oci.utoronto.ca (Jim Woodgett) wrote:
>
>> Hmmmm. It took Apple just a little while longer to "phase" out the Apple II
>> after the introduction of the Mac (at least 8 years; maybe you can still get
>> them new?).
>
>Just for the record, Apple was still producing the AppleII computer until
>just this last December when they "quietly" quit making them. They _are_
>still apparently making the Apple IIe card for the Macintosh though.
>
>It's hard to kill some things.
IMHO, I think the Apple II's lived so long because there was so much
software, and people, using them. The Macs, until reecently, couldn't run
this software. On the other hand, PPCs will be able to run Mac software,
and at their projected prices I don't see any need for Apple to keep producing
040s. I guess we'll see soon enough.
-Ian
Mark Rogowsky
down to 24% average gross margin on CPUs using 680x0 chips. At the low end,
the margins are probably closer to 10-15%. They cannot and will not sell
68040 for $500. Nor can they sell those machines for over $1,000 when a
low-end PowerPC sells for just over $1,000.
Therefore, the following scenario seems likely:
1) Both types are readily available in the business (non-Performa) channel
for some months.
2) If large quantities of 610s, 650s and 800s remain, they go back to Apple
and get logic board swaps at the factory. This happened with the Centris
discontinuation a few months back.
3) Once the market accepts the PowerPC in the business channel (and Apple's
sales projections of one million PPC Macs in 1994 presume better than 90%
adoption almost immediately) the 68040 Macs leave the business line. This
is completed late in 1994 or early in 1995 by the 603-based PowerMac in the
605 enclosure.
4) The 030 and 040 survive in the Performa channel until sometime in the
middle or end of 1995. The designs are done and free. 030s and LC040s are
so cheap to Apple, and the Performa margins a bit higher, that the less
sensitive home segment still buys these machines for awhile because they
are cheap compared to the business line and because these buyers know no
better. Rambling run-on there, sorry. In short, they survive because
they're there.
5) The Powerbook lines, Blackbird and Duo II, got PowerPC 603s late in 1994
and almost immediately see their 040 counterparts phased out. Again, the
040s will be no cheaper and simply slower. However, low-end Powerbooks
still support the 030 -- assuming the MacWeek-rumored 145, '94 style sees
the light of day.
6) The PowerPC accounts for nearly 50% of Macs sold in 1994, perhaps 80-85%
of Macs in '95 and 100% of Macs in 1996. Keep in mind that the 603 is
designed to be cheap, as in sub-$100 by the time it hits full production.
The systems designed around PowerPC will be cheap, too, save the slightly
higher DRAM and hard disk costs. But the latter two are going to drop
appreciably in 1994.
Reasoned speculation, nothing more.
Mark Rogowsky
ro...@forsythe.stanford.edu
Arne Maletzky
> slansk...@macmail1.cig.mot.com (Doug Slansky) writes:
>
> >Any netland speculation on the end of the 030/040 product line from
> >Macintosh? The reason I bring this up is twice now I have seen
> >"unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
> >INCLUDING monitor and keyboard. This figure seems quite low, given that
> >high end 040 systems are selling for quite a bit more than this (in similar
> >configurations). Even the least expensive 040 is around $1300 including
> >monitor and keyboard.
>
there code to gain the advantage of the new PPS macs. They probably
will not develop the code for 'old' macs . In the end we all have
cheap old macs but no software ( except what exsists now )
Must be like Erik Speckman (espe...@reed.edu) wrote It's Bye bye to
the old macs Sorry !
--
Tax ?? what TAX ?? Oh TAXI !!
----------------------------------------------------------------------
Arne Maletzky m...@mh.nl via internet backbones
Multihouse Automatisering uucp: ..!{uunet,sun4nl}!mh.nl!mlz
Doesburgweg 7 Gouda, The Netherlands
----------------------------------------------------------------------
Say Cheese and smile again - say smile and.... Cheese again?!?!?
Ken Stevens
> In article <slansky_doug...@mac-an-46.cig.mot.com>,
> slansk...@macmail1.cig.mot.com (Doug Slansky) wrote:
>
> > Any netland speculation on the end of the 030/040 product line from
> > Macintosh? The reason I bring this up is twice now I have seen
> > "unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
> > INCLUDING monitor and keyboard. This figure seems quite low, given that
> > high end 040 systems are selling for quite a bit more than this (in similar
> > configurations). Even the least expensive 040 is around $1300 including
> > monitor and keyboard.
>
>
> Well, there is one good reason to buy a fast 040 machine now (650 or 800).
> They will run the existing software > 2 times as fast as the emulated
> software on the Power PC's. (see Macworld issue on Power PC)
>
> What's the point of buying a Power PC and having it run your old software
> slower than an LCIII? I don't get it!
>
> Yes, when the bulk of existing software goes to native format (which will
> take a year at least), then it makes sense to upgrade to Power PC.
>
> At least in this aspect, Pentium is far superior. It runs all the existing
> software much faster than the 486 machines. I wish Power PC Mac's did the
> same!
>
>
>
>
> --------------
> Sridhar Mahadevan
> Tampa, Florida 33647
> (813)974-3260
> maha...@guardian.csee.usf.edu
Those speed results were based on a 60 mhz version of the PowerMac.......I
wonder what the emulation speed would be for the 8100-80....If it is a
linear funtion, then the Quadra 800 and the 8100 would be close to the same
speed with the 800 slightly edging the 8100-80 (this assumes that the L2
cache doesn't affect the emulation speed).
--
Ken Stevens | kste...@srs.gov |
Sridhar Mahadevan
slansk...@macmail1.cig.mot.com (Doug Slansky) wrote:
> Any netland speculation on the end of the 030/040 product line from
> Macintosh? The reason I bring this up is twice now I have seen
> "unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
> INCLUDING monitor and keyboard. This figure seems quite low, given that
> high end 040 systems are selling for quite a bit more than this (in similar
> configurations). Even the least expensive 040 is around $1300 including
> monitor and keyboard.
Well, there is one good reason to buy a fast 040 machine now (650 or 800).
They will run the existing software > 2 times as fast as the emulated
software on the Power PC's. (see Macworld issue on Power PC)
What's the point of buying a Power PC and having it run your old software
slower than an LCIII? I don't get it!
Yes , when the bulk of existing software goes to native format (which will
Dennis Clark
: What's the point of buying a Power PC and having it run your old software
: slower than an LCIII? I don't get it!
: Yes, when the bulk of existing software goes to native format (which will
: take a year at least), then it makes sense to upgrade to Power PC.
: At least in this aspect, Pentium is far superior. It runs all the existing
: software much faster than the 486 machines. I wish Power PC Mac's did the
: same!
Um. Not really. Around here 486 code usually ran at 486 speed on the
current pentiums. Code re-compiled for pentium runs at pentium speeds.
DLC
------------------------------------------------------------------------------
| Dennis Clark (303)229-4313 telnet 1-229-4313 email d...@fc.hp.com |
| Hewlett Packard ESD Perf. Lab, 3404 East Harmony Rd. Ft. Collins CO 80525 |
| Be well, Do good work, stay in touch -- Garrison Keiller |
------------------------------------------------------------------------------
Mark Basinski
(Sridhar Mahadevan) wrote:
>
> At least in this aspect, Pentium is far superior. It runs all the existing
> software much faster than the 486 machines. I wish Power PC Mac's did the
> same!
What's wrong with this picture? Why do you think Apple came up with the
PowerPC architecture? Why do apps have to be recompiled? Because the
change from CISC to RISC is a big deal!
The PPC601 is the first of new type of architecture.
The Pentium is the LAST of the old 80x86 architecture.
That is why the Pentium can run old apps with no recompilation, and that
is also why the Pentium is doomed! The *first* PPC chips keep pace with
it (if the apps are recompiled, yes) and the succeeding chips (esp the
604 and 620 coming next year) will kill it totally....for this kind of
breakthru via a different architecture, OF COURSE there is a problem
making the transition, and suporting the emulation of the 040 is a
pretty good solution. The long term solution is recompiliation. If you
insist on backward compatibility forever, eventually (i.e. Pentium)
you'll reach a level of performance that cannot be raised.
----------------------------------------------------------------------------
Mark Basinski Internet: basi...@biosci.arizona.edu
Andrew Geweke
ihoc...@mail.sas.upenn.edu writes:
>IMHO, I think the Apple II's lived so long because there was so much
>software, and people, using them. The Macs, until reecently, couldn't run
>this software. On the other hand, PPCs will be able to run Mac software,
>and at their projected prices I don't see any need for Apple to keep
producing
>040s. I guess we'll see soon enough.
What about this:
PowerPC Macs emulating Apple IIs in software...
Want an Apple IIe emulator running at something like twenty or thirty
times its original speed? Geeze, the IIe emulator card didn't have a 6502
on it (or whatever it was), and it ran at full speed on a _LC_. Think
what a PowerMac could do to it. Want an Apple IIe with a 300MB HD (jeeze,
you could probably put all Apple II software ever made on it) and 16 MB
RAM?
--- Andrew Geweke / gewe...@studentg.msu.edu / Computer Engineering ---
"Love is a temple / Love the higher law" - U2, Achtung Baby
Andrew Geweke
maha...@csee.usf.edu writes:
>What's the point of buying a Power PC and having it run your old software
>slower than an LCIII? I don't get it!
Er...perhaps this is just me, but I have a Speedometer record set that
has a few entries for something called "PDM 66" with the comment
"PowerPC". It sure sounds like the $2000 PPC to me. And it comes in about
6.0 on what I think of as the standard scale -- a Classic is 1.0. This is
emulated me on the cheapest PowerPC out there. The LC III is only a 4.0
or slightly higher. The C610 is about a 9.0.
Ian Hochman
>What about this:
>
>PowerPC Macs emulating Apple IIs in software...
>
>Want an Apple IIe emulator running at something like twenty or thirty
>times its original speed? Geeze, the IIe emulator card didn't have a 6502
>on it (or whatever it was), and it ran at full speed on a _LC_. Think
>what a PowerMac could do to it. Want an Apple IIe with a 300MB HD (jeeze,
>you could probably put all Apple II software ever made on it) and 16 MB
>RAM?
