IBM/Apple PR: 120 Mhz Power Mac demoed at Comdex & WWDC
R S Rodgers
Leigh Chao <cha...@sluaxa.slu.edu> wrote:
Oh, lookie here, lots of Mac "facts" blown away, if this is true:
>This week's MacWeek says Apple is shipping a 120MHz PPC601 based Mac, code
>named Tsunami next January! It will come in a tower case larger than that of
>8100. It will have 6 PCI slots,
Mac whine #1: "You don't need a lot of slots. Why would you need a lot of
slots? You only need slots because PeeCees suck!"
>512K L2 cache,
Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
NEED cache."
>Geoport, ethernet, fast SCSI,
>16-bit stereo sound, DMA and SIMM slots for upto 768MB of RAM. It will not
>have built-in video because graphics users will want an accelerated video card
>anyway.
Mac whine #3: "We don't NEED accellerated graphics because our graphics
are fast as is, and plus, we have them INTEGRATED and not like stupid
PeeCees that need a card added on."
>It will not have FireWire because not many peripherals will support
>FireWire by then,
Mac whine #4: "Plus, we'll have FIREWIRE _soon_ and you PeeCee users wont!!"
>but Apple supposedly will supply PCI-based FireWire cards later.
Mac whine #5: "And you CAN'T get it because FireWire is so fast that it wont
work as a plug-in card."
--
"If I went apeshit in here, you'd be in a lot of trouble, wouldn't you? I
could screw your head off and place it on the table to greet the guard."
-- Edmund Kemper, convicted serial killer, to
Robert K. Ressler ("Whoever Fights Monsters")
Dennis O'Connor
103t_e...@west.cscwc.pima.edu writes:
] In article <>, rsro...@wam.umd.edu (R S Rodgers) writes:
] > Mac whine #1:,
] > Mac whine #2:
] > Mac whine #3:
] > Mac whine #4:
] > Mac whine #5:
]
] I think that you mentioned PowerPC once in your entire message. Care to
] explain why you bothered to post it to comp.sys.powerpc? You didn't even
] bother to cross-posst to comp.sys.mac.advocacy...
Where as you didn't mention Intel in your entire message. Care to explain
why you bothered to post it to comp.sys.intel. You didn't even bother
to cross-post to rec.humor.
[ I shall gratuitously mention the Apple Macintosh, which is a venerable
old system but has been end-of-lifed, and the PowerPC, which makes a nice
low-end RS6000, which might be all IBM ever really wanted. ]
:-):-):-):-):-)
--
Dennis O'Connor doco...@sedona.intel.com
Intel i960(R) Microprocessor Division Solely responsible for what I do.
"When you discover a fluke, first determine if a whale is attached to it."
Jason Newquist
: Oh, lookie here, lots of Mac "facts" blown away, if this is true:
[obfuscated list of insipidly simplistic "facts" deleted]
Apple is addressing the needs of a sector of the computer market and you
FLAME it? GIVE ME A BREAK. What does it have to do NOT to get flamed?
--
/ / Jason R. Newquist, University of California, Davis /\ /\ "Silence is
/\ /\ jrnew...@ucdavis.edu Shields Library Reserves \/ \/ best left
\/ \/ Quotations Listserver Admin. / / unspoken."
R S Rodgers
Jason Newquist <ez03...@dale.ucdavis.edu> wrote:
>R S Rodgers (rsro...@wam.umd.edu) wrote:
>: Oh, lookie here, lots of Mac "facts" blown away, if this is true:
>[obfuscated list of insipidly simplistic "facts" deleted]
>
>Apple is addressing the needs of a sector of the computer market and you
^ belatedly
>FLAME it?
No, actually, I was flaming the typical Mac whiners who have been
using the BS excuses that I listed for literally _years_ to excuse
Apple's refusal to make some very obvious, very necessary changes
to their hardware line.
103t_e...@west.cscwc.pima.edu
>
> Mac whine #1:,
>
> Mac whine #2:
>
> Mac whine #3:
>
> Mac whine #4:
>
> Mac whine #5:
I think that you mentioned PowerPC once in your entire message. Care to explain
why you bothered to post it to comp.sys.powerpc? You didn't even bother to
cross-posst to comp.sys.mac.advocacy...
Lawson
Marc Nimchuk
: Jason Newquist <ez03...@dale.ucdavis.edu> wrote:
: >R S Rodgers (rsro...@wam.umd.edu) wrote:
: >: Oh, lookie here, lots of Mac "facts" blown away, if this is true:
: >[obfuscated list of insipidly simplistic "facts" deleted]
: >
: >Apple is addressing the needs of a sector of the computer market and you
: ^
: ^ belatedly
: >FLAME it?
: No, actually, I was flaming the typical Mac whiners who have been
: using the BS excuses that I listed for literally _years_ to excuse
: Apple's refusal to make some very obvious, very necessary changes
: to their hardware line
I wont argue that there isnt a place in Apples lineup for multi slot, video
accelerated machines, however Id argue that the VAST majority of Apple's
consumer base doesnt need more than 3 slots (a common configuration) nor
accelerated graphics that would basically only benefit large screen systems.
For standard 640x480 displays (800 x 600 too) the performance of the Powermacs
internal video is MORE than adequate.
R S Rodgers
Terje Rydland <terje....@ifi.unit.no> wrote:
>> >512K L2 cache,
>>
>> Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
>> NEED cache."
>
>Anyway the PPC is RISC. 80x86 and 680x0 are CISC chips. There is a
>difference. :-)
Hey, cool. An all new misapplication of RISC vs. CISC.
>> Mac whine #3: "We don't NEED accellerated graphics because our graphics
>> are fast as is, and plus, we have them INTEGRATED and not like stupid
>> PeeCees that need a card added on."
>
>The built-in graphics on a Mac IS fast enough for most users. You can get
>faster graohics by using *expensive* third party video cards. This is what
>graphical agencies use when they need to move 20 MB+ graphical files
>in 32-bit colour around on a 21" screen with a resolution of 1600*1200.
I consider the built-in MAc graphics pretty miserable in comparison
to PC graphics these days, and the difference is that an ATI GPT is
$600 for a 4MB truecolor to 1280x1024/16bit in 1600x1024, whereas
a slower card from SuperMac or Radius starts at twice that. One
is (almost) affordable and relatively inexpensive if bundled with
the system, the other isn't.
Joe Bennett
>R S Rodgers (rsro...@wam.umd.edu) wrote:
>: In article <CqqDM...@ucdavis.edu>,
>: Jason Newquist <ez03...@dale.ucdavis.edu> wrote:
>: >R S Rodgers (rsro...@wam.umd.edu) wrote:
>: >: Oh, lookie here, lots of Mac "facts" blown away, if this is true:
>: >[obfuscated list of insipidly simplistic "facts" deleted]
>: >
>: >Apple is addressing the needs of a sector of the computer market and you
>: ^
>: ^ belatedly
>
>: >FLAME it?
>
>: No, actually, I was flaming the typical Mac whiners who have been
>: using the BS excuses that I listed for literally _years_ to excuse
>: Apple's refusal to make some very obvious, very necessary changes
>: to their hardware line
>
>I wont argue that there isnt a place in Apples lineup for multi slot, video
>accelerated machines, however Id argue that the VAST majority of Apple's
>consumer base doesnt need more than 3 slots (a common configuration) nor
Well, this then goes contrary to everything most Mac-whiners I hear
talk about. To set the stage:
I have a PCI/ISA 486 system. It has 5 ISA slots and 3 PCI slots. Currently,
two PCI slots are used, and one ISA slot. With everything I have, I cannot
for the life of me think of what else I will put in either an ISA slot
or a PCI slot. But I'm glad they're there.
Similarly, I have argued about why multiple-monitor support exists for EVERY
Mac that is sold. I have since been bombarded that "yeah, most people
don't use it, but I for one am glad it's there".
The arguments are similar. The perceived value is what is important.
>accelerated graphics that would basically only benefit large screen systems.
>For standard 640x480 displays (800 x 600 too) the performance of the Powermacs
>internal video is MORE than adequate.
I guess that depends. I can't stand the tiny (standard) Mac monitors, no
more than I can stand tiny PC monitors. I can't imagine burning THE ONLY
SLOT I HAVE (the 6100 has one, right?) for an accelerated graphics card.
Joe
ba...@rock.enet.dec.com
>> named Tsunami next January! It will come in a tower case larger than that of
>> 8100. It will have 6 PCI slots,
> Mac whine #1: "You don't need a lot of slots. Why would you need a lot of
slots? You only need slots because PeeCees suck!"
If you had read the article, you would have found out that the machine is aimed
at desktop publishers and others who need the fastest video around.(among
other things) It makes sense to have more slots when there is no on-board
video. The reason people have criticized PC's in the past is that almost
everything is an add-on! The common home user would probably rather not have
to add a lot to get the machine they want. The traditional mac is great for
that. Just plug in the keyboard, mouse, monitor, and power and turn it on!
High-end users probably want more.
>>512K L2 cache,
>Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
NEED cache."
Since when was this a common "Mac whine"??? PowerPC is a new RISC chip. RISC
chips *need* more cache to stay fast.
>Mac whine #3: "We don't NEED accellerated graphics because our graphics
are fast as is, and plus, we have them INTEGRATED and not like stupid
PeeCees that need a card added on."
A majority of users *don't* need accelerated graphics. Only people doing
high-end graphics work do.(or people who want the fastest possible) The
Macintosh graphics engine QuickDraw, is generally recognized to be better
than the one in Windows anyway. Argue with that all you like.
>Mac whine #4: "Plus, we'll have FIREWIRE _soon_ and you PeeCee users wont!!"
>>but Apple supposedly will supply PCI-based FireWire cards later.
>Mac whine #5: "And you CAN'T get it because FireWire is so fast that it wont
> work as a plug-in card."
Now, I think you are arguing just to flame away. I think most people are
satisfied with SCSI and current peripheral connection. I don't see people
complaining.
Why do you feel the need to get all defensive? Is it because this will
blow away any Pentium machine on the market when it comes out?( Don't count
on Intel to have anything faster in machines that are shipping by then)
-SB
eric hancock
processors?
I don't remember a thing about this from the press releases.
E-
Philip Machanick
Rodgers) wrote:
> No, actually, I was flaming the typical Mac whiners who have been
> using the BS excuses that I listed for literally _years_ to excuse
> Apple's refusal to make some very obvious, very necessary changes
> to their hardware line.
This cuts both ways. Have you noticed how every time MS issues a new
release of Windows, it's announced as "finally your PC can look just like a
Mac"? And suddenly a whole bunch of features that have been standard on
Macs since 1984 move from being ridiculed by PC users to "must have"?
A really essentially change Apple needs to introduce (and here we are off
hardware, so please follow up to mac.advocacy) is the ability to log on as
a user to a machine and have it come up with your personality (a la UNIX
home directory). This would make a vast difference to maintaining shared
machines and is much more important to me than changes in slot
architecture, or even "true" multitasking.
--
Philip Machanick phi...@cs.wits.ac.za
Department of Computer Science, University of the Witwatersrand
2050 Wits, South Africa
phone 27(11)716-3309 fax 27(11)339-7965
Joe Bennett
>>>512K L2 cache,
>>Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
> NEED cache."
>
>Since when was this a common "Mac whine"??? PowerPC is a new RISC chip. RISC
>chips *need* more cache to stay fast.
Then you haven't been paying attention. The current PMac whine is that
no extra cache is needed (the 6100 doesn't have one) and benchmarks put
it in the Pentium-60 range. Then there is all kinds of extrapolation,
which is inherently wrong, about how this affects the system price, and
other extrapolations as to just how much faster the PMac *should* be
if you put the cache in.
