#Conspiracies
20 messages in this thread
You are rubbing a sore spot now when you say:
> Sounds like the ol' 200 mpg carburator story
During the arab oil embargo one of my instructors at a company
training seminar told me he was laid off as an automotive engineer
and took a job teaching but still attended seminars to keep himself
current. He had just returned from one given by Carter Carburator
where they demonstrated something called the Dresserator, a radical
new carurator device. A company in California named Dresser had been
designing nozzles for Nasa rockets but lost thier contract when the
moon program shut down, so they decided to have a go at carbs. He
told me that Carter engineers told him this design could get 40-60
mpg from ordinary V8 american cars. When asked when it would be
released the reply supposedly was to the effect "we bought the
patent, but we have to sell the current design until the warehouses
are empty at a minimum, and probably much longer than that". He was
understandably furious but admitted that he knew of other cases where
more efficient designs were bought and surpressed because they were
not in the best interest of someone somewhere.
This all tells me that it is not so silly to believe that the 68060
could indeed be surpressed if Motorola stood to make more money
selling PowerPC chips to Apple and IBM and the strength of the 68060
CISC chip could embarass their main customers. For what? For the sake
of a relatively small group of buyers making special purpose
computers for the idiot fringe. It is possible the 68060 release is
not in their best interest. alternate: jonest@hfsi.com I never met a
man I didn't like…Will Rogers
"if Motorola stood to make more money selling PowerPC
chips to Apple and IBM and the strength of the 68060 CISC chip could embarass
their main customers"
Over the entire run of the 680×0 series, Apple has accounted for less than
10% of the output. They are a big customer, but not so big that Motorola would
really give a hoot. If anything, it's Motorola that has the leverage here
since Apple has little leeway in terms of switching vendors and maintaining
backwards compatibility.
In any event, it would appear that IBM actually owns the primary rights to
the PPC – and their consumption of the CPU's at the moment is close to zero.
Considering the capabilities of the PPC, why should Motorola devote any
further manufacturing resources to the 68060, particularly if they believe that
CISC CPU's are at the end of the line?
Freud said that sometimes a cigar is just a cigar – and sometimes
business decisions are just business decisions, not acts of
conspiracy.
Isn't it more like less than 1%?
>sometimes business decisions are just business decisions
Yes, but just because you are paranoid doesn't mean that no one is out
to get you.
"Isn't it more like less than 1%?"
The number I saw in the trade press about the time the PowerMac was
released was 10%. Don't recall the source at this point. Whether it's
1% or 10%, Apple is still fairly small potatoes to Motorola.
"Yes, but just because you are paranoid doesn't mean that no one is
out to get you"
And, alternatively stated, "even paranoids have enemies". <G>
Jerry
> Over the entire run of the 680×0 series, Apple has accounted
> for less than 10% of the output. They are a big customer, but
> not so big that Motorola would really give a hoot.
That fact distorts the current situation. Rephrase that as how much
of Moto's production of '030 and higher for DESKTOP systems has Apple
bought? Your comment includes early 68000 chips which are
_everywhere_, as well as embeded applications which rarely have use
of high end chips. Apple is the biggest high end customer Moto has.
>If anything, it's Motorola that has the leverage here since Apple
> has little leeway in terms of switching vendors and maintaining
> backwards compatibility.
Apple is the reason Moto was brought in as a way to make up for the
loss of the 88K which Apple had commited to buy. They, in effect,
"bought off" Moto by getting them the PPC contract.
> Considering the capabilities of the PPC, why should Motorola devote
> any further manufacturing resources to the 68060, particularly if
> they believe that CISC CPU's are at the end of the line?
Very true.
> Freud said that sometimes a cigar is just a cigar – and sometimes
> business decisions are just business decisions, not acts of conspiracy.
If more people belived that bit of common sense, we would not have theories
about maybe the driver is the one who killed Kennedy. (As we all know it _has_
to be someone other than the logical choice).
I view the '060 as a nice chip with little room for expansion. By removing all
of the roadblocks that the architecture brings, Moto opened up much more
powerful directions to go.
"Apple is the biggest high end customer Moto has."
I don't know. Could be. Regardless, Apple is definitely not the tail wagging
the Motorola dog.
"Apple is the reason Moto was brought in as a way to make up for the loss of
the 88K which Apple had commited to buy. They, in effect, "bought off" Moto by
getting them the PPC contract."