Please, my mind can only take so much before exploding! =>8-(
-Ian
Andy Pearlman
>Largely lost in this thread is the question of margins. Apple is already
>down to 24% average gross margin on CPUs using 680x0 chips. At the low end,
>the margins are probably closer to 10-15%. They cannot and will not sell
>68040 for $500. Nor can they sell those machines for over $1,000 when a
>low-end PowerPC sells for just over $1,000.
A thing to remember is that a 33 Mhz 68040 is still going to be considerably
faster for a while than a lot of the PowerPCs. Especially if vendors charge
for upgrades...
>Therefore, the following scenario seems likely:
>3) Once the market accepts the PowerPC in the business channel (and Apple's
>sales projections of one million PPC Macs in 1994 presume better than 90%
>adoption almost immediately) the 68040 Macs leave the business line. This
>is completed late in 1994 or early in 1995 by the 603-based PowerMac in the
>605 enclosure.
I think Apple would be foolish not to make some kind of bundle arrangement
with Insignia. Charge an extra $100 and tack Windows emulation onto *every*
PPC. I think they'd do pretty well if they did that.
Andy Pearlman
--
Andy Pearlman
apea...@panix.com
"Someone stole all the paintings in a museum, leaving a building as a ransom
note." "What did it say?" "I don't know, the writer's a deconstructionist."
Matthew T. Russotto
}What about this:
}
}PowerPC Macs emulating Apple IIs in software...
}
}Want an Apple IIe emulator running at something like twenty or thirty
}times its original speed? Geeze, the IIe emulator card didn't have a 6502
}on it (or whatever it was), and it ran at full speed on a _LC_.
It had a 65C02 all right, as well as the soft-switching and
screen-mangling hardware (they might have even used a Mega-II,
essentially an Apple-II-in-a-chip).
}Think
}what a PowerMac could do to it. Want an Apple IIe with a 300MB HD (jeeze,
}you could probably put all Apple II software ever made on it) and 16 MB
}RAM?
16MB RAM? How do you bank-switch it ?
--
Matthew T. Russotto russ...@eng.umd.edu
Some news readers expect "Disclaimer:" here.
Just say NO to police searches and seizures. Make them use force.
(not responsible for bodily harm resulting from following above advice)
eunjoon lee
>there code to gain the advantage of the new PPS macs. They probably
>will not develop the code for 'old' macs . In the end we all have
>cheap old macs but no software ( except what exsists now )
Apple has been selling Macs for ten years. Installed user base for existing
Macs are huge, although not as big as PCs. No matter how amazing new PPC macs
are, it'll take a considerable amount of time for the new PPC macs to outnumber
existing 680x0 based Macs. It means market for 680x0 will exist until user
base for 680x0 based Macs becomes negligibly small. Software develpers make
softwares to sell it to those who *Have* machines, not those who *Will have*
machines.
Eunjoon
Erik A. Speckman
>In article <rogo-260...@tip-mp4-ncs-5.stanford.edu> ro...@forsythe.stanford.edu (Mark Rogowsky) writes:
>>Largely lost in this thread is the question of margins. Apple is already
>>down to 24% average gross margin on CPUs using 680x0 chips. At the low end,
>>the margins are probably closer to 10-15%. They cannot and will not sell
>>68040 for $500. Nor can they sell those machines for over $1,000 when a
>>low-end PowerPC sells for just over $1,000.
>
>A thing to remember is that a 33 Mhz 68040 is still going to be considerably
>faster for a while than a lot of the PowerPCs. Especially if vendors charge
>for upgrades...
I think 'considerably faster' is an overstatement but we will see.
However, within a year I think a 33 MHz 040 will be slow compaired to a
low end PowerPC system under emulation.
Mark Rogowsky
wrote:
>
> In article <rogo-260...@tip-mp4-ncs-5.stanford.edu> ro...@forsythe.stanford.edu (Mark Rogowsky) writes:
> >Largely lost in this thread is the question of margins. Apple is already
> >down to 24% average gross margin on CPUs using 680x0 chips. At the low end,
> >the margins are probably closer to 10-15%. They cannot and will not sell
> >68040 for $500. Nor can they sell those machines for over $1,000 when a
> >low-end PowerPC sells for just over $1,000.
>
> A thing to remember is that a 33 Mhz 68040 is still going to be considerably
> faster for a while than a lot of the PowerPCs. Especially if vendors charge
> for upgrades...
Right a 33MHz 68040 is going to be faster than a PowerPC Mac in heavily
emulated apps. But a 25MHz 040 -- the low end -- will not be faster than
the PowerPC, even in the vast majority of emulation. The low end 040s will
dry up and the high end apps will be the first to go native -- see
Freehand, Photoshop, Pagemaker, Quark, Illustrator -- all within 6 weeks
(except probably Quark since they are notorious for slow upgrades).
>
> >Therefore, the following scenario seems likely:
> >3) Once the market accepts the PowerPC in the business channel (and Apple's
> >sales projections of one million PPC Macs in 1994 presume better than 90%
> >adoption almost immediately) the 68040 Macs leave the business line. This
> >is completed late in 1994 or early in 1995 by the 603-based PowerMac in the
> >605 enclosure.
>
> I think Apple would be foolish not to make some kind of bundle arrangement
> with Insignia. Charge an extra $100 and tack Windows emulation onto *every*
> PPC. I think they'd do pretty well if they did that.
Except for the millions who don't care about Windows and don't want to pay
$100 more.
>
Mark Rogowsky
ro...@forsythe.stanford.edu
Erik A. Speckman
eunjoon lee <eun...@seas.gwu.edu> wrote:
>
>Apple has been selling Macs for ten years. Installed user base for existing
>Macs are huge, although not as big as PCs. No matter how amazing new PPC macs
>are, it'll take a considerable amount of time for the new PPC macs to outnumber
>existing 680x0 based Macs. It means market for 680x0 will exist until user
>base for 680x0 based Macs becomes negligibly small. Software develpers make
>softwares to sell it to those who *Have* machines, not those who *Will have*
>machines.
I am going to suggest that machines older than 2 years are "economically
inert" from the point of view of software developers. Much of the
software that is going to be written is going to be too bloated or slow
to run on older machines. This eliminates the incentive for developers
to write software for part of the installed base of 68k machines if it is an
inconvenience.
Within 3 years all of the 68k based machines will probably be too out of
date for most of the latest software because even if they sell 040 machines
for the next year and a half they aren't getting much faster.
Seth D. Kadesh
> IMHO, I think the Apple II's lived so long because there was so much
> software, and people, using them. The Macs, until reecently, couldn't run
Still are. The Apple II isn't quite dead.
> this software. On the other hand, PPCs will be able to run Mac software,
> and at their projected prices I don't see any need for Apple to keep
> producing 040s. I guess we'll see soon enough.
The thing that everyone seems to be missing in this discussion is that
- until everything is native PPC code, the PPC emulating a Mac is
slower than the fastest 040. [+ demand]
- there will be people that want the assurance of a "proven"
technology [+ demand]
- Apple can't afford to sell the machines for less than they're worth
So I think we'll see a slowdown in the production of the 040 Macs, but
I don't expect any firesales....
Seth ---> th...@cmu.edu It is not enough to watch.
"I said 'Forget Topsoil, Think Tinfoil, There's tons of kids who want
to get tanned.'" - Think Tree
103t_e...@west.cscwc.pima.edu
David B. Enfield
basi...@biosci.arizona.edu (Mark Basinski) wrote:
>the 604 and 620 coming next year) will kill Pentium totally.
Enough is known about 601 emulation, 1994 product lines, cpu comparisons and
PPC code development that the immediate future is fairly obvious to anyone
reading the trades. Thus, this thread is becoming boring...
Now, let's really take the thread into the realm of speculation: A year from
now we'll have boxes with much faster 604's, a screaming PCI bus
architecture, AV & Windows capabilities, major software running in PPC
native mode, and prices that have settled down to attractive, competitive
levels. Time to buy, right?
Q: A year from now: Will the speculation then shift to the 620 and whether
one should invest in the wonderful ("bird-in-hand") 604 boxes or wait for
the ("two-in-the-bush") 620? We know ONE answer, of course: "Buy when you
need it or you'll wait forever for what's around the corner."
Seriously though, is a 620 "for the rest of us" going to be an option? Or is
it going to be so expensive that only rocket scientist (RS) dweebs will use
them in a workstation architecture? Will it be offscale compared to the 604
end of the line? Will competition and technology be tempting us with other
kinds of improvements to those same 604 boxes, leaving the 620s to the RS's?
Or, will AV and other technologies be clamoring for 620 power, prompting
Apple to put out a middle-of-the-line 620 box at a reachable price?
Try that one on for size...
****************************************************************************
* David B. Enfield | "Every CRISIS is an OPPORTUNITY" *
* NOAA Oceanographer | Philosophy implicit in the Chinese words *
* enf...@ocean.aoml.erl.gov | symbolizing both: "HUE JI" <==> "JI HUI" *
****************************************************************************
Kurt A. Seiffert
ro...@forsythe.stanford.edu writes:
>> I think Apple would be foolish not to make some kind of bundle
arrangement
>> with Insignia. Charge an extra $100 and tack Windows emulation onto
*every*
>> PPC. I think they'd do pretty well if they did that.
>
>Except for the millions who don't care about Windows and don't want to
pay
>$100 more.
I don't know. Apple could make a very strong position statement that
every PowerMac is DOS, Windows, and Mac application complient. It would
make a decent case to penny-watching managers who are concerned about
precious DOS applications.
The question boils down to what would get you more customers, $100
lower price point or software compatiblity with 80-90% of all the
software sold for the personal computer desktop market out of the box
with no hassles. (Well other than the normal hassles of dealing with
Windows and DOS.)
Of course, I would love to see an arrangment that you get a purchase
deal on SoftWindows with every PowerMac bought. Then you could choose
whether or not you want to spend the $100 on SoftWindows or on VRAM.
(According to the latest MacWeek rumoring, the PowerMacs will come sans
VRAM as in a IIsi.)
*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*--*-
Kurt A. Seiffert internet: seif...@ucs.indiana.edu
UCS, Multimedia Technologist office ph: (812) 855-5746
IU Bloomington,IN
"No, no. I am an earthling. I just can't prove it." -- Zonker
DISCLAIMER: I don't speak for IU and IU doesn't speak for me.