>
>>Mac whine #3: "We don't NEED accellerated graphics because our graphics
> are fast as is, and plus, we have them INTEGRATED and not like stupid
> PeeCees that need a card added on."
>
>A majority of users *don't* need accelerated graphics. Only people doing
>high-end graphics work do.(or people who want the fastest possible) The
BINGO! Most everybody wants fast graphics. Look at the market growth for
PC acclerators. Most people buying these are NOT hi-end graphics users.
Even many except the most die-hard Macfanatics realize that PC graphics
are surperior due to the accelerators.
Joe
Brian Quinlan
>BINGO! Most everybody wants fast graphics. Look at the market growth for
>PC acclerators. Most people buying these are NOT hi-end graphics users.
>Even many except the most die-hard Macfanatics realize that PC graphics
>are surperior due to the accelerators.
>Joe
Everyone wants the fastest graphics possible BUT are they really doing to
improve on the built-in video on their macs? I could probably get a better
video card for my mac for say $300. But thats $300 that I could spend
on another 8Mb of RAM or put it in my gig fund or buy some software I want
or replace my double-speed CD with a triple-speed or...
Fast graphics are good but there comes a point when it's best to leave it
and get other things that you want. I'd have to spend a lot of money
on other things before I could justify replacing my built-in video.
--
------------------------------------------------------------
Brian Quinlan
qui...@sfu.ca
"...but people also thought that the world was flat"
Issa El-Hazin
>In article <terje.rydland...@majestix.ifi.unit.no>,
>Terje Rydland <terje....@ifi.unit.no> wrote:
>>> >512K L2 cache,
>>>
>>> Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
>>> NEED cache."
>>
>>The 601 is made by IBM. :-)
>>Anyway the PPC is RISC. 80x86 and 680x0 are CISC chips. There is a
>>difference. :-)
> Hey, cool. An all new misapplication of RISC vs. CISC.
God. That Apple brain washing machine surely done a number on a lot of
people.
>>> Mac whine #3: "We don't NEED accellerated graphics because our graphics
>>> are fast as is, and plus, we have them INTEGRATED and not like stupid
>>> PeeCees that need a card added on."
>>
>>The built-in graphics on a Mac IS fast enough for most users. You can get
>>faster graohics by using *expensive* third party video cards. This is what
>>graphical agencies use when they need to move 20 MB+ graphical files
>>in 32-bit colour around on a 21" screen with a resolution of 1600*1200.
> I consider the built-in MAc graphics pretty miserable in comparison
> to PC graphics these days, and the difference is that an ATI GPT is
> $600 for a 4MB truecolor to 1280x1024/16bit in 1600x1024, whereas
> a slower card from SuperMac or Radius starts at twice that. One
> is (almost) affordable and relatively inexpensive if bundled with
> the system, the other isn't.
The latest issue of computer shopper showed that a P5-90 with the Matrox
card (a lesser card than the ATI Mach64) out performed an 8100/80 PM over
6 times in graphic performance.
Also, not to mention that the "PeeCees" almost always come with PCI EIDE
HDs and which will trance the best PMs in throughput many times over.
>--
>"If I went apeshit in here, you'd be in a lot of trouble, wouldn't you? I
> could screw your head off and place it on the table to greet the guard."
> -- Edmund Kemper, convicted serial killer, to
> Robert K. Ressler ("Whoever Fights Monsters")
--
=======================================================================
Issa El-Hazin (is...@unomaha.edu)
University of Nebraska at Omaha
Technical Specialist (C&DC)
Ian Hin Yun Chan
>rsro...@wam.umd.edu (R S Rodgers) writes:
>The latest issue of computer shopper showed that a P5-90 with the Matrox
>card (a lesser card than the ATI Mach64) out performed an 8100/80 PM over
>6 times in graphic performance.
Ummm... there was a thread started by me asking if this was true (PowerPC vs
Pentium), and there answer is: Computer Shopper did the benchmarks with
SoftWindows whenever possible, so that's mostly emulated performance you're
talking about (read the text carefully; it doesn't say that anywhere on the
graphs. Rather nasty of them). I'm told that in reality the 8100/80 video is
faster than any Pentium video (but not by huge factors).
>Also, not to mention that the "PeeCees" almost always come with PCI EIDE
>HDs and which will trance the best PMs in throughput many times over.
Clueless 'bout this one.
>>--
>>"If I went apeshit in here, you'd be in a lot of trouble, wouldn't you? I
>> could screw your head off and place it on the table to greet the guard."
>> -- Edmund Kemper, convicted serial killer, to
>> Robert K. Ressler ("Whoever Fights Monsters")
>--
>=======================================================================
>Issa El-Hazin (is...@unomaha.edu)
>University of Nebraska at Omaha
>Technical Specialist (C&DC)
- Ian
sea...@ac.dal.ca
> In article <2shf7c$g...@wupost.wustl.edu>,
> Leigh Chao <cha...@sluaxa.slu.edu> wrote:
>
> Oh, lookie here, lots of Mac "facts" blown away, if this is true:
Gawd you're cantankerous. This is a Mac aimed at the high-end publishing
market. Apple already sells machines for this market--950s loaded to the
brim with cards.
>
>>This week's MacWeek says Apple is shipping a 120MHz PPC601 based Mac, code
>>named Tsunami next January! It will come in a tower case larger than that of
>>8100. It will have 6 PCI slots,
>
90% of Mac users don't need slots. Those with 2 16" monitors (oh sorry, can't
do that on a PC...), accelerated video cards with DSP accelerators, Fast&
Wide SCSI II cards, etc., do, and they deserve a specialized machine.
> Mac whine #1: "You don't need a lot of slots. Why would you need a lot of
> slots? You only need slots because PeeCees suck!"
>
>>512K L2 cache,
>
Third-party secondary caches have always been available, and have always
been favourably reviewed in Mac mags.
> Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
> NEED cache."
>
>>Geoport, ethernet, fast SCSI,
>>16-bit stereo sound, DMA and SIMM slots for upto 768MB of RAM. It will not
>>have built-in video because graphics users will want an accelerated video card
>>anyway.
>
> Mac whine #3: "We don't NEED accellerated graphics because our graphics
> are fast as is, and plus, we have them INTEGRATED and not like stupid
> PeeCees that need a card added on."
Again, this is a specialized machine. Your typical user does not need dual
DSP Photoshop acceleration, hardware CMYK-RGB conversion, etc., but high-
end DSP/pre-press does. Recent Macs have video on a PDS card anyways (7100
and 8100). If you're going to sink $2500-$3500 on a specialized card it
doesn't make sense to just throw out one Apple gives you. And those are the
kind of cards people who get this machine will be adding.
>
>>It will not have FireWire because not many peripherals will support
>>FireWire by then,
>
> Mac whine #4: "Plus, we'll have FIREWIRE _soon_ and you PeeCee users wont!!"
Most Mac users have never heard of FireWire. I don't think Apple intends on
keeping FireWire a Mac-only thing anyways.
>>but Apple supposedly will supply PCI-based FireWire cards later.
>
> Mac whine #5: "And you CAN'T get it because FireWire is so fast that it wont
> work as a plug-in card."
>
Don't think I've heard that one before. In fact, I think that MacWeek has
specifically mentioned PCI cards with FireWire on it.
Try again.
Joseph Hall
>BINGO! Most everybody wants fast graphics. Look at the market growth for
>PC acclerators. Most people buying these are NOT hi-end graphics users.
>Even many except the most die-hard Macfanatics realize that PC graphics
>are surperior due to the accelerators.
Personally I'm used to seeing graphics crawl on PCs and zip along pretty
well on Macs. Furthermore it is somewhat easier to design good accelerators
for the Mac because QuickDraw isn't too hard to hook, and there is no need
for developers to maintain a huge, ever-changing library of card drivers.
If you feel the need for speed, try out a Radius LeMans GT.
--
Joseph Nathan Hall | Joseph's Law of Interface Design: Never give your users
Software Architect | a choice between the easy way and the right way.
Gorca Systems Inc. | jos...@joebloe.maple-shade.nj.us (home)
(on assignment) | (602) 732-2549 (work) Joseph_Ha...@email.mot.com
Piner
Joe Bennett <jben...@frx146.intel.com> wrote:
[crap deleted]
>>I wont argue that there isnt a place in Apples lineup for multi slot, video
>>accelerated machines, however Id argue that the VAST majority of Apple's
>>consumer base doesnt need more than 3 slots (a common configuration) nor
>
>Well, this then goes contrary to everything most Mac-whiners I hear
No wonder we get tired of many Intel employees. They like to refer to Mac
users as a bunch of whiners, even though PC users do more than their share
of bitching.
>talk about. To set the stage:
>
>I have a PCI/ISA 486 system. It has 5 ISA slots and 3 PCI slots. Currently,
>two PCI slots are used, and one ISA slot. With everything I have, I cannot
>for the life of me think of what else I will put in either an ISA slot
>or a PCI slot. But I'm glad they're there.
Since I'll assume you have a sound card in the ISA slot, how about Ethernet,
additional modem, data aquisition card, professional video I/O system
(some take more than one slot), multi-port serial card(s) for running a
larger BBS system, fiber optic card, MPEG and/or JPEG card, interface to
VME bus...etc etc... of course, your average user won't use most of
these, but not everyone is an average user...
>
>Similarly, I have argued about why multiple-monitor support exists for EVERY
Because it doesn't really cost anything to have. So, you have a little code
in the ROMs and system software to handle it, but I am sure it doesn't take
much.
>Mac that is sold. I have since been bombarded that "yeah, most people
>don't use it, but I for one am glad it's there".
>
>The arguments are similar. The perceived value is what is important.
>
>>accelerated graphics that would basically only benefit large screen systems.
>>For standard 640x480 displays (800 x 600 too) the performance of the Powermac
>>internal video is MORE than adequate.
>
>I guess that depends. I can't stand the tiny (standard) Mac monitors, no
This argue is so tired and old and doesn't even apply anymore. Exactly
what is a "standard" Mac monitor? Most Mac users get a 14" monitor, just
like a PC user. And many Mac users use VGA monitors anyways, so I don't
understand your argument. If you think the resolution is too low, just
get a multisync VGA and switch the resolution higher.
I have a Sony 1430 14" VGA monitor on my Mac.
>more than I can stand tiny PC monitors. I can't imagine burning THE ONLY
>SLOT I HAVE (the 6100 has one, right?) for an accelerated graphics card.
This is your only good argument. Yes, one slot is too little, especially
when Apple forces you to buy a NuBus adapter just to install a NuBus card!
That is why I hope Apple makes computers with 3 or more full-size PCI
slots in most models.
>
>Joe
-Brian
103t_e...@west.cscwc.pima.edu
> In article <issa.770596278@cwis>, Issa El-Hazin <is...@cwis.unomaha.edu> wrote:
>>rsro...@wam.umd.edu (R S Rodgers) writes:
>
>>The latest issue of computer shopper showed that a P5-90 with the Matrox
>>card (a lesser card than the ATI Mach64) out performed an 8100/80 PM over
>>6 times in graphic performance.
>
> Ummm... there was a thread started by me asking if this was true (PowerPC vs
> Pentium), and there answer is: Computer Shopper did the benchmarks with
> SoftWindows whenever possible, so that's mostly emulated performance you're
> talking about (read the text carefully; it doesn't say that anywhere on the
> graphs. Rather nasty of them). I'm told that in reality the 8100/80 video is
> faster than any Pentium video (but not by huge factors).