I've read maybe four different versions of how Motorola got into the
PPC deal, this being one of them. I've also read it the other way
around where Motorola forced Apple to take the PPC deal because
Motorola planned no further development of the 680×0 which would have
left Apple in deep water viz a viz the planned Intel (at that time)
586. And then there are the really weird stories about IBM . . . <G>
I don't pretend to know the realities of it, just that no one is
talking about PPC as an Intel killer any more. Of course that might
have changed over the past few weeks of Pentimania. <G>
"maybe the driver is the one who killed Kennedy"
Geez, I coulda told you: Colonel Mustard did it. <G>
Jerry
> I've read maybe four different versions of how Motorola got
> into the PPC deal, this being one of them. I've also read it
> the other way around where Motorola forced Apple to take the
> PPC deal because Motorola planned no further development of
> the 680×0 which would have left Apple in deep water viz a viz
> the planned Intel (at that time) 586. And then there are the
> really weird stories about IBM
I have spoken to several of the people directly involved in the
initial exchange between Apple and IBM.
Apple sent a delagation to IBM. Talks went well. Moto was invited in
by Apple as a way to ensure mutliple venders and to keep relations
good for the 88K which was not going well at that time. Apple and
Moto have been pretty tight over the years and I think they both
realize how faithful the other has been.
The ranking Apple RISC designed tells a funny story about how they
wore suits on their first meeting with IBM in an attempt to "dress
up" and impress IBM. Meanwhile, IBM was having a dress down Friday
that day. It led to a rather interesting role reversal.
"The ranking Apple RISC designed tells a funny story about how they wore suits
on their first meeting with IBM in an attempt to "dress up" and impress IBM.
Meanwhile, IBM was having a dress down Friday that day. It led to a rather
interesting role reversal."
Yeah. I read that in Stephen Levy's "Insanely Great". Different twist, of
course, but that's to be expected with Levy who sees reality through his very
own peripheral vision.
The whole PPC thing has interested me. Talk about partnerships born of
desperation.
Jerry
> The whole PPC thing has interested me. Talk about partnerships
> born of desperation.
To a certain real extent it was born of desperation. Had Apple not done
something of that magnitude, they would not long have survived. Windows was
catching up with the "ease of use" issue. Pentium would destroy any slim
chance at a performance lead.
I think they've done rather well. Their OS needs a lot of work, but all agree
they've handled the move to PPC nicely. Almost all software workds "okay" in
emulation, which helped bridge the gap until lots of native stuff was out.
I have been looking at the mail order catalogs over the last few weeks, and
almost every significant program (> 90%) are native. The only ones that are
not are either games, smaller educational things, or older unsupported
products.
I don't even own a PowerMac, yet half of what I own is a "fat" binary which
contains a PPC native version. I have not gone out of my way to do this, just
that most of the stuff I've bought has come that way, and most of the shareware
I've down loaded has also. (Also upgrades in versions also).
I only hope that their OS stuff goes as well. They need a lot of work, but
have shown that they are serious.
As for the IBM part of the deal. Talk about the tail wagging the dog. IBM
can't seem to get their act together. Sort of the twist on the old joke…
Q: "What do you get when you combine IBM and Apple?"
A: "Apple.:
I wish them well. At least one major company is trying to offer an alternative
to the Microsoft and Intel hemogeny.
There is one thing everyone must admit about Apple. They were willing to bet
the company on a radical idea, and they put their money where their mouth is.
They are the only company to go through two CPU changes and survive. 6502 ->
68K -> PPC.
Oh well. Here's for the future. Cheers!
—- Andrew
"They [Apple] are the only company to go through two CPU changes and survive.
6502 -> 68K -> PPC."
Interesting point, Andrew. But I think it should also be remembered that they
weren't provided a path with built-in backward compatibility ala the Intel 08x
series.
Apple didn't have a real choice in the matter: change or die.
In any event, Apple also faced a couple of "near-death" experiences
which shouldn't be forgotten. The Mac almost pulled the company into
oblivion. Only PageMaker, PostScript and the LaserWriter saved it.
You are essentially correct, though, the folks at Apple do have the guts to
play "bet the company". I think the jury is still out on the PowerMac. PPC
devices are not gaining new marketshare, but rather selling into the existing
Mac market (and there's nothing wrong with that: at least Macintosh owners do
seem to buy newer models of the product). Apple's overall market share
continues to decline – and that is not a good sign. New Mac adherents are not
being made in the numbers required.
I've been beta on Windows95 since they started shipping them and I must admit
it's starting to look better than the Mac in terms of its interface. Apple
won't have their new OS out until at least 1996. I finally got a Pentium 90 in
here – and it absolutely rips – at about two-thirds the price of an equivalent
PowerMac. So we're continuing the process of converting from being primarily
Mac shop to becoming a Windows environment. Apple's years of price-gouging and
high-end model abandonment left me with a bad taste.