We both like it that way. ;-)
Sridhar Mahadevan
Speckman) wrote:
> Within 3 years all of the 68k based machines will probably be too out of
> date for most of the latest software because even if they sell 040 machines
> for the next year and a half they aren't getting much faster.
>
> --
I seriously doubt this scenario. People do not upgrade computers every year
(much as we would want them to). 3 years down the road, plenty of homes and
schools will continue to have millions of 68K machines, and any vendor who
does not ensure that his/her program runs on the old machines will be
foolishly throwing away a big chunk of profits. (Heck, there are plenty of
the old Mac Classics around still, and those are close to 10 years old!).
----
Sridhar Mahadevan
Department of Computer Science
University of South Florida
Eric Smith
>Enough is known about 601 emulation, 1994 product lines, cpu comparisons and
>PPC code development that the immediate future is fairly obvious to anyone
>reading the trades. Thus, this thread is becoming boring...
>Now, let's really take the thread into the realm of speculation: A year from
>now we'll have boxes with much faster 604's, a screaming PCI bus
>architecture, AV & Windows capabilities, major software running in PPC
>native mode, and prices that have settled down to attractive, competitive
>levels. Time to buy, right?
Right, exactly. The question shouldn't be whether the 040 Macs are dead
end products, but whether the PPC 601, Nubus-based Macs are the real
dead end products. I think the best buying strategy is pretty clear -
the 040 Macs are proven technology and where all the software runs with
minimal problems NOW, and 1.5 to 2 years from now PCI bus, PPC 604 Macs
will be the fast machines everybody wants, when the majority of software
has finally migrated to the PPC. Apple, for obvious reasons, is not
marketing the initial PPC Macs as an interim product line but it seems
obvious that that's the case.
-----
Eric Smith
er...@netcom.com
er...@infoserv.com
CI$: 70262,3610
Thomas M Gooding
>In article <2i7ra3$c...@scratchy.reed.edu>, espe...@reed.edu (Erik A.
>Speckman) wrote:
>> Within 3 years all of the 68k based machines will probably be too out of
>> date for most of the latest software because even if they sell 040 machines
>> for the next year and a half they aren't getting much faster.
>>
>> --
>I seriously doubt this scenario. People do not upgrade computers every year
>(much as we would want them to). 3 years down the road, plenty of homes and
>schools will continue to have millions of 68K machines, and any vendor who
>does not ensure that his/her program runs on the old machines will be
>foolishly throwing away a big chunk of profits. (Heck, there are plenty of
>the old Mac Classics around still, and those are close to 10 years old!).
Actually a better example would be the Apple // line. Apple finally
discontinued the //e a couple months ago, however it's technology was 11 years
old and software wasn't being developed. The Mac Classic could still run most
programs. 680x0 Macs will still be sold as long as their cheaper than a PPC
Mac. Initially the PPCs will be more expensive, due to demand.
Question: Is the 68040 chip more/less expensive than a PPC?
--
Tom Gooding
tgoo...@iastate.edu
Iowa State University
Peter Kocourek
AM> so what if Apple continues to sell 'old' macs Developers must
AM> rewrite there code to gain the advantage of the new PPS macs.
Rewrite? Hardly. Some minor changes, but basically, you just recompile your
source code.
AM> They probably will not develop the code for 'old' macs . In
AM> the end we all have cheap old macs but no software ( except
AM> what exsists now )
Sooner or later, developers will stop compiling 680x0 versions of their
software, but with the installed base of "old" Macs, it's hardly likely this
situation will trap you in the immediate future. By the time this happens, you
Mac will be too old anyway to do anything useful with it.
AM> Must be like Erik Speckman (espe...@reed.edu) wrote It's Bye
AM> bye to the old macs Sorry !
Yep. The PowerPC is the future, and let's rejoice at the prospect.
YHS:QSI!
Noah Price
wrote:
>
> ...the 040 Macs are proven technology...and 1.5 to 2 years from now...
> ...the fast machines everybody wants...
> Apple, for obvious reasons, is not
> marketing the initial PPC Macs as an interim product line but it seems
> obvious that that's the case.
C'mon, if you look at things that way _every_ computer is part of an
interim product line. Last years model is always proven, and there will
_always_ be a better/faster/cheaper computer out in two years.
noah
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Noah Price (not the opinions of) Apple Computer, Inc.
Macintosh AV Hardware 20525 Mariani Ave., MS 60-TNT
no...@apple.com Cupertino CA 95014
L.H....@lut.ac.uk
>
>I seriously doubt this scenario. People do not upgrade computers every year
>(much as we would want them to). 3 years down the road, plenty of homes and
>schools will continue to have millions of 68K machines, and any vendor who
>does not ensure that his/her program runs on the old machines will be
>foolishly throwing away a big chunk of profits. (Heck, there are plenty of
>the old Mac Classics around still, and those are close to 10 years old!).
The Mac Classic was introduced in October 1990. That makes it 3.5 years old.
L.
Mark Basinski
enf...@ocean.aoml.erl.gov (David B. Enfield) wrote:
>
> Seriously though, is a 620 "for the rest of us" going to be an option? Or is
> it going to be so expensive that only rocket scientist (RS) dweebs will use
>
Wiil there be a 620 system for the rest of us?
Depends on your bank balance, doesn't it? In truth, the answer for most of
us wil be a resounding NO. And why shouldn't it? Why should any company
provide their most powerful system a prices that undercut the rest of the
product line? I mean, I'd like to have the fastest thing just like you
would,
but I confess to being both puzzled and amused by the idea that a) Apple's
cheapest system should do everything and have every option or b) Apple's
fastest, best system should be cheap...
Apple has stated that the 620 (a multi processing system) is targeted at
the high-end, semi-workstation market (though it really won't compete
directly with Sun or SGI).
Jesse Hertzberg
> Seriously though, is a 620 "for the rest of us" going to be an option? Or is
> it going to be so expensive that only rocket scientist (RS) dweebs will use
> them in a workstation architecture? Will it be offscale compared to the 604
> end of the line? Will competition and technology be tempting us with other
> kinds of improvements to those same 604 boxes, leaving the 620s to the RS's?
> Or, will AV and other technologies be clamoring for 620 power, prompting
> Apple to put out a middle-of-the-line 620 box at a reachable price?
>
> Try that one on for size...
Your redirection of this thread is to be commended. What a wonderful Net
it would be
if everyone made thoughtful comments instead of flaming about IBM.
In my opinion, it is critical for Apple to continue putting power on
EVERYone's desktop.
Sure, when the 620 comes its cost is going to be to the 601 or 604 what the
040 was to
the 030. But people will always want that kind of power, and will pay for
it. And as
people pay for it the price comes down and then we all buy it. When I
bought my
IIsi I though an 040 was the most out of my reach item in the world. And
see?
I also think the AV and Heandwriting technologies are going to demand that
power. But
with that power, we may be closer to Star Trek than we realize.
--
hert...@wharton.upenn.edu
-------------------------------------------------------------------------------------------------------
Set the
gearshift for the high gear of your soul;
You've got to
run like an antelope out of control.
-------------------------------------------------------------------------------------------------------
Jesse Hertzberg
University of Pennsylvania
Philadelphia, PA 19104
Doug Slansky
> Now you've got me wondering. Suppose that Apple was so eager to ditch
> the '040s that the price would be worth it not even considering the
> price of the mother-board. One might be able get one of these models
> then upgrade the guts (CPU? mother-board? every board?) to PPC when they
> came out and still wind up money ahead. Plus you'd have the use of a
> very fine machine in the mean time (upgrades are usually the last thing
> to come available). The question is, of course, are any of the current
> models upgradable to PPC? Would Apple say if they were?
>
> --
> david wiley "God is dead." - Nietzsche
> Intergraph Corporation "Nietzsche is dead." - God
> (205) 730-6390
> wi...@wiley.b11.ingr.com (Shamelessly stolen from Signals Catalog)
Motorola has publicly committed to upgrade paths for the following
configurations that I am aware of: The C/Q610 (and presumably C/Q660 since
they use the same box). The Q650 and IIvx (again, same case). The
Q800/840 will also be upgradeable. In addition, the usual round of
accelerator board suspects have indicated that they will provide
accelerator or mother board upgrades for most of the II series and
remaining Quadra series. Note also that some of the performa/LC series are
equivalent to some of the II series (P600/LC475/IIvx for example). So, I'm
willing to bet that upgrade paths will be available for all except the
bottom of the current offering barrel. As far as your logic goes, I have
been thinking the same thing -> namely, wait for the Q660AV price to come
down for clearance, snatch one up, then upgrade when my budget (read: WIFE)
allow....
As to one of the other posters, I don't agree that the 040 will survive
longer than the '030. The way I see it, margins are killing Apple right
now, and to reduce the 040 prices further for anything other than clearance
will put them in a loss mode. The 030's, on the other hand could still
sell to the education and low-end home market at significant discounts to
the PPC, while still eeking out a (meager) profit. Unless Apple goes into
NEXTmode, and only supplies high-end gear (unlikely based upon their
history in the edu market), I think they will continue to ship low cost 030
machines even after the 040 is gone...
Again, I could be (and frequently am) wrong....
-doug
PETER K. ZEITLER
il1.cig.mot.com (Doug Slansky) writes:
>Any netland speculation on the end of the 030/040 product line from
[info about harbingers of change deleted}
>--
>Doug Slansky
>Motorola Cellular Infrastructure Group
>Advanced Products Division
>AKA: slvrs...@aol.com
Following up on this thread, what will this brave PPC world be like for those
who have various committments to 030 and 040 Macs? I ask because I'm about to
automate my laboratory (using LabView), and I'll be shelling out significant
bucks. If before long, PPC's rule, what happens if one of my older Macs packs
it in? On the peecee side of things, I have no doubt that there will be x86
clonoids available for the rest of time. But with Macs, dunno. I can't really
wait for the PPC line to shake itself out (and for all the appropriate
ahrdware and software to be released). I suppose there will used 040's out on
the market for a while. What's the scoop? Does anyone have predictions about
the likely decay curve for older Macs?
--
Peter K. Zeitler (pk...@lehigh.edu)
Earth & Environmental Sciences, Lehigh University, Bethlehem, PA
Chuck Simciak
espe...@reed.edu writes:
>Within 3 years all of the 68k based machines will probably be too out of
>date for most of the latest software because even if they sell 040
machines
>for the next year and a half they aren't getting much faster.