But... If you read the description, you will find that they mention the
CodeWarrior C compiler from MetroWerks, saying that they used it to do their
benchmarks...
Either way, it is interesting: PC Magazine identifies certain tasks as testing
the video sub-system and says that the PowerMacs come out ahead in those tasks.
Computer Shopper says that they tested "common Windows graphics routines" using
CodeWarrior (?) and reports that the PowerMac is 1/6 the speed of the 90mhz
Pentium...
Makes you wanna throw out just about all benchmarks anyways, eh?
Lawson
103t_e...@west.cscwc.pima.edu
Wouldn't at least the first incarnation of FireWire work on NuBus? What is the
actual throughput of NuBus and NuBus-90 on the Mac?
Lawson
Joe Bennett
>In article <2sks2c$c...@inews.intel.com>,
>Joe Bennett <jben...@frx146.intel.com> wrote:
>>I have a PCI/ISA 486 system. It has 5 ISA slots and 3 PCI slots. Currently,
>>two PCI slots are used, and one ISA slot. With everything I have, I cannot
>>for the life of me think of what else I will put in either an ISA slot
>>or a PCI slot. But I'm glad they're there.
>
>Since I'll assume you have a sound card in the ISA slot, how about Ethernet,
>additional modem, data aquisition card, professional video I/O system
>(some take more than one slot), multi-port serial card(s) for running a
>larger BBS system, fiber optic card, MPEG and/or JPEG card, interface to
>VME bus...etc etc... of course, your average user won't use most of
>these, but not everyone is an average user...
Well, if you care to come over to my house and show me where my ethernet
port is located, I'll go buy a card and stick it in a slot to use ethernet.
The rest of your "what abouts" are nearly useless. There is a market for
this, but it is NOT the general market, and I for one don't want to pay
for stuff I'm not going to use.
In the Mac world, you don't get that choice, so you make rationalizations
as to the goodness of those things being there.
>>Similarly, I have argued about why multiple-monitor support exists for EVERY
>
>Because it doesn't really cost anything to have. So, you have a little code
>in the ROMs and system software to handle it, but I am sure it doesn't take
>much.
I don't know the exact cost, but I *do* know it is not for free. But
there you go with the MacRationalizing, "I'm sure it doesn't take much".
Joe
ba...@rock.enet.dec.com
>>
>>A majority of users *don't* need accelerated graphics. Only people doing
>>high-end graphics work do.(or people who want the fastest possible) The
>BINGO! Most everybody wants fast graphics. Look at the market growth for
>PC acclerators. Most people buying these are NOT hi-end graphics users.
>Even many except the most die-hard Macfanatics realize that PC graphics
>are surperior due to the accelerators.
>Joe
PowerMacs already give fast graphics. Tests have shown when using the DRAM
based video, that the performance hit is only ~3% compared to VRAM based. The
PMac is better on the plug-in card included with the 7100 and 8100. The truth
is, Macintosh built-in video is about equivalent to local-bus video on PC's.
As far as accelerated video goes, you need to compare on a case by case basis.
PC cards do tend to be cheaper, though. This says nothing about real
performance. If the video performance you refer to is based on the Computer
Shopper article, check what they were testing. I think the tests where they
claimed a 6x difference was with SoftWindows.
-SB
Piner
R S Rodgers <rsro...@wam.umd.edu> wrote:
>In article <2shf7c$g...@wupost.wustl.edu>,
>Leigh Chao <cha...@sluaxa.slu.edu> wrote:
>
>Oh, lookie here, lots of Mac "facts" blown away, if this is true:
Yes, you must think you have Mac users figured out. Oh, yeah, and
all Mac users are the same.
>
>>This week's MacWeek says Apple is shipping a 120MHz PPC601 based Mac, code
>>named Tsunami next January! It will come in a tower case larger than that of
>>8100. It will have 6 PCI slots,
>
>Mac whine #1: "You don't need a lot of slots. Why would you need a lot of
> slots? You only need slots because PeeCees suck!"
In a high-end Mac like the one described, you want a lot of slots so you
can put in cards to do all sorts of things. In a PC, you need several
slots to get things like video, fast SCSI, Ethernet, sound cards... On a Mac,
it is nice to have things that you need already there, and be able to use
the slots for doing "extra" stuff, not stuff that you need just to run your
computer.
>>512K L2 cache,
>
>Mac whine #2: "Unlike badly designed Intel chips, Motorola chips don't
> NEED cache."
No, that is NOT the point. The point is that PowerPC machines seem to perform
the same without cache as Intel machines with cache. Of course, PowerPC
chips have more internal cache than Pentiums, so this would help explain
this.
>
>>Geoport, ethernet, fast SCSI,
^^^^^^^
Personally, I hate the Geoport. Software-based Apple modems are scary.
And not much support for them, not to mention many unreliable pods.
>>16-bit stereo sound, DMA and SIMM slots for upto 768MB of RAM. It will not
>>have built-in video because graphics users will want an accelerated video
>>anyway.
>
>Mac whine #3: "We don't NEED accellerated graphics because our graphics
> are fast as is, and plus, we have them INTEGRATED and not like stupid
> PeeCees that need a card added on."
It IS convienent to have the video already there on the motherboard. As
far as speed, I want as much as I can get. The 040 and PowerPC Macs are local
bus as it is, so they are fast.
As far as not being built-in, this is a top-end machine and Apple found that
most high-end users were buying accelerated cards anyways, so they decided to
save money on it I guess. Sure, it won't be as nice to have it built-in,
but if it will lower system cost, and the cost third-party video boards, I
am all for it. Cheap video cards is a thing about PC's I like.
>
>>It will not have FireWire because not many peripherals will support
>>FireWire by then,
>
>Mac whine #4: "Plus, we'll have FIREWIRE _soon_ and you PeeCee users wont!!"
Never heard this before. As far as I know, IBM was going to have FireWire
around the same time as Apple...
>
>>but Apple supposedly will supply PCI-based FireWire cards later.
>
>Mac whine #5: "And you CAN'T get it because FireWire is so fast that it wont
> work as a plug-in card."
This is funny, since IBM has said several times that it will have FireWire
cards. I even read Apple said it would have FireWire cards a few months
ago.
I don't recall the speeds of FireWire, but think of this in terms of Ethernet.
Ethernet has a real-world speed significantly lower than its max speed.
I would assume this would be the same for FireWire, so it should be no
problem.
And PCI WILL increase in speed later, so this will handle the future increased
FireWire speeds. I don't see much of a problem here.
Maybe you need to talk to some different Mac users. Also, IBM users have
spewed their share of BS/false claims as well. Both sides need to learn
as much as possible about both kinds of computers.
-Brian
sea...@ac.dal.ca
Intel-whining thread that was supposed to terminate boring Mac/PC debate?
But it seems that you are the chief perpetrator in the continuance of this
thread, flogging non-issues. You have some valid points but this isn't one
of them.
Multiple monitors have been supported since 1987 (at least) with Color
QuickDraw. It is entirely a software issue! Do you propose that Apple produce
two versions of Color QuickDraw, one that has such support and one that
doesn't?? Now that is a ridiculous proposition. The notion that it costs
Apple to include this makes as much sense as saying that it costs Apple to
have colour support built in because some users only have monochrome screens.
Having 2+ monitor support integrated into the imaging model (i.e., it doesn't
care how many/what size screens are attached) rather than kludged on (as
happens in the PC world) makes for reliable behaviour with any application
in a multi-monitor setting, with multi-depth screens arranged how the user
chooses.
Multiple monitor support is not some far-out thing that only a few users
might possibly use. Take PowerBooks, for instance. You complain about having
to look at "tiny" Mac monitors (neverminding that all desktop Macs can
drive 21" monitors out of the box), but what about PC notebooks? Some do
support that 800x600 mode on an external (love that low refresh rate)
but your internal screen becomes useless. PowerBooks allow you to utilize
both the portable screen and an external monitor up to 832x624. Why should
your $2000 active matrix colour display become useless just because you're
working at your desktop? Furthermore, the latest MacUser, in an article
advising people what Mac to pick based on various needs, mentions a number
of times the usefulness and practicality of multi-monitor set-ups. It's
economical for a programmer to have a 2-page mono/grey display and a 14"
colour display. I myself am seriously considering, as my next peripheral
purchase, a second 14" monitor for my set-up. That'll let me play F/A-18
in panorama mode too :)
I think that the fact that Color QuickDraw abstracts screen space to an
arbitrary size & shape and hence allows multi-monitor support isn't much
of a con on the Mac side of the debate.
Sean
Michael Peirce
In article <2sq729$e...@inews.intel.com> (comp.sys.powerpc,comp.sys.mac.hardware,comp.sys.intel), jben...@frx146.intel.com (Joe Bennett) writes:
> Well, if you care to come over to my house and show me where my ethernet
> port is located, I'll go buy a card and stick it in a slot to use ethernet.
> The rest of your "what abouts" are nearly useless. There is a market for
> this, but it is NOT the general market, and I for one don't want to pay
> for stuff I'm not going to use.
Ethernet is an excellent way to send data quickly to a printer.
That's what *I* do in *my* house.
Macs have always had a network port built in (even the original
128K Mac!). It really wouldn't be a Mac without AppleTalk. One
of the biggest complaints has always been that the localTalk
port was simply too slow for many people. As ethernet prices
have come down, Apple started building it in.
Most Macs are sold into businesses, all the business oriented
Mac have ethernet built in. Some of the low end consumer
models don't - yet. But even the new portables (the 500
series) has ethernet as a standard feature.
I say, it's about time!
__ Michael Peirce __ pei...@outpost.sf-bay.org
__ Peirce Software, Inc. __ 719 Hibiscus Place, Suite 301
__ __ San Jose, California USA 95117-1844
__ Makers of: Smoothie & __ voice: +1.408.244.6554 fax: +1.408.244.6882
__ Peirce Print Tools __ AppleLink: peirce & AOL: AFC Peirce
Joe Bennett
>In article <2sq729$e...@inews.intel.com>, jben...@frx146.intel.com (Joe Bennett) writes:
>> In article <1994Jun04.0...@zeus.aix.calpoly.edu> bp...@harp.aix.calpoly.edu (Piner) writes:
>>>In article <2sks2c$c...@inews.intel.com>,
>>>Joe Bennett <jben...@frx146.intel.com> wrote:
>>
>>>>Similarly, I have argued about why multiple-monitor support exists for EVERY
>>>
>>>Because it doesn't really cost anything to have. So, you have a little code
>>>in the ROMs and system software to handle it, but I am sure it doesn't take
>>>much.
>>
>> I don't know the exact cost, but I *do* know it is not for free. But
>> there you go with the MacRationalizing, "I'm sure it doesn't take much".
>>
>No, Joe, you're Mac-Whining. Wasn't it you who starting this
>Intel-whining thread that was supposed to terminate boring Mac/PC debate?
No.
>Multiple monitor support is not some far-out thing that only a few users
>might possibly use. Take PowerBooks, for instance. You complain about having
1) Very few people have notebooks and desktop machines.
2) Of those that do, even fewer get their panties in a bunch if they
can't use both screens at once.
3) I have never argued that it is USELESS. I have argued that Mac people
stop jumping up and down as if this is some inherent Mac advantage,
given to us by the Great god that is Apple, when only the very few
actually NEED or WANT it. It's the same thing as getting all starry
eyed as Mac people do over the built in Ethernet, when I have yet
to find the Ethernet connection in my house that warrants the use of
it.