Overall, none of this stuff really bothers me one way or the other.
Regardless of the OS or the CPU manufacturer, I still get the job done and
that's all that really counts.
Jerry
>Interesting point, Andrew. But I think it should also be remembered that
>they weren't provided a path with built-in backward compatibility ala the
>Intel 08x series.
I think their 68K emulation went well enough to ensure backward compatibility.
And don't forget, Intel will put their people through some kind of
change with P7. We have no idea how they plan on it. It will either
cost $ for extra circuits to make 80×86 work, or they will emulate.
But they will be forced to emulate a P6 where expectations will be
much higher than today.
> PPC devices are not gaining new marketshare, but rather selling into
> the existing Mac market (and there's nothing wrong with that: at least
> Macintosh owners do seem to buy newer models of the product). Apple's
> overall market share continues to decline – and that is not a good
> sign. New Mac adherents are not being made in the numbers required.
I think this is a flawed argument. PPC is doing very well in places where Macs
were previously not welcome.
And "marketshare" means nothing when the market is growing as fast as
it is. Sales mean everything. Each quarter their number of
computers sold has increased very nicely. That represents attracted
customers. There are enough buyers to support many companies. Apple
knows they will never dominate, but that there is a very large
"niche" for their style of computer. Just as many won't use an Apple
for one reason or another, there are those who feel the other way.
There are many companies in many industrys who would _LOVE_ to own
8-14% of a market, with increasing sales each quarter.
You comment that you have a Pentium 90 which is much cheaper than an
similar PowerMac. What brand? Are you talking Gateway 2000 or
Compaq/IBM? I would bet that the Macs are not radicilly higher in
price than the big name brand PCs. People love to compare the
cheaper clones against the Mac, yet never mention that you can get a
66Mhz PowerMac (7100) for _hundreds_ less than a high end 486 from
IBM. IBM charges so much for their name brand, yet I never hear
people saying they are going to lose. There are "layers" to every
market. Different brands offer features that different segments of
the market look for.
I will also assert that at the low end, I can get a PowerMac for at
least $300 less than a similar Pentium. I can buy a 60Mhz PowerMac
for $1,200 today (monitor sep), or complete system with bundled
software for under $1,500.
Put another way, they are selling as fast as Apple can make them.
They must be doing something right. Either way, no matter what we
think, the market will decide. The quarter ends this week. We
should see the first numbers in two or three weeks.
"PPC is doing very well in places where Macs were previously not
welcome."
I'd be interested in knowing what sources you're citing here.
Everything I've read to date says no.
"I would bet that the Macs are not radicilly higher in price than the
big name brand PCs."
Wanna bet? <G> The Pentium I purchased has 32Meg RAM, 512K cache,
2Meg VRAM, 1Gig HD, CD-ROM, DAT tape, Ethernet card for less than
what I paid for a PowerMac 8100 with a 500Meg HD and 8Meg RAM – and
which would have required an additional $1,400 video card to drive a
large monitor. I've been comparison-pricing Mac's vs. Intel machines
for a long time now and Mac's are always more expensive for less
capability.
But, yes, I do agree that buying some brand names could rachet the
price up.
"I can buy a 60Mhz PowerMac for $1,200 today (monitor sep), or
complete system with bundled software for under $1,500"
And you can buy a Pentium 90 system with a larger drive (but no bundled
software) for $1,625.
The Wall St. Journal reported today that U.S. computer shipments for the
fourth quarter were up 32% from a year earlier at 5.8-Million units.
I think your analysis of Apple's happy state of marketshare is
somewhat off: Apple's marketing cost is significantly higher than
those for Intel based machines. Their R&D expenses are also higher.
Having 8 to 14% market share doesn't mean much if you can't turn a
profit. Finally, Apple's sales are not increasing each quarter:
they've been up and down like a yo-yo in recent years, although the
last two quarters have been up. The real question is when do they
reach the end of the string on the PowerMac's.
"Put another way, they are selling as fast as Apple can make them."
Not quite. Some models (such as the PowerBook 540c) are in
chronically short supply, while others (the PowerMac 6100) sit in
unsold stacks at their warehouses. Apple has one of the worst records
for consistently failing to match production with demand.
"Either way, no matter what we think, the market will decide."
For sure. I'm not predicting doom and gloom for Apple, just a steady slide
into a smaller and smaller market position. You can still make money as Number
10 or Number 50 in the market if your management is good. People just don't a
lot about you or buy your stock and stuff like that. Sort of like IBM. <G>
Jerry
> And you can buy a Pentium 90 system with a larger drive (but no bundled
software) for $1,625.