And why not?? I understand that the chip, going from '30 to '40, makes
the biggest difference in performance. But the fastest chip out is only
the 40Mhz in the Quadra 800. (and the AV I think) There are 486DX2
chips out there that do 66Mhz, in addition I heard that the DX2
architecutre supports the addition of another chip, doubling performance.
Can Motorola squeeze some more juice out of the '40 before they have to
retire it??? Say something like 70Mhz....
Anyway, has anyone heard of Intel's plans for their next generation of
chips?? Can't see how they can go beyond a 586 other than to go RISC.
Chuck Simciak !"Now, Boys, you won!t see this operation performed
w...@po.cwru.edu ! very often and there!s a reason for that.. You see
sim...@ccsmtp.ccf.org! it has absolutely no medical value." - Dr. Benway
((WRUW 91.1 FM in Cleveland - over 25 years of diversity!!!))
Andrew C. Freeman
: I think 'considerably faster' is an overstatement but we will see.
: low end PowerPC system under emulation.
I'm not so optimistic. I'm looking forward to buying a PowerPC, but I'm not
looking forward to making the choice between running software at LC III speeds
or shelling out the money for upgrades. [Apple reps please skip forward to
the next message] I think Apple would be within reason to keep producing
high-end '040 Macs for people who have already invested in high-end software.
How much do you think the upgrade will be for PPC Electric Image (a $7000
program) or Authorware Professional (a $5000 program)?
Those of us who lived through the lean years of Mac software aren't looking
forward to the lean years of native PPC software. It's a trade-off: large
software base vs. better performance. Whattya gonna do.
Andy Pearlman
>In article <2i799q$e...@panix.com>, apea...@panix.com (Andy Pearlman)
>wrote:
>> In article <rogo-260...@tip-mp4-ncs-5.stanford.edu> ro...@forsythe.stanford.edu (Mark Rogowsky) writes:
>> A thing to remember is that a 33 Mhz 68040 is still going to be considerably
>> faster for a while than a lot of the PowerPCs. Especially if vendors charge
>> for upgrades...
>Right a 33MHz 68040 is going to be faster than a PowerPC Mac in heavily
>emulated apps. But a 25MHz 040 -- the low end -- will not be faster than
>the PowerPC, even in the vast majority of emulation. The low end 040s will
>dry up and the high end apps will be the first to go native -- see
>Freehand, Photoshop, Pagemaker, Quark, Illustrator -- all within 6 weeks
>(except probably Quark since they are notorious for slow upgrades).
I think a 33 Mhz Quadra 605 will probably show its head at some point. Get
rid of chips and all that. Quark has a horribly bad tech support department.
If a program has a problem, people are going to call the software, not the
hardware company. But by delaying the software's intro, they get to find out
the bugs of mass production. After all, Quark 3.3 still runs on a PowerPC,
just not that fast.
>> I think Apple would be foolish not to make some kind of bundle arrangement
>> with Insignia. Charge an extra $100 and tack Windows emulation onto *every*
>> PPC. I think they'd do pretty well if they did that.
>
>Except for the millions who don't care about Windows and don't want to pay
>$100 more.
Here's an interesting story, and my impression is that it is a common one.
A company I do freelance graphics for will only approve the buying of Macintosh
for graphics. Every other computer bought has to be a Windows machine. There
are numerous people in the company who would love to have a Mac, but because
it isn't on the 'approved' list, you can't get one.(this is to keep the need
for file exchange experts to a minimum :-)
Then, because occasionally people have to do work at home, they have to
buy a computer. So they get a Windows machine, because there is no point to
getting a Mac.
I doubt any serious Macintosh person would not buy the Mac because of that.
What are they going to do? Buy a windows machine? But there are millions of
sales that require the words:"Yes, it is a Mac, it costs about the same as a
Pentium-based Clone, and you can run all your Windows software on it as fast
as you can now." Get the companies that make applications for Windows and
Macs to offer competitive upgrades from a Windows version to a Mac Version.
And while this might make the initial cost of PPC slightly higher, in the
long run, it would probably reduce prices, due to higher volume. It also
gives the Mac user access to certain niche programs, such as games or cheap
programs that one would like to have, but just can't justify buying the Mac
version at present(word processing being the obvious one)
On the other hand, if the businessperson has to pay an extra $350-400 for
the ability to do Windows, they are just going to go out and get a Windows
machine.
Thomas E. DeWeese
Chuck Simciak <sim...@ccsmtp.ccf.org> wrote:
>In article <2i7ra3$c...@scratchy.reed.edu> Erik A. Speckman,
>espe...@reed.edu writes:
>>Within 3 years all of the 68k based machines will probably be too out of
>>date for most of the latest software because even if they sell
>>040 machines >for the next year and a half they aren't getting
>>much faster.
>And why not?? I understand that the chip, going from '30 to '40, makes
>the biggest difference in performance. But the fastest chip out is only
>the 40Mhz in the Quadra 800. (and the AV I think) There are 486DX2
>chips out there that do 66Mhz, in addition I heard that the DX2
>architecutre supports the addition of another chip, doubling performance.
> Can Motorola squeeze some more juice out of the '40 before they have to
>retire it??? Say something like 70Mhz....
A 40Mhz 68040 is very compairable (faster in floating point
even) to a 486DX2 running at 66Mhz. The Clock speed of the chip is an
even worse indicator of processor speed than the much balyhood MIPS
rating.
The only 486 that is significantly faster than a 40MHz 68040
is the fabled 486DX3 that runs at 99MHz (and as far as I know that is
still vaporware). Now a Pentium running recompiled code kills a 40Mhz
68040 (running unrecompiled code though I think it is only 20% or so
faster).
>Anyway, has anyone heard of Intel's plans for their next generation
>of chips??
>Can't see how they can go beyond a 586 other than to go RISC.
> Chuck Simciak
My understanding is that they are going to go RISC, and that
the Pentium is the first step in that direction. They are just going
to try and do it in a more backwards compatable way, than Motorolla
(personally I think backwards compatability is a big trap, but that's
me).
--
dewe...@rdrc.rpi.edu Thomas DeWeese
"Most people don't realize that large pieces of coral,
which have been painted brown and attached to the skull by
common wood screws, can make a child look like a deer."
Brendon Towle
> I understand that the chip, going from '30 to '40, makes
> the biggest difference in performance. But the fastest chip out is only
> the 40Mhz in the Quadra 800. (and the AV I think) There are 486DX2
> chips out there that do 66Mhz, [...]
Yes, but Motorola chips use an internal frequency doubler, where
Intel chips don't. So, a 40 MHz clock speed 68040 is actually running
internally at 80MHz, where a 66MHz clock speed 80486 is actually running
internally at 66MHz.
Also, MHz is meaningless. It's like saying, "My car runs at 7000
rpm." Well, fine, but until you know about the transmission, the shocks,
and the gear ratios, that doesn't tell you beans about how fast the car
will go. That analogy holds more with computers, where manufacturers have
been known to cripple high-speed components by interfacing them with low
speed components.
--
Brendon Towle | to...@ils.nwu.edu | "Jesus only told half the
Institute for the | | story. The truth *will*
Learning Sciences | "Think of it as | set you free. But, first
1890 Maple Ave. | evolution in action." | it's going to piss you
Evanston IL 60201 | -Tony Rand | off." -Solomon Short
--------------------------------------------------------------------------
Erik A. Speckman
Brendon Towle <to...@ils.nwu.edu> wrote:
>In article <1994Jan28.1...@bme.ri.ccf.org>, Chuck Simciak
><sim...@ccsmtp.ccf.org> wrote:
>
>> I understand that the chip, going from '30 to '40, makes
>> the biggest difference in performance. But the fastest chip out is only
>> the 40Mhz in the Quadra 800. (and the AV I think) There are 486DX2
>> chips out there that do 66Mhz, [...]
>
> Yes, but Motorola chips use an internal frequency doubler, where
>Intel chips don't. So, a 40 MHz clock speed 68040 is actually running
>internally at 80MHz, where a 66MHz clock speed 80486 is actually running
>internally at 66MHz.
>
> Also, MHz is meaningless. It's like saying, "My car runs at 7000
>rpm." Well, fine, but until you know about the transmission, the shocks,
>and the gear ratios, that doesn't tell you beans about how fast the car
>will go. That analogy holds more with computers, where manufacturers have
>been known to cripple high-speed components by interfacing them with low
>speed components.
I really wish that this myth would die. It is sort of like saying your
your car runs at 14,000 rpm because some idler gear in the valve train
spins at 14,000 rpm while the crankshaft only turns at 7,000.
A 66 MHz 486 can retire as many as 66 million integer instructions a
second, a 40 MHz 040 can only retire 40 million integer instructions a
second, not 80.
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
Workstation- A high-performance microcomputer designed to run benchmarks.
Erik A. Speckman
>espe...@reed.edu writes:
>>Within 3 years all of the 68k based machines will probably be too out of
>>date for most of the latest software because even if they sell 040
>machines
>>for the next year and a half they aren't getting much faster.
>
>And why not?? I understand that the chip, going from '30 to '40, makes
>the biggest difference in performance. But the fastest chip out is only
>the 40Mhz in the Quadra 800. (and the AV I think) There are 486DX2
>chips out there that do 66Mhz, in addition I heard that the DX2
>architecutre supports the addition of another chip, doubling performance.
> Can Motorola squeeze some more juice out of the '40 before they have to
>retire it??? Say something like 70Mhz....
There are two issues here.
First, is motorola going to make faster chips. I don't think so. I
don't think they have even delivered small production quantities of a 50
MHz 040 and thier future lies with the PowerPC.
Second, I don't think it is worth apple while to do much reingineering on
the 040 machines to use them with faster chips. Perhaps if they ran a 50
or 66 MHz chip on a half speed system bus they could get away with it
cheaply.
Personally, I don't think it is in apples interest to bring out 040
machines that compete with the PowerPC machines. There will be a market
for new fast 040s after the PowerPC machines ship but I think it is a
niche market and better left to accelerator card makers.
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
Workstation- A high-performance microcomputer designed to run benchmarks.
Charlie Kuehmann
> Seriously though, is a 620 "for the rest of us" going to be an option? Or is
> it going to be so expensive that only rocket scientist (RS) dweebs will use
> them in a workstation architecture? Will it be offscale compared to the 604
> end of the line? Will competition and technology be tempting us with other
> kinds of improvements to those same 604 boxes, leaving the 620s to the RS's?