4) These may be things that some Mac users really like. I have never
argued that they shouldn't like it. However, if Apple truly wants
to lure customers from the PC to the Mac, switching the processor
like they did isn't the right strategy. If they would only look at
the rest of the computer industry, they might realize, "Oh, so THIS
is why we are only at 11% of the market!"
>I think that the fact that Color QuickDraw abstracts screen space to an
>arbitrary size & shape and hence allows multi-monitor support isn't much
>of a con on the Mac side of the debate.
Until somebody in the Mac arena can argue that it comes for free, I have
to continue to discount hand waving arguments like you make. Since you
do NOT know how much some of the built-in-only-used-by-the-minority
functions actually cost, how do you know the 6100 (sans monitor and
keyboard) wouldn't cost, say, $1300, instead of $1700, and thus giving
Apple lovers something worth actually arguing about in a cost/performance
debate?
You'll probably never get the chance to really know, because Apple will
dictate to you what you need, and not let you choose for yourself.
>Sean
Joe
The Man with the Glass Teapot
>>Multiple monitor support is not some far-out thing that only a few users
>>might possibly use. Take PowerBooks, for instance. You complain about having
>
>1) Very few people have notebooks and desktop machines.
>2) Of those that do, even fewer get their panties in a bunch if they
> can't use both screens at once.
So what? Having the notebook as your only computer is all the more reason
you would want to have dual screen capability. I know students here who
bought a PowerBook and a monitor for that reason. I don't think you
really understand the point....
Followup directed to comp.sys.mac.advocacy.
Chris Umbricht, M.D.
jben...@frx146.intel.com writes:
>1) Very few people have notebooks and desktop machines.
really? So unless its billions and billions sold, it's very few?
>2) Of those that do, even fewer get their panties in a bunch if they
> can't use both screens at once.
Nice truism.
>3) I have never argued that it is USELESS. I have argued that Mac people
> stop jumping up and down as if this is some inherent Mac advantage,
> given to us by the Great god that is Apple, when only the very few
> actually NEED or WANT it. It's the same thing as getting all starry
> eyed as Mac people do over the built in Ethernet, when I have yet
> to find the Ethernet connection in my house that warrants the use of
> it.
Well, it is an "inherent Mac advantage". I sympathize with you that
jumping up and down is irritating. And ethernet (on the Mac) is really,
really nice.... and the standard inclusion drives down the price
fiercely.
>4) These may be things that some Mac users really like. I have never
> argued that they shouldn't like it. However, if Apple truly wants
> to lure customers from the PC to the Mac, switching the processor
> like they did isn't the right strategy. If they would only look at
> the rest of the computer industry, they might realize, "Oh, so THIS
> is why we are only at 11% of the market!"
Sorry, I must be dense, but I don't follow your argument (if there is
one). Sounds more like FUD to me...
>>I think that the fact that Color QuickDraw abstracts screen space to an
>>arbitrary size & shape and hence allows multi-monitor support isn't much
>>of a con on the Mac side of the debate.
>
>Until somebody in the Mac arena can argue that it comes for free, I have
>to continue to discount hand waving arguments like you make. Since you
>do NOT know how much some of the built-in-only-used-by-the-minority
>functions actually cost, how do you know the 6100 (sans monitor and
>keyboard) wouldn't cost, say, $1300, instead of $1700, and thus giving
>Apple lovers something worth actually arguing about in a cost/performance
>debate?
Why would I care how much a built-in function costs? What matters is the
price/performance and completeness of the final product. If you really
don't think Apple has something worth arguing about in price/performance
debates, you must be on a desert island, and oblivious to Intel's current
scrambling to try to fight off major inroads in their marker share. But
again, I doubt this is really what you think, and is more likely part of
Intel's FUD campaign...
Rakesh Malik
but a mac-bigot. PMac video is rather pitiful compared to that of a Pentium
w/ PCI-based graphics accelerators such as the S3 928 or the Matrox MGA-II.
This is in comparison to a Pmac 8100, with the Pentium at 60 Mhz. Other
factors actually favor the Pmac, (ram, Pmac w/ 24, Pentium w/ 8, MS windows vs
Sys 7, both of which lobotimize performance big time). Also, when running
X windows atop Linux, the Pentium graphics left the Pmac's in the dust. This
is largely a result of the faster bus (PCI) and the coprocessor handling much
of the general windows-related stuff like drawing windows and bit blits.
-Rakesh
vl...@cs.jhu.edu
Ian Hin Yun Chan
Rakesh Malik <vl...@panini.cs.jhu.edu> wrote:
>Whoever said that PMac video is faster than Pentium video was OBVIOUSLY naught
>but a mac-bigot. PMac video is rather pitiful compared to that of a Pentium
>w/ PCI-based graphics accelerators such as the S3 928 or the Matrox MGA-II.
Could you give references? The only test where I saw the Pentium blow the
PMac was in Computer Shopper, and that was because the PMac was emulating
MS Windows.
>This is in comparison to a Pmac 8100, with the Pentium at 60 Mhz. Other
>factors actually favor the Pmac, (ram, Pmac w/ 24, Pentium w/ 8, MS windows vs
>Sys 7, both of which lobotimize performance big time). Also, when running
>X windows atop Linux, the Pentium graphics left the Pmac's in the dust. This
>is largely a result of the faster bus (PCI) and the coprocessor handling much
>of the general windows-related stuff like drawing windows and bit blits.
You know, a reference would really make this more believable...
Is Linux already available on the Mac? A PMac? I thought I read that Linux
was just being ported over...
>-Rakesh
>vl...@cs.jhu.edu
- Ian
Phil Harbison USG
Please change the subject line when you alter the subject matter of
a thread. When I select an article with a title like ``Re: IBM/Apple
PR: 120 Mhz Power Mac demoed...'', I have the silly expectation that
I'm going to read about a 120MHz PowerPC system. I do not expect to
read about how many monitors a Mac supports, or how bad Windoze sucks,
or how the Mac OS needs preemptive scheduling. Please be considerate
and edit the subject lines.
--
*name: Phil Harbison
*mail: harb...@zk3.dec.com
Amancio Hasty Jr
Hi,
I used to hack on X11R5 servers for 386bsd which now is FreeBSD so my eyes
are used to high speed graphics.
At any rate, when the PPC first came out, I went to Fry's in Palo Alto to
check out a few of the PPC systems. Was mostly interested on graphics
and system response. I walked away not very impressed with the overall
graphics performance of the system.
On the other hand, I saw a demo of a music video (I think it was using
motion j-peg) on a PPC with hardware assist, and I was green with
jealousy!
Amancio
--
FREE unix, gcc, tcp/ip, X, open-look, netaudio, tcl/tk, MIME, midi,sound
at freebsd.cdrom.com:/pub/FreeBSD
Amancio Hasty, Consultant |
Home: (415) 495-3046 |
e-mail ha...@netcom.com | ftp-site depository of all my work:
| sunvis.rtpnc.epa.gov:/pub/386bsd/X
Marc Nimchuk
: Whoever said that PMac video is faster than Pentium video was OBVIOUSLY naught
: -Rakesh
: vl...@cs.jhu.edu
Sorry, but thats in DIRECT conflict with what PC Magazine reported. They
compared a 100mhz Pentium machine with a Motrox MGA II PCI card. Overall,
they found the Pmac video systems (of the 8100-80) faster and implemented
better. Its all there in writing, see for yourself...
Wade Masshardt
jben...@frx146.intel.com (Joe Bennett) writes:
> 4) These may be things that some Mac users really like. I have never
> argued that they shouldn't like it. However, if Apple truly wants
> to lure customers from the PC to the Mac, switching the processor
> like they did isn't the right strategy. If they would only look at
> the rest of the computer industry, they might realize, "Oh, so THIS
> is why we are only at 11% of the market!"
Look, Joe, the 680x0 is a dead end, right? The 68060 will probably be
the last one, at least from what I read on this newsgroup & others. It
also makes sense, considering 1) they probably won't sell a whole lot
of them, and 2) they have the PowerPC to develop. So, how the hell is
Apple supposed to keep up if the DON'T switch?
+--------------------------------------------------------------------+
| "Alright you primitive screwheads, this...(holds shotgun up)...is |
| my boom-stick!" |
| Ash, from the movie "Army of Darkness" |
+--------------------------------------------------------------------+
| wa...@macc.wisc.edu |
+--------------------------------------------------------------------+
Joe Bennett
>In article <2st2k0$t...@inews.intel.com> Joe Bennett,
>>4) These may be things that some Mac users really like. I have never
>> argued that they shouldn't like it. However, if Apple truly wants
>> to lure customers from the PC to the Mac, switching the processor
>> like they did isn't the right strategy. If they would only look at
>> the rest of the computer industry, they might realize, "Oh, so THIS
>> is why we are only at 11% of the market!"
>
>Sorry, I must be dense, but I don't follow your argument (if there is
>one). Sounds more like FUD to me...
The PC industry is based upon a customer driven market. Customers
pay for what they want, and don't pay for what they don't want. This
is true even for processors, which is why there was a 486 SX to
begin with.
>Why would I care how much a built-in function costs? What matters is the
Ummm, because, like, maybe if I don't want it I don't want to pay
for it? Naw... that's too simple.
Joe
Patrick Chase
|>
|> I doubt this is really what you think, and is more likely part of
|> Intel's FUD campaign...
Do you really think Intel or any other major company would actually bother
to conduct a "FUD campaign" on USENET? If so, then you have a greatly
exagerrated sense of self-importance...
-----------------------------------------------------------------------------
Patrick Chase I speak for myself, not H-P
H-P San Diego
Joe Bennett
>In article <2st2k0$t...@inews.intel.com>
>jben...@frx146.intel.com (Joe Bennett) writes:
>
>> 4) These may be things that some Mac users really like. I have never
>> argued that they shouldn't like it. However, if Apple truly wants
>> to lure customers from the PC to the Mac, switching the processor
>> like they did isn't the right strategy. If they would only look at
>> the rest of the computer industry, they might realize, "Oh, so THIS
>> is why we are only at 11% of the market!"
>
>Look, Joe, the 680x0 is a dead end, right? The 68060 will probably be
>the last one, at least from what I read on this newsgroup & others.
Apple decided LONG before the 68060 that they were going to switch
processors. Their pullout of the 680x0 is directly related to the
680x0 dying out, not vice-versa. And Apple needed to leave the
680x0 because, like, well, RISC is so much sexier, and Apple has
a long history of flat out ditching stuff.
>It
>also makes sense, considering 1) they probably won't sell a whole lot
>of them, and 2) they have the PowerPC to develop. So, how the hell is
>Apple supposed to keep up if the DON'T switch?
They didn't have to pull out, and even when they did, there was a
nice, stable, high performing x86 out there they could have used.
They just decided they had to be different.
Joe
Adam Nash
Joe Bennett <jben...@frx146.intel.com> wrote:
>>Look, Joe, the 680x0 is a dead end, right? The 68060 will probably be
>>the last one, at least from what I read on this newsgroup & others.
>
>Apple decided LONG before the 68060 that they were going to switch
>processors. Their pullout of the 680x0 is directly related to the
>680x0 dying out, not vice-versa. And Apple needed to leave the
>680x0 because, like, well, RISC is so much sexier, and Apple has
>a long history of flat out ditching stuff.
>
>nice, stable, high performing x86 out there they could have used.
>They just decided they had to be different.
>
Look Joe,
I know you probably mean well (Then again...) Anyway, Apple did experiment
with a version of System 7 that ran on Intel. They canned it for various
reasons. They had a lot of choices for switching. Several key factors are
involved:
1) If you're going to switch, make it to a clean design. Without this
advantage, I think you'd have to agree no one can do better with a
15 year old CISC design than Intel.