A Pentium 90 for under $2,000? that must be a sight. Only recently
has the 66Mhz gone under $2,000! (resonable national brand)
What vender please. Which ad? Even Gateway 2000 costs much more than
that, and they are a mid level brand.
It is very unfair to compare the cheapest put together import clone
with a name brand. Everyone agrees you get what you pay for. I've
even seen reports in the PC mags showing that a very good 486/66 from
IBM can outperform the poorest 60Mhz Pentium machines because they
are just junk.
> Their R&D expenses are also higher. Having 8 to 14% market share doesn't
> mean much if you can't turn a profit.
Last quarter Apple had the _highest_ margins in the industry, being slightly
ahead of Compaq. Other than Compaq, almost all clone venders live on a tiny
percent. Apple made 25% profit per dollar sold. About 1% higher than Compaq.
>Finally, Apple's sales are not increasing each quarter: they've been up and
down
> like a yo-yo in recent years, although the last two quarters have been up.
Looked okay to me! (I read the annual stock holders report just today)
> Not quite. Some models (such as the PowerBook 540c) are in chronically short
> supply, while others (the PowerMac 6100) sit in unsold stacks at their
warehouses.
Not true. The 6100 is the best selling model they have. It sells more than
any other single machine in there line. It outsells the other PPC machines
2:1.
> Apple has one of the worst records for consistently failing to match
production with demand.
Very true.
It scares me to think they might not survive exactly because of the
R&D comment you made. Packard Bell does "okay", but don't spend a
dime on improving the system. They buy commodidy parts and slap them
together. The only real R&D is IBM, Microsoft, Compaq, and Apple.
People have been predicting their death for a decade. Eventually they
might be right. I think it really bothers people to see a company
try a different approach and do a few billion dollars a year of
business while at it.
Jerry,
>Apple has one of the worst records for consistently failing to match
production with demand.
No, it was Commodore aka Medhi Ali that had the worst. ;-}
Regards,
John Chang
The way I understood it, Motorola/Apple/IBM all understood that if they didn't
do it, Intel would gain supremacy by default.
Apple didn't want to have to cave in to using Intel chips, IBM didn't
want to be dependent on Intel, (at the time they didn't realise the
fight AMD and Cyrix could put up), and Motorola wanted to stay in the
CPU ball game.
Certainly Intel wasn't interested in making a Power PC chip. So who did that
leave?….
Conally killed Kennedy, shot himself to throw off suspicion, and then ate the
gun to hide the evidence. He turned Republican later…
– Alex… Celebrating 10 years on Compuserve
"Motorola/Apple/IBM all understood that if they didn't do it, Intel would gain
supremacy by default."
Ah, Intel already had supremacy – and the PowerPC sure ain't threatening
that. Even during the hottest days of Apple's PPC hype, no one ever claimed a
hope of getting more than 30% of the CPU market with the PPC. All the estimates
have been considerably scaled down since then.
"Apple didn't want to have to cave in to using Intel chips, IBM didn't
want to be dependent on Intel, (at the time they didn't realise the
fight AMD and Cyrix could put up), and Motorola wanted to stay in the
CPU ball game."
I think it was more complex than that, although it's only speculation on my
part.
Apple couldn't cave in and use an Intel chip. They have to date been notably
unsuccessful in porting the Macintosh OS to an Intel base. And switching to
Intel sure would have stolen all their thunder. I can see their advertising
now: "Better Windows From Apple". <G>
There was also a massive level of desperation at the time at IBM and Apple
regarding their respective futures.
I kind of think it was more like going home with the last woman at the bar .
. . you take what you can get.
"Conally killed Kennedy, shot himself to throw off suspicion, and then ate the
gun to hide the evidence. He turned Republican later…"
Geez, now that I know, I'm going to have to kill you. <G>
Jerry
>That fact distorts the current situation. Rephrase that as how much of
>Moto's production of '030 and higher for DESKTOP systems has Apple
>bought? Your comment includes early 68000 chips which are _everywhere_,
>as well as embeded applications which rarely have use of high end chips.
>Apple is the biggest high end customer Moto has.
I'm not so sure about that. Gobs of laser printers have '020's and 030's,
not to mention higher end i/o controllers…
Anyhow, I think there's still a market for '060's, but your right, PPC is
the way to go for the future.
Warp 3 — Engage!
Bob Comer Cheyenne, Wyoming U.S.A.