> Or, will AV and other technologies be clamoring for 620 power, prompting
> Apple to put out a middle-of-the-line 620 box at a reachable price?
As one of those "RS dweebs" you mention above I beg to remind you that with
recent events and the demise of the cold war it is getting harder and
harder for us to afford these things ourselves. :^)
--
Charles Kuehmann
Northwestern University
Steel Research Group
kueh...@nwu.edu
Mr_G._Landweber_student_tel_2-73550
>there code to gain the advantage of the new PPS macs. They probably
>will not develop the code for 'old' macs . In the end we all have
>cheap old macs but no software ( except what exsists now )
Actually, it doesn't take much work to rewrite code to work on the PowerPC
Macs. Then, if using Apple's universal header files, developers can compile
the same source code for use on both 680x0 and PowerPC platforms (as well as
whatever new ones Apple comes up with in the future).
BTW, look around at the number of developers who still support System 6.
There are quite a few considering that System 7 has been around for about
three years. And why? Because a large number of people haven't bothered
to upgrade. As long as there is a market, developers will sell software.
-- Greg
Ipokratis Papaioanu
: At least in this aspect, Pentium is far superior. It runs all the existing
: software much faster than the 486 machines. I wish Power PC Mac's did the
: same!
Not true. Much faster is not the case with Pentium. It runs existing
Win3.1 applications <75% faster, not even twice as fast (PCWorld January
benchmarks). You need to recomplile you applications for Pentium to see
a doubling in performance. That's exactly the same with the Mac applications
needing to be recompiled for PPC (and in that case 2-4 times faster)
PS keep in mind that most Win apps are still 16bit
: --------------
: Sridhar Mahadevan
: Tampa, Florida 33647
: (813)974-3260
: maha...@guardian.csee.usf.edu
--
Paki Papaioanu
Boston University My words are my own and they speak for no one
School of Management but me
Computer Support Services
Rick Worley
slansk...@macmail1.cig.mot.com (Doug Slansky) wrote:
>
>As to one of the other posters, I don't agree that the 040 will survive
>longer than the '030. The way I see it, margins are killing Apple right
>now, and to reduce the 040 prices further for anything other than clearance
>will put them in a loss mode. The 030's, on the other hand could still
>sell to the education and low-end home market at significant discounts to
>the PPC, while still eeking out a (meager) profit. Unless Apple goes into
>NEXTmode, and only supplies high-end gear (unlikely based upon their
>history in the edu market), I think they will continue to ship low cost 030
>machines even after the 040 is gone...
>
>Again, I could be (and frequently am) wrong....
>
>-doug
>--
>Doug Slansky
>Motorola Cellular Infrastructure Group
>Advanced Products Division
>AKA: slvrs...@aol.com
hogwash
The 030 will soon be replaced by the 040 just like the 286/386 has been
replaced by 486. The 68LP040 should replace the 030 in PowerBooks. When
386 clones appeared, Intel priced the 486SX low enought to kill off the 386.
The Pentium will have trouble surpassing 486 due to high cost. The
following are prices from EETimes, while somewhat current, they probably do
not reflect current market prices. The 486SX and I assume the 68LC040 are
both <$100.
Motorola
68040 25 MHz $245
68040 33 MHz $290
68040 40 MHz $393
IBM
PowerPC 601 50 $280
PowerPC 601 66 $374
PowerPC 601 80 $500
Intel
386DX 33MHz $119
486SX 20MHz<$100
486SX 25MHz $119
486DX 33MHz $300
486DX 50MHz $502
486DX2 33/66$550
Pentium 66 $850
The 68040 and 80486 have ~1.2M transistors, PPC601 and Pentium has ~2.8M
The 040 will survive, a real 040 is faster than a PPC emulating 680x0 code,
and the last time I looked all my Mac applications are 680x0 code. Until
there is native PPC applications there is no reason to buy one.
Rick Worley
WL/ELM BLDG 620
2241 AVIONICS CIRCLE SUITE 25
WRIGHT-PATTERSON AFB OH 45433-7327
Tel: (513) 255-7665
Fax: (513) 476-4807
wor...@el.wpafb.af.mil
Brendan Mahony
What about to mooted 68060 machines?
James Chuang
I've been following this thread, and one things that stinks in my mind
is that people keep on referring to the low margins on the 68040 machines
as a factor.
But couldn't Motorola lower their price? I mean, if they start selling
68040s for $50 to apple, Apple will have a lot of margin to play with.
It would be in Motorola's best interest, because every buyer of a 040 mac
is a potential buyer of a Power PC upgrade.... And the more Motorola
chips are out there, the more ammo they have against Intel...
jamesc
--
=========================================
If someone asks if you are a God, you say... YES!
H S Kim
>But couldn't Motorola lower their price? I mean, if they start selling
>68040s for $50 to apple, Apple will have a lot of margin to play with.
>It would be in Motorola's best interest, because every buyer of a 040 mac
>is a potential buyer of a Power PC upgrade.... And the more Motorola
>chips are out there, the more ammo they have against Intel...
Hmmm, let's see. The potential market for the PPC are IBMs, Apples, and
every IBM and Apple wannabe. Don't think Motorola needs any more ammo.
I've read however that they will release a 60+ mhz version of the 040.
--
-- Michael Kim
-- hs...@uxa.cso.uiuc.edu
Mr_G._Landweber_student_tel_2-73550
Heard of the P6? What about the P7? The answer is that you make it
RISC inside while preserving compatibility at the instruction set level.
Sounds like what Apple is doing with the 680x0 emulator on the PowerPC
(except that in the case you mention, it is done at the level of the chip).
-- Greg
Erik A. Speckman
<103t_e...@west.cscwc.pima.edu> wrote:
>
>> I really wish that this myth would die. It is sort of like saying your
>> your car runs at 14,000 rpm because some idler gear in the valve train
>> spins at 14,000 rpm while the crankshaft only turns at 7,000.
>>
>> A 66 MHz 486 can retire as many as 66 million integer instructions a
>> second, a 40 MHz 040 can only retire 40 million integer instructions a
>> second, not 80.
>> Erik Speckman espe...@romulus.reed.edu GBDS
>
>SPECint92 of 21; SPECpf92 of 15 (IEEE Spectrum, December '93)
>
>According to PCWeek, an 80486DX2/66 has:
>SPECint92 of 32.2; SPECfp92 of13.1 (PCWeek, Jan 10, 1994, pages 1 & 16)
>
>Multiply the 25mhz 68040 by 1.6 to get a 40mhz 68040 (yes, I know this isn't
>accurate, but we are looking for ballpark here) and you get:
>SPECint92 of 33; SPECfp92 of 24.
How, exactly, does this refute my statement that the 040 isn't clock
doubled and the 486 DX/2 is?
All this suggests is that the 040 is more efficient thatn the 486 DX2.
But even this is obscured by possible differences in system
implimentation, most likely the size of the L2 cache.
>BTW, simple arithmatic suggests that a clock-doubled 80486DX2/66 is running off
>of a 33mhz data bus. Care to tell me how it is retiring 2 integer instructions
>for every data-bus access?
By implementing such obscure concepts as registers, an on chip cache and
write buffering that have been used on mainframes since the 1960s and on
microprocessors for quite some time.
>Lawson (I can add, but I don't got no stinkin' degree, sorry)
No need to apologise. Try a BA in biology, it works for me.
Eric Smith
> My understanding is that they are going to go RISC, and that
>the Pentium is the first step in that direction. They are just going
>to try and do it in a more backwards compatable way, than Motorolla
>(personally I think backwards compatability is a big trap, but that's
>me).
Yes of course it's a trap in one sense, because you end up having to
support old technology. But it's a much bigger risk (pun intended) to
not be backwards compatible and cut loose your huge customer base, and
hope they'll all return to you. It's like the analogy of the store owner
who wants to put up a different display, but first he tells his customers
that they have to go outside for five minutes and then come back in.
It's likely that once outside at least half will go to some other store. :-)
- Eric
Eric Smith
>I understand that the chip, going from '30 to '40, makes
>the biggest difference in performance. But the fastest chip out is only
>the 40Mhz in the Quadra 800. (and the AV I think)
The 800 is 33 MHz. Only the 840av is 40 MHz.
>There are 486DX2
>chips out there that do 66Mhz, in addition I heard that the DX2
>architecutre supports the addition of another chip, doubling performance.
> Can Motorola squeeze some more juice out of the '40 before they have to
>retire it??? Say something like 70Mhz....
This sounds like the old "my MHz is bigger than your MHz" argument.
To paraphrase the old comeback, it's not the size of the MHz, it's
what you *do* with it that counts.
>Anyway, has anyone heard of Intel's plans for their next generation of
>chips?? Can't see how they can go beyond a 586 other than to go RISC.
Heard of the P6? What about the P7? The answer is that you make it
RISC inside while preserving compatibility at the instruction set level.
-----
103t_e...@west.cscwc.pima.edu
> In article <2i7ra3$c...@scratchy.reed.edu>, espe...@reed.edu (Erik A.
> Speckman) wrote:
>
>
>> Within 3 years all of the 68k based machines will probably be too out of
>> date for most of the latest software because even if they sell 040 machines
>> for the next year and a half they aren't getting much faster.
>>
>
> I seriously doubt this scenario. People do not upgrade computers every year
> (much as we would want them to). 3 years down the road, plenty of homes and
> schools will continue to have millions of 68K machines, and any vendor who
> does not ensure that his/her program runs on the old machines will be
> foolishly throwing away a big chunk of profits. (Heck, there are plenty of
> the old Mac Classics around still, and those are close to 10 years old!).
>
>
> Sridhar Mahadevan
> Department of Computer Science
> University of South Florida
Try this for a scenario: by the end of the year, the 3 watt MPC603 will be
shipping. DayStar, or some other enterprising accelerator/clone maker, will
introduce an "LC v-like" upgrade board for ALL compact Macs (128k, 512k, Plus,
SE, SE/30, Classic, Classic II) with a SCSI II port, built-in video supporting
thousands or even millions of collors, etc.
Total price: under $1000.
How many do you think they will sell? After 2 years, assuming that the above
upgrade is introduced, I predict that the only non-upgraded Macs in the USA
will be aquariums (DayStar explicitly excludes them from their advertising).