2) Be assured that whatever they picked would be accessible for current
users, not too confusing, and not lead to a mass migration to the
simpleton's, I mean, similar and unfortunately popular solution, Windoze.
Writing a fast Mac emulator for the Pentium might not have been as easy,
fast, or possible. I don't know, I didn't spen 2 years looking for a chip.
3) There is good reason to believe that RISC has inherent advantages. Even
Intel's P5 and P6 are utilizing more and more features. Rumor has it P7 is
RISC with hardware encoding of x86. So clearly Apple is not alone in
moving to RISC, and your "RISC is sexier" remark shows you are here to flame
and not discuss.
Basically, Apple is probably the last company you'd want to bail on for
abandoning customers. Has there ever been a company that changed hardware
so drastically and yet provided upgrade paths for its users. I bought a
Quadra 18 months ago. I can buy a $500 card w/PPC to double my speed or
a $1500 mother board that will boost it up to 3-4x.
For God's sake, they JUST stop making the ][e, and you can STILL get
System software for it that makes it look like a MAC System 6 finder.
(In fact, I'd say the Apple ][ System looks alot like Chicago, save the
memory protection)
Anyway, You're here to flame, not to make sense. I'm sure that there are a
lot of Intel employees, fans, etc that are kind of bummed that they have to
be on your side, given your attitude and your posts.
Intel makes good chips, that's certain. They have FANTASTIC engineers, several
of my good friends have gone to work for them. They will continue to
succeed. They STILL probably were not the best choice for Apple.
Please be wary when you post, you wouldn't want your name in too many kill
files. That can get ugly.
Thanx,
Adam
Please leave followups to email as I think this is only tangential to Intel
discussion.
Wade Masshardt
jben...@frx146.intel.com (Joe Bennett) writes:
[arguments about how 680x0 development stopped because Apple stopped
using it deleted]
> They didn't have to pull out, and even when they did, there was a
> nice, stable, high performing x86 out there they could have used.
> They just decided they had to be different.
>
> Joe
Apple started looking for a "RISC" processor years ago. They were
looking for a chip with better price/performance & room for development
than the 680x0 family had. Can you honestly say the x86 family has
better price/performance over the PowerPC?
As far as Apple using the x86 family in their computers, sure, they
could have, but if everyone was using x86 chips, where would the
competition & drive for inovation come from?
Jim Wong
>nice, stable, high performing x86 out there they could have used.
>They just decided they had to be different.
Did you ever stop to think, perhaps, that Apple really does believe
that PowerPC (and RISC architectures in general) have a brighter
future than the x86 line? Don't you think Apple considered the
various advantages that switching to the x86 would provide?
Multi-million dollar, publicly held companies don't make decsions
just to "be different".
Patrick Chase
|>
|> Multi-million dollar, publicly held companies don't make decsions
|> just to "be different".
|>
They most certainly do, especially when "being different" has protected them
from the rampant cloning prevalent in other portions of the computer industry.
Apple would have to be suicidal to have used Intel chips - they'd gave to
compete directly with the likes of Gateway, and Apple is nowhere near lean
enough to survive in that kind of market.
Chris Umbricht, M.D.
pat...@sdd.hp.com writes:
>|> I doubt this is really what you think, and is more likely part of
>|> Intel's FUD campaign...
>
>Do you really think Intel or any other major company would actually
bother
>to conduct a "FUD campaign" on USENET? If so, then you have a greatly
>exagerrated sense of self-importance...
Ad hominem attacks are always much appreciated. But what, pray tell, does
my assessment of Intel employees actions have to do with my sense of
self-importance? I guess participation in a public debate immediately
betrays hubris...
If your point is the truism that usenet participants are but a
insignificant fractions of users, it can be used as well to assess the
relevance of all the posts clamoring for a CLI or Preemptive
Multitasking...
Mark Decker (guest from Wm Jewell College)
>The PC industry is based upon a customer driven market. Customers
>pay for what they want, and don't pay for what they don't want. This
>is true even for processors, which is why there was a 486 SX to
>begin with.
I was under the impression that the 386SX, 486SX, 68LC chips came
about because of yield results. The 386SX chip came about when
someone thought, "hey, when we test these chips for yield and
the only thing that doesn't pass is the FPU part, let's sell
it with the FPU disabled." So, the company is happy, they don't
lose all there money on the chip... and the customers are happy
because they get a chip at reduced cost.
I am curious. Is this true? Or were there original plans to
produce the 386SX.
In other words... Does the 386SX and 486SX chips have their
defunked FPUs disabled or are they designed without a FPU.
I hope I made my question clear.
mkdk
Jim Wong
|> They most certainly do, especially when "being different" has protected them
|> from the rampant cloning prevalent in other portions of the computer industry.
You're right in a certain sense, but Apple isn't being different for
frivolous or trivial reasons, and that was my point. I apologize for
my unclear wording.
Karl Thomas
Where are you getting your info? A PC magazine article said that the
graphics on the p_Mac were faster but the article in Computer Shopper
that said the video was slower was comparing emulated apps.
>-Rakesh
>vl...@cs.jhu.edu
Chris Umbricht, M.D.
jben...@frx146.intel.com writes:
>The PC industry is based upon a customer driven market. Customers
>pay for what they want, and don't pay for what they don't want. This
>is true even for processors, which is why there was a 486 SX to
>begin with.
Excellent example! What a great idea to cripple a chip (making its
production costs higher if anything) to be able to offer it at a lower
cost, and maintaining a higher profit margin for the intact chip.
I'm sure nobody would have wanted the full chip at the lower price. NOT!
Stupid Apple, on the other hand, offers people a (more) complete
solution, spreading the cost thinner, and thereby creating a much larger
customer base with standard features that developers can count on being
present in their customers computers. This must be a really dumb way to
do it, especially now that the Mac's prices are on par with Intel-inside
machines.
>>Why would I care how much a built-in function costs? What matters is the
>
>Ummm, because, like, maybe if I don't want it I don't want to pay
>for it? Naw... that's too simple.
Right on. Simplistic is how I would call it.
Adam Nash
Patrick Chase <pat...@sdd.hp.com> wrote:
>Apple would have to be suicidal to have used Intel chips - they'd gave to
>compete directly with the likes of Gateway, and Apple is nowhere near lean
>enough to survive in that kind of market.
>
Even if Apple ran on Intel, Gateway could not compete at all since it
doesn't have Mac ROMS. If Apple licensed them, it would only be to certain
models. Of course, this does not apply to the other OS's, but it certainly
does invalidate your point.
Also, Apple is pretty damn lean for a company that has done and continues to
do so much R&D for the computer industry.
-Adam
Richard Vireday
That's right Patrick! Everyone knows that Intel's FUD campaigns
only happen on Compuserve and the Wall Street Journal.
USENET, schnooze-NET! C'mon, do you really think that the "infobahn"
is a reality?
:-) Richard Vireday I speak for myself, not H-P, or Intel
Tony Hamilton
> As far as Apple using the x86 family in their computers, sure, they
> could have, but if everyone was using x86 chips, where would the
> competition & drive for inovation come from?
This is a very good point. In the end, a large motivational factor for going
with PowerPC may have simply been that all of the key CEO's (from Apple,
Moto, IBM, etc...) were just sitting around saying "Look guys, this is our
last chance to try and introduce a new ISA - there will be risks, but we've
gotta go for it".
The way I look at it, going with pure RISC meant they could more quickly
port over their OS and everything than they could have with Pentium. The
downside is the risk in going with a new architecture which may or may not
be more cost effective in the long run than the Intel architecture.
The mistake, I believe, was in all the PR from over a year ago. Had Apple
not said anything (letting IBM do their AIX stuff quietly on the side, which
is all it has amounted to anyway), they could have "popped" it all on the
public all at once, causing far more speculations and FUD (good FUD for
them, that is) in the industry, and maintained their market share up to that
point. Instead, by the time they introduced the PowerMacs, the initial
anti-Intel FUD wore off, and already the analysts are in the over-cautious
mode with respect to the future of PPC.
In the end, I think they made a decent technical decision to go with PPC,
but I find their execution questionable, and I think it has lead to an even
greater risk for Apple, their shareholders, and their customers. I find it
interesting that so many speculate that Intel should be looking into its own
PPC offerings, while I personally think the "other camp" should be doing this
by having x86 fallback positions, since I still see far long-term risk with
the PPC than x86. But I'm obviously biased.
--
TONY HAMILTON ==================================================
Intel Corporation "Have compiler - will travel"
Tony_H_...@ccm.fm.intel.com -or- tham...@pcocd2.intel.com
========================(I speak only for myself ... not Intel)=
Tony Hamilton
> I was under the impression that the 386SX, 486SX, 68LC chips came
> about because of yield results. The 386SX chip came about when
> someone thought, "hey, when we test these chips for yield and
> the only thing that doesn't pass is the FPU part, let's sell
> it with the FPU disabled." So, the company is happy, they don't
> lose all there money on the chip... and the customers are happy
> because they get a chip at reduced cost.
>
> I am curious. Is this true? Or were there original plans to
> produce the 386SX.
>
> In other words... Does the 386SX and 486SX chips have their
> defunked FPUs disabled or are they designed without a FPU.
>
> I hope I made my question clear.
Yes, even though the content is a bit off. First, it is true that these
chips came about from _market_ forces, _not_ yield issues. The 486SX with
it's missing FPU (and only a small percentage of SX's ever shipped ever
had non-functional FPU's on-board; most don't have an FPU at all, including
all shipped today) was a marketing idea. Disabled FPU's was just a way of
meeting marketing demand with an existing die until a fully-SX die could
be manufactured.
The 386SX is a different issue, since the SX in that part indicates 16-bit
data bus vs. 32-bit, not FPU, since all 386's do not have FPU's. But, the
same is true here, in that the 386SX was certainly a part conceived by
marketing folks, not engineers. In the industry in general, this can be
said of about 99% of all products nowadays. Companies rarely do things
anymore just because they want to, or because it can be done.
William R Keel
>Excellent example! What a great idea to cripple a chip (making its
>production costs higher if anything) to be able to offer it at a lower
>cost, and maintaining a higher profit margin for the intact chip.
>I'm sure nobody would have wanted the full chip at the lower price. NOT!
Don't be so snobby. Apple did it with the processor in the ][gs.
Supposedly, the chip in the ][gs was crippled so it wouldn't be faster
than Apple's flagship pc, the Mac.
Reed
rk...@fnugget.intel.com
These are my personal opinions.
Joe Bennett
>In article <2svebd$c...@inews.intel.com>
>jben...@frx146.intel.com (Joe Bennett) writes:
>
>[arguments about how 680x0 development stopped because Apple stopped
>using it deleted]
>
>> They didn't have to pull out, and even when they did, there was a
>> nice, stable, high performing x86 out there they could have used.
>> They just decided they had to be different.
>>
>> Joe
>Apple started looking for a "RISC" processor years ago. They were
>looking for a chip with better price/performance & room for development
>than the 680x0 family had. Can you honestly say the x86 family has
>better price/performance over the PowerPC?
Well.... look at the systems, and try to tell me the PowerPC is
inherently better. Do _not_ quote chip lot prices, because there
is a thing called supply-demand that drives that.
I believe the x86 has at least an equal price/performance comparison
to the PPC. I would need you to try to convince me somehow otherwise,
because I just can't see what your facts are to back it up.
>As far as Apple using the x86 family in their computers, sure, they
>could have, but if everyone was using x86 chips, where would the
>competition & drive for inovation come from?