> This all tells me that it is not so silly to believe that the
> 68060 could indeed be surpressed if Motorola stood to make more
> money selling PowerPC chips to Apple and IBM and the strength of
> the 68060 CISC chip could embarass their main customers.
Moto has made their position very clear. They have layed out what
makes their RISC chips cheaper/faster than their CISC chips.
The 68K family is really very nice, but brings a lot of baggage that
weighs around it's neck like an anchor. Intel finds themselves in
the same boat, but have many more billions to throw at brute force
solutions. Even Intel has almost reached the end of the road and is
looking for a radically different approach for the P7.
This reply is a long one, but I think it important not to allow the
fear of a conspiracy to cloud peoples opinions. Moto did not do what
they did to placate Apple. Had Apple adopted the 68060 they would
have gone under in three more years. There would just not be enough
imporvement in the 68K family to keep them going.
Consider several issues which work to keep 68060 CPU's slow:
A larger cache can radicly speed up a CPU, but you need the space.
What could be eating that space?
1) Addressing modes which programmers love, but are rarely, if ever,
really used. Compilers almost never emit these fancy modes, but
transisters were given over to that.
2) Complex interupt logic which has to know how to push many different
types of stack frames.
3) Vastly overcomplex decode logic which must deal with instructions
which vary between 2 and 10, or more, bytes.
The clocks are very hard to get up on the 060. What could be holding
the clock down?
1) The clock is the length of the longest possible cycle. Variable
length instruction decode is so complex that it takes more time.
2) Microcode provides complex instructions at the expence of doing
multiple ROM lookups per instruction. This takes time.
3) Instructions which are rarely used, add the expense of gates
signals must propagate through.
There are issues about how many instructions can be executed at the
same time (Superscalar). What holds the 68K family back in this?
1) There is a condition code bottle neck. That is each ALU
instruction sets the flags depending on the result. The CPU can't
execute a future instruction, which might alter the flags, until it
is sure the flags are no longer needed.
The PPC solves this by having 8 sets of flags which can be scheduled
by the compiler to avoid conflict.
2) The variable length instruction hobgoblin. The more combinations
of instruction lengths which exisit, the harder it is to get and
decode instructions in advance to execute them in parallel.
3) Not enough registers. The 68K only has 16 general purpose
registers and of that only 8 are really general purpose. This means
the CPU spends a lot of time moveing data back and forth to scratch
memory. The PowerPC has 31 general purpose registers. This allows
more of the problem to be kept in registers which are several times
faster than scratch ram. Remember a register is twice as fast as
cache, and 4-10 times faster than RAM. On a tight inner loop this
makes a HUGE difference.
There are things which the 68K does which slow a program in general
down:
1) Two operand instructions. All 68K instructions are of the form A
= A op B. This means that the operand A is also the destination.
This means a lot of scratch values must be moved around to preserve
both operands. The PowerPC does it of the form A = B op C. This
helps to avoid register bottle necks.
2) The floating point is underdone on the 68060. The PowerPC allows
the following A = B + C * D as a single instruction. Because
multiply and add use different parts of the CPU they may be done in
parallel. This is a very common instruction for graphics and FFT.
The PPC takes two cycles to do that, but one new instruction may be
pipelined each cycle. The 68K is much slower and can't easily
overlap.
There is a lot more, but the bottom line is the 68060 is the peak of
technology of that family. Just by removing it's own faults it could
be radicly speeded up, but then would not be a 68K.
Moto did the right thing, Apple did the right thing, and the Amiga, if
it survives, must move to some RISC platform or will become nothing
but a curiosity from the late 80's.
Oh well. Could be worse. Could be rainin…..
Well, those carter warehouses should have emptied long since…
Most stories about huge mileage carbs were in the 100+ region. I
remember one tale that claimed 140+ mpg for a full size Lincoln. A
little math at the time showed that there weren't enough BTU's in a
gallon of gas to get that milage at even 100% efficiency. (Which
ofcourse no heat engine is capable of.)
Especially during the dark days of the Japanese imports drubbing the
American products, the auto companies would have paid dearly for any
device that would get them 40 to 60 mpg. They would still pay dearly
for such a device, as the requirements for mileage for autos
increases in the government regs.
Carter would be back in the thick of things as a supplier to the
auto industry, were they able to produce such a device at any sort of
reasonable price.
And last but not least, Motorola is going to sell all of everything
it can to make a buck. The average Mac user is not as technically
aware, as an average Amiga user. They don't often open a case, let
alone make a CPU upgrade. I doubt Motorola is worried that the 68060
will kill Power PC sales. If it were that good, even the more reason
to release it.