Lawson
103t_e...@west.cscwc.pima.edu
> Question: Is the 68040 chip more/less expensive than a PPC?
> --
> Tom Gooding
> tgoo...@iastate.edu
> Iowa State University
An MPC601 is more expensive to buy and use than an Mc68040. An MPC603 is less
expensive to buy and use than an Mc68040.
THe "and use" part refers to the overhead from power consumption... The 64-bit
bus costs more, but not enough to enter the equation, in my
less-than-informed opinion...
Lawson
Andrew Geweke
Speckman) wrote:
> How, exactly, does this refute my statement that the 040 isn't clock
> doubled and the 486 DX/2 is?
It doesn't. But your statement that an 80486DX2/66 can retire 66 million
integer instructions per second, versus 40 million for the 68040-40, while
true, is utterly meaningless. This _never_ actually happens, and if it did,
nobody would care. In most loops, improving branching algorithms helps a
heck of a lot more.
As people have said before: Clock speeds are meaningless.
The 68030-50 can retire 50 million integer instructions per second --
theoretically. So how can you explain that it's slower than the 68040-33,
let along the 68040-40?
> All this suggests is that the 040 is more efficient thatn the 486 DX2.
> But even this is obscured by possible differences in system
> implimentation, most likely the size of the L2 cache.
Er...if I remember correctly, nearly every 80486-based machine has a
sizeable L2 cache. 68040-based machines generally don't, and they still
compete head-to-head with 486s. This would tend to accentuate, not obscure,
the fact that the 040 is more efficient. If we ever got a 68040-66, it
would _fly_.
> Workstation- A high-performance microcomputer designed to run benchmarks.
PC- A high-performance microcomputer that you can't actually do anything
with.
--
Andrew Geweke / Computer Engineering, Michigan State University
Delphi: Single biggest source of net.idiots today. Write to help.
"Love is a temple / Love the higher law" -- U2, _One_
GAT/GT/GE d? -p+ c+ !l u e*(+) m+ s+++/ n+ h f g+(-) w+(-) t+ r- y?
103t_e...@west.cscwc.pima.edu
>
> I really wish that this myth would die. It is sort of like saying your
> your car runs at 14,000 rpm because some idler gear in the valve train
> spins at 14,000 rpm while the crankshaft only turns at 7,000.
>
> A 66 MHz 486 can retire as many as 66 million integer instructions a
> second, a 40 MHz 040 can only retire 40 million integer instructions a
> second, not 80.
>
> --
> ____________________________________________________________________________
> Erik Speckman espe...@romulus.reed.edu GBDS
>
> Workstation- A high-performance microcomputer designed to run benchmarks.
According to IEEE Spectrum, a 25mhz 68040 has:
SPECint92 of 21; SPECpf92 of 15 (IEEE Spectrum, December '93)
According to PCWeek, an 80486DX2/66 has:
SPECint92 of 32.2; SPECfp92 of13.1 (PCWeek, Jan 10, 1994, pages 1 & 16)
Multiply the 25mhz 68040 by 1.6 to get a 40mhz 68040 (yes, I know this isn't
accurate, but we are looking for ballpark here) and you get:
SPECint92 of 33; SPECfp92 of 24.
BTW, simple arithmatic suggests that a clock-doubled 80486DX2/66 is running off
of a 33mhz data bus. Care to tell me how it is retiring 2 integer instructions
for every data-bus access?
Lawson (I can add, but I don't got no stinkin' degree, sorry)
Erik A. Speckman
Andrew Geweke <gewe...@student.msu.edu> wrote:
>In article <2ikp43$o...@scratchy.reed.edu>, espe...@reed.edu (Erik A.
>Speckman) wrote:
>
>> How, exactly, does this refute my statement that the 040 isn't clock
>> doubled and the 486 DX/2 is?
>
>It doesn't. But your statement that an 80486DX2/66 can retire 66 million
>integer instructions per second, versus 40 million for the 68040-40, while
>true, is utterly meaningless. This _never_ actually happens, and if it did,
>nobody would care. In most loops, improving branching algorithms helps a
>heck of a lot more.
I was only addressing the rather annoying and persistant rumor that the
040 line is "actually clock doubled". Of course, the DX/2 66 rarely
retires 66 million instructions per second, but then the 40 MHz 040 rarely
retires 40 million instructions per second. I was using an extreme case
to highlight the difference between clock-doubled and NOT CLOCK DOUBLED.
>As people have said before: Clock speeds are meaningless.
>The 68030-50 can retire 50 million integer instructions per second --
>theoretically. So how can you explain that it's slower than the 68040-33,
>let along the 68040-40
It wasn't designed to retire one instruction per clock. It is not a scalar
processor. It has a shallow pipeline. I don't even know if their is any
theoretical case where the 030 can retire one instruction per clock.
>> All this suggests is that the 040 is more efficient thatn the 486 DX2.
>> But even this is obscured by possible differences in system
>> implimentation, most likely the size of the L2 cache.
>
>Er...if I remember correctly, nearly every 80486-based machine has a
>sizeable L2 cache. 68040-based machines generally don't, and they still
>compete head-to-head with 486s. This would tend to accentuate, not obscure,
>the fact that the 040 is more efficient. If we ever got a 68040-66, it
>would _fly_.
Er...if I remember correctly you can't run the SPEC suite on the MacOS.
So these benchmarks were probably run under some version of UNIX. I
doubt it was A/IX. It was probably run on a workstation of some sort and
probably had an L2 cache.
I don't doubt that te 040 is a more "efficient" chip. I just doubt that
the data presented supports this conclusion.
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
Workstation- A high-performance microcomputer designed to run benchmarks.
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
Workstation- A high-performance microcomputer designed to run benchmarks.
Piner
By the way, you cannot use MHz to compare two different models/brands of
chips in terms of performance. The 40MHz 68040 in the 840AV is about the
same speed in terms of raw power as a 486DX2 at 66MHz.
No, Motorola, won't be squeezing any more juice out of the 68040. The
PowerPC chips are going to be cost-effective come this fall with the 603,
so there will be no use for the 68040 beyond 40 MHz. Why build chips that
cost the same or more money that have less performance?
Yes, certain 486 motherboard have an Overdrive socket so you can add another
486 that is clock doubled. This would be like if you are upgrading a 486/25
or 486/33 and want to go to a 486DX2/50 or 486DX2/66. There is talk about
being able to use BOTH chips at the same time by the operating system
(Windows NT), instead of the new one replacing the old, but who knows if
this will ever become a reality.
Supposedly, Intel's P6 (the Pentium is a P5) is supposed to be a transition
chip, with the P7 being fully RISC. Obviously, software companies would
have to recompile to take full advantage of the P7, with older software
being emulated, just as they are doing with the PowerPC. But, if the PowerPC
is well-established by then, runs many operating systems and environments,
is made by many companies, and is cheaper and more powerful than the P7,
Intel is going to have trouble. On the other hand, if the P7 can challenge
the PowerPC on all technical fronts as well as price, and has lots of industry
support, it COULD become the second popular RISC platform (as I believe
PowerPC will be the first widespread RISC chip used in standardized hardware).
Time will tell.
-Brian
Mats Ekberg
--In article <ericsCK...@netcom.com> er...@netcom.com (Eric Smith) writes:
-- Heard of the P6? What about the P7? The answer is that you make it
-- RISC inside while preserving compatibility at the instruction set level.
The idea of using RISC in stead of CISC is the change in architecture,
if you preserve the CISC compatibility at the instruction level nothing is gained!
John L. Coolidge
>6) The PowerPC accounts for nearly 50% of Macs sold in 1994, perhaps 80-85%
>of Macs in '95 and 100% of Macs in 1996. Keep in mind that the 603 is
>designed to be cheap, as in sub-$100 by the time it hits full production.
>The systems designed around PowerPC will be cheap, too, save the slightly
>higher DRAM and hard disk costs. But the latter two are going to drop
>appreciably in 1994.
While I agree with much of Mark Rogowsky's post, these numbers don't
work for 1994. The current run rate for Mac sales seems to be
1M/quarter, give or take a bit. Even assuming a couple lean quarters,
that's around 3M Macs sold in 1994. The "1 million PowerPC
Macintoshes in the first year" statement would make, at best, 33% of
Macs sold in 1994 PowerPC-based.
Of course, either Apple could be planning on selling a lot more than
1M or the bottom could drop out of the Mac market. But 50% PowerPC
Macs in 1994 sounds a bit high (although I for one wouldn't be upset :-)).
--John
Thanks to the diligence of the FBI, this particular vacuum
cleaner won't fall into enemy hands.
-- Howard Hughes, _The_Rocketeer_
+++John L. Coolidge++++++++...@apple.com+++++++++++++++++++++++
I speak for myself, not for Apple Computer. Copyright 1993 John L. Coolidge.
Copying allowed only if attributed, and if all copies may be further copied.
Piner
>
>> slansk...@macmail1.cig.mot.com (Doug Slansky) writes:
>>
>> >Any netland speculation on the end of the 030/040 product line from
>> >Macintosh? The reason I bring this up is twice now I have seen
>> >"unofficial" pricing on the base PowerMac 6100-60 in MacWeek to be $2000
>> >INCLUDING monitor and keyboard. This figure seems quite low, given that
>> >high end 040 systems are selling for quite a bit more than this (in similar
>> >configurations). Even the least expensive 040 is around $1300 including
>> >monitor and keyboard.
>>
>so what if Apple continues to sell 'old' macs Developers must rewrite
>there code to gain the advantage of the new PPS macs. They probably
>will not develop the code for 'old' macs . In the end we all have
>cheap old macs but no software ( except what exsists now )
>
>the old macs Sorry !
>--
>Tax ?? what TAX ?? Oh TAXI !!
>----------------------------------------------------------------------
>Arne Maletzky m...@mh.nl via internet backbones
>Multihouse Automatisering uucp: ..!{uunet,sun4nl}!mh.nl!mlz
>Doesburgweg 7 Gouda, The Netherlands
>----------------------------------------------------------------------
> Say Cheese and smile again - say smile and.... Cheese again?!?!?
Sorry. Software WILL continue to be available for the old Macs for at
least several more years. The way it will work is that most Mac software
packages will include both PPC and regular Mac versions. You run an
installer than will install a program so that it will run on an old Mac
only, a PPC Mac only, or BOTH a PPC Mac and old Mac.