AMD, Cyrix, IBM, TI, NexGen, SGS Thompson, etc. Do you need more?
Joe
Charles Owen
>I was under the impression that the 386SX, 486SX, 68LC chips came
>about because of yield results. The 386SX chip came about when
>someone thought, "hey, when we test these chips for yield and
>the only thing that doesn't pass is the FPU part, let's sell
>it with the FPU disabled." So, the company is happy, they don't
>lose all there money on the chip... and the customers are happy
>because they get a chip at reduced cost.
>I am curious. Is this true? Or were there original plans to
>produce the 386SX.
Not at all. First, the 386SX is a completely different part. It has
a 16 bit external bus rather than the 32 bit bus, so it requires less
external hardware (or did before the days of high density PC boards and
relatively cheap memory chips). It was introduced to provide a low
end 386 machine. Neither the 386SX or DX have floating point processing.
The 486SX is the equivalent of a 486DX without the floating point. It's
not easy to "fix" a broken part to disable the floating point. The
CPU has to not only not recognize the port, but report to software that
it is a 486SX, not a DX. While I'm sure laser surgery could do this,
I doubt if it's as economical as many feel it could be. Intel has always
said that they studied the DOS software market place and fewer than
1% of all programs use enough floating point to see any increase in
speed. Hence, a non-floating point version was an obvious product.
Since the floating point takes up a lot of space on the die, leaving it
off allowed either a smaller die or a less complex process and, hence
a cheaper part. It's sound business sense.
So, why is there not a PentiumSX? I don't know. I think the Pentium is
aimed, still, at the high end of the market and is not as cost driven
as the 486, so there's not as much need for a cheaper version, but that's
just a guess.
>In other words... Does the 386SX and 486SX chips have their
>defunked FPUs disabled or are they designed without a FPU.
One other thought about this: If you have a part that has failed test,
does it make sense to disable the section that failed and send that
part out, assuming the rest of the part is okay? I would think that
that would lead to quality problems since the failed part may have more
than just a localized problem.
--
| Charles B. Owen Charles...@dartmouth.edu |
| Dartmouth College Home: 603-448-5677 |
| 6211 Sudikoff Laboratory, Rm 108 |
| Hanover, NH 03755 |
Steve Kanefsky
Joe Bennett <jben...@frx146.intel.com> wrote:
>The PC industry is based upon a customer driven market. Customers
>pay for what they want, and don't pay for what they don't want. This
>is true even for processors, which is why there was a 486 SX to
>begin with.
No it's not true. There are plenty of things you have no choice with on
a PC. See below.
>>Why would I care how much a built-in function costs? What matters is the
>
>Ummm, because, like, maybe if I don't want it I don't want to pay
>for it? Naw... that's too simple.
What if I don't want a parallel port or serial ports or a mouse port or
SIMM slots or an overdrive socket or ISA slots, etc. etc.? Should each of
these be optional equipment and only provided if specified, or is it
beneficial to have a certain minimum level of functionality? In today's
world, it's beneficial to have a certain minimum level of capability to
communicate with other computers. Ethernet is now cheap enough that it
can serve as this minimum level. There are unexpected times when it's
really nice to have, even if you never thought you'd need it. I had a
friend drop by and he plugged right into my system and was effortlessly
printing to my LaserJet, transferring files to and from my machine, and
even logging into my machine and ftp-ing some files via my SLIP connection
to the Internet while I was doing other stuff on my machine (he was able
to do the latter because I was running A/UX).
You don't need an ethernet jack in your wall to make good use of ethernet.
You only need the jack if you want to cleanly run ethernet cabling between
rooms. I happen to have three network devices all in the same room
(although it would be nice to have the freedom to put them in different
rooms). A few cable companies are starting to offer ethernet connections
to the Internet, however.
--
Steve Kanefsky
--
Steve Kanefsky
Wade Masshardt
jben...@frx146.intel.com (Joe Bennett) writes:
> In article <2svo85$5...@news.doit.wisc.edu> wa...@macc.wisc.edu (Wade Masshardt) writes:
> >In article <2svebd$c...@inews.intel.com>
> >jben...@frx146.intel.com (Joe Bennett) writes:
> >
> >[arguments about how 680x0 development stopped because Apple stopped
> >using it deleted]
> >
> >> They didn't have to pull out, and even when they did, there was a
> >> nice, stable, high performing x86 out there they could have used.
> >> They just decided they had to be different.
> >>
> >> Joe
>
> >Apple started looking for a "RISC" processor years ago. They were
> >looking for a chip with better price/performance & room for development
> >than the 680x0 family had. Can you honestly say the x86 family has
> >better price/performance over the PowerPC?
>
> Well.... look at the systems, and try to tell me the PowerPC is
> inherently better. Do _not_ quote chip lot prices, because there
> is a thing called supply-demand that drives that.
Apple would be buying chips from you, not systems. I don't know what
Apple pays per 601, but I 'd be willing to bet it is cheaper than what
you are charging for Pentiums. The gap between the production costs of
the 601 & the Pentium (do I have to say TM here somewhere???) is
probably not as large as price per 1000 would indicate, but the only
info I have to go on is the published prices for the two chips.
> I believe the x86 has at least an equal price/performance comparison
> to the PPC. I would need you to try to convince me somehow otherwise,
> because I just can't see what your facts are to back it up.
I was just looking at the price of 601 & Pentium, since I thought we
were discussing chips, not systems. I still think that the PowerMac
family competes extremely well with Intel based PCs on a
price/performance basis, especially when you consider the cost over the
lifetime of the machine.
> >As far as Apple using the x86 family in their computers, sure, they
> >could have, but if everyone was using x86 chips, where would the
> >competition & drive for inovation come from?
>
> AMD, Cyrix, IBM, TI, NexGen, SGS Thompson, etc. Do you need more?
Well, yes actually. I think different competing chip architectures
does more to stimulate innovation & development. I have no data to
back this up, but if everyone was using x86 type chips nobody would
have a reason to switch. Seems to me things would tend to stagnate.
Now, mind you I am not saying that the Pentium is a bad chip. In fact,
I think it is good even though I probably will never buy a machine with
one. Competition is good. Keeps the Apple/IBM/Motorola group on their
toes. Visa versa as well, don't you agree?
> Joe
Bruce Hoult
> >It
> >also makes sense, considering 1) they probably won't sell a whole lot
> >of them, and 2) they have the PowerPC to develop. So, how the hell is
> >Apple supposed to keep up if the DON'T switch?
>
> They didn't have to pull out, and even when they did, there was a
> nice, stable, high performing x86 out there they could have used.
> They just decided they had to be different.
Sure.
The x86 is dying every bit as much as the 68k, and would be moreso if
it wasn't for herculean efforts, at enourmous expense, by Intel.
I'm really, really impressed by the Intel engineers who've managed to
take the x86 as far as they have, but it's getting to be like shovelling
sh*t uphill -- thee comes a point when it's easier to just divert a
river and make a clean start.
-- Bruce
Michael Peirce
In article <2t0990$s...@nyx.cs.du.edu> (comp.sys.powerpc,comp.sys.mac.hardware,comp.sys.intel), kath...@nyx.cs.du.edu (Karl Thomas) writes:
> >This is in comparison to a Pmac 8100, with the Pentium at 60 Mhz. Other
> >factors actually favor the Pmac, (ram, Pmac w/ 24, Pentium w/ 8, MS windows vs
> >Sys 7, both of which lobotimize performance big time). Also, when running
> >X windows atop Linux, the Pentium graphics left the Pmac's in the dust. This
> >is largely a result of the faster bus (PCI)
Both built in video options on the Power Macs don't go over over a interface
bus (PCI on the Pentium or NuBus on the PowerMac).
One of the video options uses main memory as the video option and
memory has as fast or faster access as you'll get over a any I/O bus
be it PCI or NuBus. 64 bit wide access helps alot.
I'm not sure what the other video option uses, but it's clearly not
the NuBus (nor will it be the PCI bus in the next generation Power
Macs).
If you need some sort of special purpose video card or simply want
a third monitor (like one my friends has - he has THREE 19" monitors
on his desk hooked to his Power Mac: two using on board video, the
third using a NuBus card) you can use the NuBus for that and it will
be slower than using on board video or PCI.
__ Michael Peirce __ pei...@outpost.sf-bay.org
__ Peirce Software, Inc. __ 719 Hibiscus Place, Suite 301
__ __ San Jose, California USA 95117-1844
__ Makers of: Smoothie & __ voice: +1.408.244.6554 fax: +1.408.244.6882
__ Peirce Print Tools __ AppleLink: peirce & AOL: AFC Peirce
John W. Komp
: Sure.
I've been following this discussion for a while now and I'm RTFL. To
say that the x86 series is a good or great product is sheer lunicy.
I've used most Intel and Motorola products from 8080, 8051 through
80386's and 6800, 6811s through 601's. Even when Intel gets a feature
set right they still hose up the implementation (compare the great
features of the 8051 with Dallas Semi's correct and much faster
layout).
After having used all these products and then to see Intel employees
state that the 601 is only on par with the x86 family is too funny.
Lets face it. Intel has got some serious problems. AMD and the other
x86 clone makers are putting a lot of pressure on them for that
market. The Pentium currently only has wins (that I know of) in the
desktop market while the PPC601 has several wins in the controller
market besides multiple desktop vendors (and not just Apple and IBM).
Intel had a great ride with a mediocre product in the x86 family. I
wish them well but I've definately got some reservations about their
ability to hit another home run.
-John
SB Industries
****************************************************
This morning I shot six holes in my freezer
I think I've got cabin fever
Somebody sound the alarm
- J. Buffett
****************************************************
David Lerner
|> In article <2svdov$c...@inews.intel.com> Joe Bennett,
|> jben...@frx146.intel.com writes:
|>
|> >The PC industry is based upon a customer driven market. Customers
|> >pay for what they want, and don't pay for what they don't want. This
|> >is true even for processors, which is why there was a 486 SX to
|> >begin with.
|>
|> Excellent example! What a great idea to cripple a chip (making its
|> production costs higher if anything) to be able to offer it at a lower
|> cost, and maintaining a higher profit margin for the intact chip.
|> I'm sure nobody would have wanted the full chip at the lower price. NOT!
|>
.....
|>
|> Right on. Simplistic is how I would call it.
Your understanding of the microprocessor business appears to be minimal.
Joe Bennett is right- the PC market is customer driven like all others. The
Intel486SX(tm) is not "crippled," and it was created to provide a low-cost solu-
tion for customers. To understand this better, learn about the relationships
between defect density, die area, and yield. Since you probably don't have
the time or desire, I'll sum up- for a given defect density, yield increases
exponentially as die area decreases. (the models are actually a good deal
more complicated than this, but this is essentially true) Therefore, for
even a relatively small decrease in die are, i.e. remove the FPU, yield
AND total die per wafer increases, resulting in a dramatic increase in good
die per wafer.
So the Intel486(SX) is cheaper to make than the Intel486DX(tm). How much
so? I do not know.
-David Lerner
============================================================================
David Lerner --------------------------------------------
Intel Corporation | The views contained within this post are |
2200 Mission College Blvd | a result of my own ridiculous opinions. |
Santa Clara, CA 95052 | Some settling may have occured while in |
(408) 765 5326 | transit, so interpret with caution. |
dle...@t12sys.intel.com --------------------------------------------
____________________________________________________________________________
Ibrahim Mohammed
: In article <2svo4m$6...@larry.rice.edu>, jim...@ruf.rice.edu (Jim Wong) writes:
: |>
: |> Multi-million dollar, publicly held companies don't make decsions
: |> just to "be different".