-Brian
-Remember, if you don't like something about Macs, you always have the
option of getting a PC. :)
John F. Woods
>Anyway, has anyone heard of Intel's plans for their next generation of
>chips?? Can't see how they can go beyond a 586 other than to go RISC.
According to the latest EE Times, they are starting to drop rumors about
the P6 (Psexium? :-) which they threaten to be 3 times the speed of the
Pentium, to ship next year.
John F. Woods
>The
>following are prices from EETimes, while somewhat current, they probably do
>not reflect current market prices.
>PowerPC 601 66 $374
>Pentium 66 $850
A-HAH! Motorola LIED! The 601 is NOT half the cost of a Pentium!
It is CHEAPER! :-)
Note also that the 601/66 is also cheaper than the 486DX2/66, whose performance
it exceeds by a good margin. (Especially since the DX2/66 will be placed in
a 33MHz system, which in many real-world cases is really a 25MHz motherboard
with mandatory wait states.) This coming year should be a good time for
watching DOS lemmings squirm.
John F. Woods
>THe "and use" part refers to the overhead from power consumption... The 64-bit
>bus costs more, but not enough to enter the equation, in my
>less-than-informed opinion...
Well, board space is extremely expensive, believe it or not. The
manufacturing costs are not negligible, but more important is the fact that
you have a limited amount of real-estate, and every square inch taken up by
wires is a square inch not covered with an IC; getting the same functionality
then requires bigger and bigger ASICs, with lower yields and higher development
costs.
Andrew Geweke
(Mr_G._Landweber_student_tel_2-73550) wrote:
It always seemed to me that this was the whole bit about the RISC/CISC
debate. The whole advantage of RISC -- or so I've heard -- is that the
compiler does a lot of the work that the chip did before.
I'm just a sophomore CE student, so correct me if I'm wrong. But:
The "microcode" in a chip refers to the transistors on the chip that turn
the instructions it receives into instructions it can actually execute --
for instance, there's not hardware to _directly_ support all these weird
CISC addressing modes, instructions, and so forth. Rather, you send the
instruction, and the microcode turns it into several smaller instructions.
The microcode determines what the chip can do -- for example, Cyrix's 386
isn't identical to Intel's, but the microcode is. Thus it is compatible.
CISC chips typically have a large microcode. I read someplace that ought to
know what they're talking about -- I forget where, sorry -- that the
Pentium's microcode sprawls across three-quarters of the chip. The internal
engine they call RISC.
If you took the Pentium's internal engine out of the rest of the chip, it
would be a RISC chip. It also wouldn't run 80x86 programs. So they have the
microcode. And microcode does _not_ execute infinitely fast; that's what's
retarding the chip's progress.
RISC chips, OTOH, have very little -- if any -- microcode. Rather, it's the
compiler that turns complex instructions into very small ones that the
engine (integer unit, FP unit, BP unit, etc.) can understand directly.
That's why they have the potential to be so much faster. That's also why
the PPC 601 has 1.2 million transistors, less than HALF of what the Pentium
has. (Someone here stupidly said they were equal).
Now, there are other differences, too, I'm sure of that. But the P6 and P7,
to go faster, are going to have to cut out the microcode. And to do that,
you're going to have to make a complete binary-compatibility break with the
80x86 instruction set. Apple's doing it earlier, rather than later -- IMHO,
a _very_ wise decision. The PPC 601's 1.2M transistors are accounted for in
a large part by its (relatively quite large) cache. The Pentium's 2.8M are
mostly microcode.
Some of you may be right, but many people seem to think that the Pentium is
RISC. As far as I can tell, it's no more RISC than the 80486, 80386, 68040,
68030, and so far were.
And for those of you who dispute Motorola's schedule for PPC releases: we
can dispute Intel's schedule back, too, as well as their performance
claims. IMO, given RISC architecture's phenomenal speed growth recently and
CISC's relatively small one, plus Pentium's lagged schedule and PPC's
on-time (if not better) one, Motorola/IBM/Apple are easier to believe.
NOTE THE FOLLOWUP-TO COMP.SYS.MAC.ADVOCACY.
--
Andrew Geweke / Computer Engineering, Michigan State University
Andrew Geweke
Ekberg) wrote:
Thank you. This is what I've been trying to say all along.
You simply cannot switch from CISC to RISC without killing compatibility.
--
Andrew Geweke / Computer Engineering, Michigan State University
Delphi: Single biggest source of net.idiots today. Write to help.
Chuck Browne
: In article <mahadeva-2...@131.247.2.65>, maha...@csee.usf.edu (Sridhar Mahadevan) writes:
: Try this for a scenario: by the end of the year, the 3 watt MPC603 will be
: shipping. DayStar, or some other enterprising accelerator/clone maker, will
: introduce an "LC v-like" upgrade board for ALL compact Macs (128k, 512k, Plus,
^^^^
A 128k???? I have one of those! If Daystar came out with a card (motherboard)
that would fit into a 128k mac, I'd certainly buy it, for the irony of it all...
Can you imagine, the original "never upgradable, even memory" mac, suddenly
50 million times faster...
: SE, SE/30, Classic, Classic II) with a SCSI II port, built-in video supporting
: thousands or even millions of collors, etc.
: Total price: under $1000.
Now, what can Daystar do with my Commodore 64 (1Mhz 6510, 64k memory)?
Chuck
: How many do you think they will sell? After 2 years, assuming that the above
Eric Smith
>er...@netcom.com (Eric Smith) writes:
> Heard of the P6? What about the P7? The answer is that you make it
> RISC inside while preserving compatibility at the instruction set level.
>Sounds like what Apple is doing with the 680x0 emulator on the PowerPC
>(except that in the case you mention, it is done at the level of the chip).
Yes, but since there will be an actual hardware compatibility mode older
programs will run even faster because of the increased CPU performance,
instead of slower as in the PowerPC Macs, where the 680x0 instruction set
has to be emulated in software.
John L. Coolidge
Sure, but there's an inherent problem with compatibility modes:
people use them! The big problem Intel faces is the installed base
problem: there are A LOT of 80x86-based PC's out there. All of them
pretty much run the same programs (well, that's no longer true of
8086-based systems, and won't be true for 80286 systems much longer).
Developers like this because their program runs on lots of machines
and so they have a huge potential market. System software vendors like
this because they don't have to do multiple OS versions.
Now suppose Intel, in their hypothetical Px (where x>6) chip, produces
the kind of hybrid chip that people are talking about -- a RISC chip
with its own instruction set that also has a compatibility mode for
80x86 instructions. Intel's pitch is, no doubt, to tell developers
that they will get massive speedups by going to the new RISC
instruction set.
However, by doing that they face exactly the same problem that Mac
developers face -- their new programs no longer run on the installed
base of machines, only on machines with the Px chip. To get backwards
compatibility, they need to do "fat apps" or stock two versions. To be
able to do partial ports (only recompiling speed-critical code) they
need something like the Mixed Mode Manager. In short, they need
_exactly_ the same infrastructure that Apple has built, except that
they don't need a processor emulator.
Wait a minute - sure they do. That part of the chip devoted to 80x86
instruction processing sure sounds like an emulator to me, just one
pushed into hardware. This means that the chip uses a lot more
transistors than it could have to get the same performance levels for
a single-mode processor (RISC or CISC). It also means that, if you
had devoted those transistors to cache, say, or another integer unit,
or a better branch processor, you might have made a real screamer. So
there goes those "massive speedups" developers were told to expect...
What kind of message does all this send? By making a single clear-cut
transition statement ("The Future of Macintosh is PowerPC"), Apple
makes it clear that backward compatibility is a temporary thing. The
implied message is that important developers will offer native-mode
solutions quickly, so that if you need native speed you'll have it.
By focusing the work of the chip designers on building a new, clean,
fast architecture and not wasting transistors on compatibility modes,
PowerPC systems are as fast as possible at a given price-point.
On the other hand, this hybrid Px doesn't give a clear message. It in
effect says that Intel isn't very comfortable that important
developers will adopt the new RISC instruction set, so you'll need to
be able to run all your old apps quickly. Since the 80x86 core will
most likely be faster than previous 80x86's (if not, why bother
putting it in silicon -- you'd probably do just as well with a
software emulator), there will be a lot of pull on developers to just
do an 80x86 version -- there's that huge installed base to serve, and
Intel sure doesn't look convinced about this RISC stuff. This hybrid
Px will be slower than other RISC processors since it's blowing its
transistor budget on compatibility stuff. Either that, or it'll be a
gigantic (read: expensive) chip. And the kicker is that, even with the
RISC core, it'll still probably be a slower 80x86 than a pure 80x86
knockoff from Cyrix, AMD, or one of the other cloners. They can use
the same process technology and the same design techniques, but they
don't have to blow their transistor budget on the RISC instruction set
decoder or the CISC/RISC translator.
The point is that hardware isn't free. The silicon used to provide
backward compatibility on this Px could be used to make it a faster
RISC chip instead; the silicon used to provide RISC decode could be
used to make it a faster CISC chip. Instead, you've got a chip that
fails to be the best at either. Why would developers want to port to a
slower RISC chip with no installed base instead of the (by then)
several years mature PowerPC with a decent installed base? Especially
when one vendor is 100% behind their RISC product and the other is
hedging their bets?
I'm not discounting Intel -- I think they're a very formidable
competitor, if for no other reason than they can (and do) outspend
everyone else in R&D. But I think the hybrid RISC/CISC-does-both Px
won't fly, and I don't expect Intel to try it. It might have flown
several years ago (before the 486 came out); it would be a disaster
now.
--John
Every program has at least one bug and can be shortened by
at least one instruction -- from which, by induction, one
can deduce that every program can be reduced to one instruction
which doesn't work. -- found in fortune
+++John L. Coolidge++++++++...@apple.com+++++++++++++++++++++++
I speak for myself, not for Apple Computer. Copyright 1994 John L. Coolidge.
Eric Smith
>--er...@netcom.com (Eric Smith) writes:
>-- Heard of the P6? What about the P7? The answer is that you make it
>-- RISC inside while preserving compatibility at the instruction set level.
>The idea of using RISC in stead of CISC is the change in architecture,
>if you preserve the CISC compatibility at the instruction level nothing is gained!
No, not "nothing" is gained. Compatibility is gained, and that's the
important issue. Of course you don't get the full benefits of the
underlying architecture, but since the emulation is in hardware and
the CPU is much faster than the emulated CPU (runs at higher MHz, has
internal optimizations and large on-chip caches), applications will
run faster without modification, unlike the PowerPC Macs where the
emulation is done in software.
But that's only for compatibility mode. The real gains are of course in
programming the chip in RISC mode, which is a different API. (I wasn't
saying that the RISC level of the chip wouldn't be available to the user.)
The original poster was asking what Intel is doing. That's what they're
doing.
Castor Fu
It strikes me as strange to see all these people looking forward to
cheap 68K based macintoshes when power PC comes out.
If Apple wants to phase out the 68K line, rather than dump the remaining
machines on the market at low prices, I would think they should instead
RAISE the prices for 68K machines (to pay for the higher inventory costs), and
continue production (at a lower rate) for those using macs in
critical applications.
--
Castor Fu, (f...@leland.stanford.edu)
103t_e...@west.cscwc.pima.edu
According to last years Wall St. Journal, prices for a single chip, in lots of
1000, are:
66mhz Pentium: $ 650
80mhz MPC601: $ 550
66mhz MPC601: $ 450 (from Motorola ads)
The benchmarks for the PowerPC 601 are on a half speed (33mhz for 66mhz) data
bus. An Intel engineer informs me that I am incorrect about the needs of the
Pentium, so I won't comment on the system needs of one.
Lawson
Andrew Geweke
wrote:
Right. But look at the Pentium. Assuming its cache (half the size of the
601's) is as transistor-efficient as the 601's, it's taking up 0.8 million
transistors. This means the chip itself is 2.3 million transistors. The 601
is only 2.1 million _with_ the cache; the Pentium could thus make a RISC
engine with a _huge_ cache with the number of transistors it has.
In other words, for a given size/price, the CISC compatibility is going to
cost you a _lot_ of performance. Or, for a given performance level, the
CISC-compatible chip will cost you a lot more and generate a lot more heat.
The Pentium is behemoth. Keeping the CISC compatibility will just cost you
more.
I think Apple's approach is better. Someone here posted a very valid
argument about why you'll never get full performance out of a chip with the
hardware compatibility -- why bother with fat binaries, RISC compiling, and
all that junk, when you can still run faster than the last chip (though not
nearly as fast as possible) by coding 80x86 code still? Apple's given all
programmers and developers a kick in the butt to get them moving. Its
software emulation is very impressive, and Apple's guaranteeing that we'll
all be RISC soon.
Quinn
writes:
> Yes, but since there will be an actual hardware compatibility mode older
> programs will run even faster because of the increased CPU performance,
> instead of slower as in the PowerPC Macs, where the 680x0 instruction
set
> has to be emulated in software.
Not necessarily. A lot of the speed of the Mac emulator comes from
trapping out system software calls and making those calls run native.
Under DOS there aren't any system software calls worth trapping so
running DOS programs is really going to stress the emulator.
How much that balances the gains made by putting the emulator in
hardware is an open question.
btw another advantage of a software emulator is that it makes it
*much* easier to deal with bugs (:
--
Quinn "The Eskimo!" <qu...@cs.uwa.edu.au> "Support HAVOC!"
Department of Computer Science, The University of Western Australia
Eric Smith
>j...@ksr.com (John F. Woods) >wrote:
>> Chuck Simciak <sim...@ccsmtp.ccf.org> writes:
>Right, they still haven't shipped the P54C.... which is now due second
>quarter. They still haven't shipped the 66MHz P5 in quanity... which was
>due six months ago...
>But, they're going to ship the P6 next year.
Well, don't forget that the Pentium and P6 (and P7 for that matter) are
completely separate development projects. And the release date for the
P6 was at one time pushed *back* in order to give the 486 and the Pentium
longer life spans.
Eric Smith
>er...@netcom.com (Eric Smith) writes:
> Quinn <qu...@cs.uwa.edu.au> writes:
> >Not necessarily. A lot of the speed of the Mac emulator comes from
> >trapping out system software calls and making those calls run native.
>This is mostly true, but the connotation is wrong. First of all, the
>emulator is a 680x0 emulator, not a Mac emulator. This new machine is
>a Macintosh, and it supports the same toolbox and system software.
>Secondly, the emulator is not "trapping out" toolbox calls. "Traps"
>describes the mechanism by which programs on the Mac call toolbox
>routines. On the PowerPC Macs, many of these routines are written in
>native PowerPC code. The good bit is that a 680x0 program that calls
>these routines will get the speed boost.
> Well, there's that! But also on the Mac you really only have to worry
> about one software emulator - for the Mac OS.
>The emulator is not emulating the Mac OS. It is emulating the 680x0.
>The Mac OS is in ROM just like in a 680x0 Macintosh, except that now
>much of it is in native PowerPC code.
Sorry I wasn't clearer - it doesn't emulate the Mac OS, but it runs
under the Mac OS.
> (I guess there's A/UX,
> but I haven't heard anything about emulation for applications that
> run under that. In fact I don't even think A/UX will run on my 840av,
> and it's not even a new instruction set.)
>I don't believe that A/UX will run on the PowerMacs.
That's exactly my point; they only have to make the emulator run under
one system.
John L. Coolidge
>cool...@sirius.lansys.apple.com (John L. Coolidge) wrote:
>> While I agree with much of Mark Rogowsky's post, these numbers don't
>> work for 1994. The current run rate for Mac sales seems to be
>> 1M/quarter, give or take a bit.
>>
>that in Q1/94 as PowerPC paralyzes the market. When Apple quoted one
>million, the IDC and Apple projections for sales of Macs in 1993 were about
>two million. I'm stunned, it went up so much in Q4... perhaps 3 million
>Macs were sold last year.
You're completely right; I was misremembering 1992 numbers. Mac sales
were around 2.25 million in 1993, I believe.
>I think that the PowerPC sales will exceed 1 million easily in 1994 if the
>603-based Powerbooks reach the market in September... They could blow past
>1 million if the Q605/LC475 also goes PPC at that time.
I'll agree with this (hopefully not just wishful thinking -- I'd like
the profit sharing checks :-)). Of course, the Q605/LC475 upgrade has
been announced now, I think...
--John
By analysis of usenet source, the hardest part of C to use
is the comment.
-- Bill Davidsen (davi...@crd.ge.com)
+++John L. Coolidge++++++++...@apple.com+++++++++++++++++++++++
I speak for myself, not for Apple Computer. Copyright 1994 John L. Coolidge.
sea...@ac.dal.ca
> no...@apple.com (Noah Price) writes:
>
>>er...@netcom.com (Eric Smith) wrote:
>
> [ deleting my severely edited commentary ]
>
>>> Apple, for obvious reasons, is not
>>> marketing the initial PPC Macs as an interim product line but it seems
>>> obvious that that's the case.
>
>>C'mon, if you look at things that way _every_ computer is part of an
>>interim product line. Last years model is always proven, and there will
>>_always_ be a better/faster/cheaper computer out in two years.
>
> See, I told you they'd have a different story. :-) No, just kidding.
> Of course there are always faster and cheaper things just down the
> road, but when you're dealing with the transition from one bus structure
> to another, which threatens to obsolete a large portion of the third
> party products out there, and many software applications too, that's
> a little more serious than just another case of "better/faster/cheaper".
>
I was under the impression that there could be PCI -> NuBus adaptors to
avoid this problem. Is this true?
Sean
elliotte m harold math stnt
Mark Rogowsky <ro...@forsythe.stanford.edu> wrote:
>
>I think that the PowerPC sales will exceed 1 million easily in 1994 if the
>603-based Powerbooks reach the market in September... They could blow past
>1 million if the Q605/LC475 also goes PPC at that time.
Those are some BIG ifs. I doubt we'll see any volume in 603 machines
until early 1995. I also think that propaganda to the contrary, Apple won't
be able to meet the early demand for PPC 601 machines. The debacle is
going to exceed the unavailability of LCIII's last year and be of the
same order as IBMs ThinkPad troubles. IBM can't ship as many PPC
workstations as they want because there aren't enough chips, and Apple's
going to need two orders of magnitude more chips than IBM. I don't know
where they expect to get them from. I'm predicting that it's going
to be at least TWO months before all the orders taken in the first
week of "availability" are fulfilled.
--
Elliotte Rusty Harold Department of Mathematics
elh...@shock.njit.edu New Jersey Institute of Technology
emh...@hertz.njit.edu Newark, NJ 07102
Mark Rogowsky
(elliotte m harold math stnt) wrote:
>
> In article <rogo-030...@president-provost-04.stanford.edu>,
> Mark Rogowsky <ro...@forsythe.stanford.edu> wrote:
> >
> >I think that the PowerPC sales will exceed 1 million easily in 1994 if the
> >603-based Powerbooks reach the market in September... They could blow past
> >1 million if the Q605/LC475 also goes PPC at that time.
>
> Those are some BIG ifs. I doubt we'll see any volume in 603 machines
> until early 1995. I also think that propaganda to the contrary, Apple won't
> be able to meet the early demand for PPC 601 machines. The debacle is
> going to exceed the unavailability of LCIII's last year and be of the
> same order as IBMs ThinkPad troubles. IBM can't ship as many PPC
> workstations as they want because there aren't enough chips, and Apple's
> going to need two orders of magnitude more chips than IBM. I don't know
> where they expect to get them from. I'm predicting that it's going
> to be at least TWO months before all the orders taken in the first
> week of "availability" are fulfilled.
I believe this "shortage" of PowerPC chips is a myth. Do you have any
evidence (I'll accept "a friend at IBM" because I haven't even seen
anything that good to back the myth).
When I talked to IBM Microelectronics folk at Comdex, I got a real strong
impression that chip yields were excellent. The 601s been made for months
already. Where has it gone? I think there are hundreds of thousands of 601s
manufactured already.
I also think Apple will not have the debacle you state. It's a brave new
world. On March 14, 1994, Apple will introduce PowerMacintosh and you'll
see why 1994 won't be like 1984.
As for the 603, if Apple releases a Powerbook based on the 603 is October,
they'll sell 200,000-300,000 before the year is out. And if the LC goes
603, watch it...
Mark Rogowsky
ro...@forsythe.stanford.edu