: |>
: They most certainly do, especially when "being different" has protected them
: from the rampant cloning prevalent in other portions of the computer industry.
: Apple would have to be suicidal to have used Intel chips - they'd gave to
: compete directly with the likes of Gateway, and Apple is nowhere near lean
: enough to survive in that kind of market.
While I agree that differentiation is essential for any company that
spends money on R&D, I disagree that Apple would not be able to compete
with the Gateways of the world if Macs ran on x86 arch. The key to the
success of the Mac has always been the OS. Just because the CPU would be
the same doesn't mean that the Mac OS could run on x86 because of the Mac
ROMs. Without Mac ROMs no clone manufacturer would be able to compete
with Apple by delivering the same product for less money because it would
be impossible for them to deliver the same product.
Ibrahim Mohammed.
: -----------------------------------------------------------------------------
LEE BRIAN
> >It
> >also makes sense, considering 1) they probably won't sell a whole lot
> >of them, and 2) they have the PowerPC to develop. So, how the hell is
> >Apple supposed to keep up if the DON'T switch?
>
> They didn't have to pull out, and even when they did, there was a
> nice, stable, high performing x86 out there they could have used.
> They just decided they had to be different.
just to be different? geez... oh yeah, you work for intel :)
anyway, you look at the intel architecture... it has its roots
in the 8080, 8085. don't tell me that you are going to have a
clean and fast 32/64 bit architecture from those old 8 bit processors!
anyway, if you want something fast, you've got to drop all ties from
the weird past products and then design something from the very beginning
again. and, as most people can see, it has to be RISC. and of course,
most RISC chips are structurally the same (there is only so much you can
do in a RISC chip)!
apple wanted to do emulation, so you speed up bit-manipulation, etc.
-> there you go, you've got the PPC
I know that the intel architecture has the huge software base on its
side, but you have to admit, the x86 is too old. and engineers
designing any new computer WON'T pick it as the main processors.
seriously... do you really believe what you said?
"decided they had to be different?" ha!
actually, even intel microcontrollers are bad... but that's another story.
later,
bjl
(if you have any comments, don't email me... post them, or I'll bounce it
back!)
--
Brian Jonathan Lee (aka "hojo") | "Beef satay?!?! Not beef satay again!!!!"
b...@ecf.toronto.edu | "XMen! XMen! Rescue Kitty from the caves!"
b...@eecg.toronto.edu | "Evil thy name is NETREK!"
Michael Shandony
William R Keel <rk...@fnugget.intel.com> wrote:
>In article <2t03v...@jhunix.hcf.jhu.edu> Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> writes:
>>Excellent example! What a great idea to cripple a chip (making its
>>production costs higher if anything) to be able to offer it at a lower
>>cost, and maintaining a higher profit margin for the intact chip.
>>I'm sure nobody would have wanted the full chip at the lower price. NOT!
>
>Supposedly, the chip in the ][gs was crippled so it wouldn't be faster
>than Apple's flagship pc, the Mac.
>
Apple did not cripple the 65816 in the Apple IIGS. They just refused to
go with a CPU that ran faster than 2.8 MHz. Supposably, the company that
designed the 65816 (Western Design?) had versions of the chip that ran
at ~25 MHz and also a dual CPU version. If I remember correctly, the
fastest 68000 in a Mac at that time was 16 MHz. Further, the 65816 was
slightly more efficient per cycle. Also, there was a press conference
at that time where Apple representatives where explaining that the
Mac was their future because they could not make the IIGS run any
faster. After that comment, supposably the president of the company
that designed the CPU charged the stage yelling "That's a lie! You
know that's not true!" Therefore your theory on not letting the IIGS
go faster than the MAC is probably true.
But, there is a difference between not letting a faster CPU into your
machine and deliberately breaking the connection your CPU so you can
sell a slower version. Deliberately breaking your CPU may be a
wise marketing move, but it sure seems strange from an engineering
perspective.
Mike
=============================
Mike Shandony
Bell-Northern Research, Inc.
vanh...@bnr.ca
====================================================================
The opinions expressed are my own and not necessarily those of BNR.
====================================================================
David Harwood
There is no controversy about this subject among academic computer
scientists and EE's.
103t_e...@west.cscwc.pima.edu
> They didn't have to pull out, and even when they did, there was a
> nice, stable, high performing x86 out there they could have used.
> They just decided they had to be different.
>
> Joe
>
>
Um, Joe? Care to inform us of the relative merits of running software designed
for the 68K series of chips on an ix86?
In other words, how fast can you emulate a program designed for a 68020 on a
P5?
Lawson
103t_e...@west.cscwc.pima.edu
> In article <2sthmj...@jhunix.hcf.jhu.edu> Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> writes:
>>In article <2st2k0$t...@inews.intel.com> Joe Bennett,
>
>>>4) These may be things that some Mac users really like. I have never
>>> argued that they shouldn't like it. However, if Apple truly wants
>>> to lure customers from the PC to the Mac, switching the processor
>>> like they did isn't the right strategy. If they would only look at
>>> the rest of the computer industry, they might realize, "Oh, so THIS
>>> is why we are only at 11% of the market!"
>>
>>Sorry, I must be dense, but I don't follow your argument (if there is
>>one). Sounds more like FUD to me...
>
> The PC industry is based upon a customer driven market. Customers
> pay for what they want, and don't pay for what they don't want. This
> is true even for processors, which is why there was a 486 SX to
> begin with.
>
>
> Ummm, because, like, maybe if I don't want it I don't want to pay
> for it? Naw... that's too simple.
>
Joe!
As a prospective ix86 user, I *only* want to run 32-bit programs. Why am I
***FORCED*** to be backwards-compatible with applications that will run on
ancient chips?
I never, ever want to use them.
Seems to me that Intel is *forcing* users to pay for backwards-compatibility.
OTH, since the support for multiple monitors on the Mac is software-only (you
can have a Mac that only supports one monitor), then the R & D for this has
been paid for long ago, unlike the ix86 situation where the user is STILL
paying for the extra hardware and real estate on the latest batch of ix86
chips to keep them backwards-compatible with the 8086 & co...
Lawson
103t_e...@west.cscwc.pima.edu
> In article <2sthmj...@jhunix.hcf.jhu.edu>, Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> writes:
> |>
> |> I doubt this is really what you think, and is more likely part of
> |> Intel's FUD campaign...
>
> Do you really think Intel or any other major company would actually bother
> to conduct a "FUD campaign" on USENET? If so, then you have a greatly
> exagerrated sense of self-importance...
>
Maybe not the companies, but certainly the employees of various companies feel
compelled to disseminate it.
Lawson
Rakesh Malik
based on a '68040 SX', a '040 w/o math co-processor. VERY much like the
486sx, if you ask me (or anyone else).
Seems to me that both companies are just trying to attack both the 'high-end'
and the 'low-end' markets. Which makes a lot of sense.
-Rakesh
vl...@cs.jhu.edu
Rakesh Malik
I got my info by trying them out. I have never trusted PC mag, to be honest.
Also, if you run windows on either machine, neither machine will reveal its
true performance capabilities (emulation or no, ms windows will ream performance
quite badly, which is why I run Linux and X instead, it is faster if the system
has enough memory)
>>-Rakesh
>>vl...@cs.jhu.edu
>
David J. Ayers
>One other thought about this: If you have a part that has failed test,
>does it make sense to disable the section that failed and send that
>part out, assuming the rest of the part is okay? I would think that
>that would lead to quality problems since the failed part may have more
>than just a localized problem.
I think we need to consider the cost of testing in this thread. Testing modern
chips is a very expensive proposition. You'd have to do the testing of the
original up, then sort out the parts that failed the co-processor tests only.
Next, you'd need to run them through a separate battery of tests, etc. (or
similar scheme). The result is probably too expensive to be profitable.
David
---------------------
Any opinions expressed are purely my own, not Intel's or anyone else's.
---------------------
Patrick Chase
|> The x86 is dying every bit as much as the 68k, and would be moreso if
|> it wasn't for herculean efforts, at enourmous expense, by Intel.
Then how do you explain the efforts by NexGen, AMD, and Cyrix? If the x86
is truly dying, why aren't they all designing chips around the PowerPC or
SPARC ISAs? If published reports of the Nx586, K5, and M1 are at all correct,
then there is existence proof that advancing the x86 does not require
"herculean efforts" by Intel...
|> I'm really, really impressed by the Intel engineers who've managed to
|> take the x86 as far as they have, but it's getting to be like shovelling
|> sh*t uphill -- thee comes a point when it's easier to just divert a
|> river and make a clean start.
True up to a point, in that it is easier to work from a clean sheet. One
question: what do you think will happen to PowerPC in, say, 5-10 years when
new implementation techniques have been devised which it was not architected
to support? Will Apple/IBM/Moto attempt to "shovel sh*t uphill", or will
they make yet another clean break?
|>
|> -- Bruce
----------------------------------------------------------------------------
Karl Thomas
That's true the IIgs had the capability to be a better Mac than the Mac.
It had a color version of the Finder before Macs did, better sound
capability and with a rather inexpensive accelerator it was actually
faster than some of the B&W Macs.
Patrick Chase
|>
|> Both built in video options on the Power Macs don't go over over a interface
|> bus (PCI on the Pentium or NuBus on the PowerMac).
Actually, the "Processor Direct Slot" in the PowerMacs does not truly reside
on the 601's system bus: instead, it provides an '040 bus interface, which in
turn is bridged onto the system bus. Given the resulting throughput
limitations, it's certainly no better than PCI. This is the interface used
for PDS video and AV cards, BTW.
|> One of the video options uses main memory as the video option and
|> memory has as fast or faster access as you'll get over a any I/O bus
|> be it PCI or NuBus. 64 bit wide access helps alot.
The DRAM based option is also the worst of the lot. To see why, consider the
bandwidth requirements for video. Refresh bandwidth is a fixed requirement,
equal to the pixel rate (in pixels/sec) times the pixel depth (in bytes).
For 8-bit video on a 16" monitor (I'm calling it 1024x768, although I know
it's actually less than that) at 60 Hz, the required refresh bandwidth is
47.2 MB/sec. Because DRAM is single-ported, this bandwidth must be supplied
by the system bus. The second requirement is for update: this is the pixel
data that the OS and/or app put into the frame buffer.
Now consider the system bus: it's 64 bits wide, and clocks at 30 MHz on
a 6100, which is the only PMac with DRAM-based video on the motherboard.
Assuming a fairly ideal DRAM system with interleaving (which the PMacs
don't have, but I'll give it the benefit of the doubt), you could do
4-2-2-2 cycle bursts over this bus. Refresh accesses are always sequential,
so bursting will be the predominant access mode. This burst moves 32 bytes
in 10 cycles, so the aggregate bandwidth of the bus is (32/10)*30 = 96
MB/sec. This is already slower than 32-bit PCI's bandwidth in burst mode.
Half of this bandwidth (47 MB/sec) is immediately gone for refresh. The
CPU gets the remainder, both for line loads (cache misses) and for video
update. DRAM-based video suddenly doesn't look all that impressive, and
this was for a comparatively wimpy video mode with an (unacceptably) low
refresh rate.
The 7100 and 8100 are in better shape: because their VRAM is dual-ported,
refresh occurs over a separate bus. They therefore get the entire 105 MB/sec
or 128 MB/sec bandwidth of the system bus to play with for accessing either
DRAM or VRAM. Also, since the VRAM is now decoupled from system memory
completely, they don't have to use 4-2-2-2 bursts as I assumed above, and
can play some additional games to get up closer to the 210 MB/sec or 256
MB/sec theoretical limits of the system busses in the 7100 and 8100,
respectively. I cannot imagine how the CPUs in these systems could possibly
generate video update information quickly enough to overwhelm even a
100 MB/sec bus interface, however.
|> I'm not sure what the other video option uses, but it's clearly not
|> the NuBus (nor will it be the PCI bus in the next generation Power
|> Macs).
It uses an '040 PDS, as described above. Your second comment is plainly
wrong: from all public mention, the Tsunami (601-100 based tower system
for high-end applications) will use PCI for all video. You'd better call
apple right away and tell them they're screwing up :-)
|> If you need some sort of special purpose video card or simply want
|> a third monitor (like one my friends has - he has THREE 19" monitors
|> on his desk hooked to his Power Mac: two using on board video, the
|> third using a NuBus card) you can use the NuBus for that and it will
|> be slower than using on board video or PCI.
True.
|> __ Michael Peirce __ pei...@outpost.sf-bay.org
|> __ Peirce Software, Inc. __ 719 Hibiscus Place, Suite 301
|> __ __ San Jose, California USA 95117-1844
|> __ Makers of: Smoothie & __ voice: +1.408.244.6554 fax: +1.408.244.6882
|> __ Peirce Print Tools __ AppleLink: peirce & AOL: AFC Peirce
------------------------------------------------------------------------------
Patrick Chase
|>
|> The Pentium currently only has wins (that I know of) in the
|> desktop market while the PPC601 has several wins in the controller
|> market besides multiple desktop vendors (and not just Apple and IBM).
That's because the Pentium targets only the desktop market. If you want to
discuss the microcontroller market, why not also count i960 design wins?
I can rattle off quite a few for you of which I have immediate knowledge :-)
|> -John
|> SB Industries
|>
|> ****************************************************
|> This morning I shot six holes in my freezer
|> I think I've got cabin fever
|> Somebody sound the alarm
|> - J. Buffett
|> ****************************************************
------------------------------------------------------------------------------
LEE BRIAN
>Apple did not cripple the 65816 in the Apple IIGS. They just refused to
>go with a CPU that ran faster than 2.8 MHz. Supposably, the company that
>designed the 65816 (Western Design?) had versions of the chip that ran
>at ~25 MHz and also a dual CPU version. If I remember correctly, the
>fastest 68000 in a Mac at that time was 16 MHz. Further, the 65816 was
>slightly more efficient per cycle. Also, there was a press conference
>at that time where Apple representatives where explaining that the
>Mac was their future because they could not make the IIGS run any
>faster. After that comment, supposably the president of the company
>that designed the CPU charged the stage yelling "That's a lie! You
>know that's not true!" Therefore your theory on not letting the IIGS
>go faster than the MAC is probably true.
just a thought... maybe they wanted the system to be closer to the speed
of the original apple II's. I mean, since you are able to use apple II
hardware and boards (I think). Also, the software and things.
people who bought IIgs's bought it to run apple II software, not mac software.
who's going to pay for a 16/25MHz system (the board will be very expensive)
to run apple II software? lots of games and things I guess... that's all.
and they already run at decent speeds.
(just a thought).
later,
bjl
Patrick Chase
|>
|> But, there is a difference between not letting a faster CPU into your
|> machine and deliberately breaking the connection your CPU so you can
|> sell a slower version. Deliberately breaking your CPU may be a
|> wise marketing move, but it sure seems strange from an engineering
|> perspective.
They don't sell crippled chips. The 486SX has a redesigned die, sans FPU.
Some of the very first ships had disabled FPUs, but that was a marketing
move intended to get something on the market as quickly as possible. They
sacrificed their profit margin for a while in order to establish a market
for low-end 486 boxes. This in turn hosed AMD, which by that time was just
figuring out how to clone a 386...
|>
|> Mike
|>
|> =============================
|> Mike Shandony
|> Bell-Northern Research, Inc.
|> vanh...@bnr.ca
|> ====================================================================
|> The opinions expressed are my own and not necessarily those of BNR.
|> ====================================================================
-----------------------------------------------------------------------------
Srinivasan Seshan
: -David Lerner
You should also include packaging costs. I have heard the the die
size (and thereby heat dissipation) of the SX allows intel to use
plastic packaging (PLCC chips) instead of ceramic PGA packages. This
translates to a end user price difference of about $50.
--
Srinivasan Seshan s...@cs.Berkeley.EDU
University of California Berkeley Ph: +1(510)642-8248
Computer Science Division, 571 Evans Hall Fax: +1(510)642-5775
Berkeley CA 94720
Erik A. Speckman
Patrick Chase <pat...@sdd.hp.com> wrote:
>In article <CNjbKKK...@outpost.SF-Bay.org>, pei...@outpost.SF-Bay.org (Michael Peirce) writes:
>|>
>|> Both built in video options on the Power Macs don't go over over a interface
>|> bus (PCI on the Pentium or NuBus on the PowerMac).
>
>Actually, the "Processor Direct Slot" in the PowerMacs does not truly reside
>on the 601's system bus: instead, it provides an '040 bus interface, which in
>turn is bridged onto the system bus. Given the resulting throughput
>limitations, it's certainly no better than PCI. This is the interface used
>for PDS video and AV cards, BTW.
I am going to call your bluff. Here is what I have, can you beat it.
The video circutry on the AV card is taken from the 040 based AV
machines and adapted for the Power MAcs.
Since the A/V video hardware was designed for the 040 protocol conversion
must be performed between the CPU and video hardware on the PowerMAc AVs
but, the conversion circutry is located on the video card.
The interface between the card and the PowerMAc motherboard is a PowerPC
PDS, not an 040 PDS.
The PDS VRAM video cards in the 7100 and 8100 machines were designed to be
used with the 601, not the 040, thus they escape the overhead of
protocol conversion and they take advantage of a 64 bit data path to the
processor.
This is one of the resons the VRAM video is faster than the AV video.
I should also correct the misconception that only the 6100 has onboard,
DRAM based video.
All the machines have onboard DRAM based video but the 7100 and 8100
systems also ship with VRAM based video on a card in the PDS slot. It is
possible for the 7100 and 8100 to support two monitors out of the box by
attaching one to the connectior for DRAM video and the other to the
connector for VRAM based video.
--
____________________________________________________________________________
Erik Speckman espe...@romulus.reed.edu GBDS
John W. Komp
: In article <2t1ubl$r...@gazette.medtronic.COM>, jk0...@medtronic.COM (John W.
: |>
: |> The Pentium currently only has wins (that I know of) in the
: |> desktop market while the PPC601 has several wins in the controller
: |> market besides multiple desktop vendors (and not just Apple and IBM).
: That's because the Pentium targets only the desktop market. If you want to
: discuss the microcontroller market, why not also count i960 design wins?
: I can rattle off quite a few for you of which I have immediate knowledge :-)
I don't count the i960 because it is yet again another design (ie not
the pentium). What I was getting at is that appearance that the PPC
core is generic enough to be targetted at a variety of applications.
My reason for the statement was in counter to those that felt the PPC
was only on par with the x86 and had was of only limited usefulness
(basically some Apple products and a few IBM workstations).
-John
SB Industries
--
Paul Winterrowd
I believe the original poster is referring to the rumor that the original
486SX were actually 486DX's with the math coprocessor disabled. In other
words, the dies were the same size.
Don't ask me if this rumor was true, I can't remember where I read it and I
don't really care (I myself have a 486DX and a PowerMac). Being from Intel
perhaps you could expound upon it.
: -David Lerner
: ============================================================================
: David Lerner --------------------------------------------
: Intel Corporation | The views contained within this post are |
: 2200 Mission College Blvd | a result of my own ridiculous opinions. |
: Santa Clara, CA 95052 | Some settling may have occured while in |
: (408) 765 5326 | transit, so interpret with caution. |
: dle...@t12sys.intel.com --------------------------------------------
: ____________________________________________________________________________
:
Paul W.
pa...@crl.com
--
-----------------------------------------------------------------
I only speak for myself.
Mark J. Lilback
>In article <2sthmj...@jhunix.hcf.jhu.edu>, Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> writes:
>|>
>|> I doubt this is really what you think, and is more likely part of
>|> Intel's FUD campaign...
>
>Do you really think Intel or any other major company would actually bother
>to conduct a "FUD campaign" on USENET? If so, then you have a greatly
>exagerrated sense of self-importance...
Actually, some are. There is a commerical service that offers to make X
number of posts to relevent newsgroups talking up a companies software
-- like recommendations that we see everyday. It's really not a bad
marketing idea -- if someone asks for advice on purchasing a products
and gets multiple responses from different sites advocating a particular
brand, they might just buy that brand. As marketing people like to say,
the best advertising is word of mouth...
Not that this means that Intel, Apple, or any other company is doing
this with their employees (which I think would be a bad idea), but
companies are doing this.
************************************************************************
* Mark J. Lilback ****** Computer Guru, Who Cares Magazine *
* mlil...@cec.org ****** A Journal of Service and Action *
* mlil...@seas.gwu.edu ****** whoc...@cec.org or AOL: WhoCares3 *
************************************************************************
Patrick Chase
|> Patrick Chase (pat...@sdd.hp.com) wrote:
|> :That's because the Pentium targets only the desktop market. If you want to
|> :discuss the microcontroller market, why not also count i960 design wins?
|> :I can rattle off quite a few for you of which I have immediate knowledge :-)
|>
|> I don't count the i960 because it is yet again another design (ie not
|> the pentium). What I was getting at is that appearance that the PPC
|> core is generic enough to be targetted at a variety of applications.
|> My reason for the statement was in counter to those that felt the PPC
|> was only on par with the x86 and had was of only limited usefulness
|> (basically some Apple products and a few IBM workstations).
From the standpoint of philosophical purity, I'll agree that the PPC's
demonstrated versatility is quite nice. From a practical standpoint (which
is all that really matters IMHO) I don't see how it matters all that much,
unless Ford is planning on running MacOS or PowerOpen on ever car they
build.
The Pentium definitely pays a significant performance and/or cost penalty
for backwards compatibility. Intel believes that this penalty is justified
for the desktop market. We'll see whether they're right or not. Binary
compatibility is much less important in embedded-control applications.
Since the PowerPC carries no such baggage, it is suitable in modified
form as an embedded controller. Relatively unsuccessful attempts like
the 80376 aside, the x86 family is not. Therefore Intel did what they had
to and went with a new architecture. It seems to me that they made the
right business choice given the state of the market and of their existing
products. What's the problem?
|>
|> -John
|> SB Industries
|> --
|> ****************************************************
|> This morning I shot six holes in my freezer
|> I think I've got cabin fever
|> Somebody sound the alarm
|> - J. Buffett
|> ****************************************************
----------------------------------------------------------------------------
Erik A. Speckman
Joe Bennett <jben...@frx146.intel.com> wrote:
>They didn't have to pull out, and even when they did, there was a
>nice, stable, high performing x86 out there they could have used.
>They just decided they had to be different.
Yeah, in retrospect, it was a pretty stupid choice, wasn't it?
After all, they picked a chip that gives them the same performance at
what, 1/2 the price? At the same time the 601 gave them decent
performance without the extra cost of a 60-66 MHz L2 cache subsystem
for an even greater savings.
Sure, they could have gone out of the hardware business alltogeather but
then they would have had to contend with porting their software to a
million different motherboards, video cards, disk controllers, etc.
What else does the x86 offer Apple, besides higher prices for the
same performance? Exactly the wrong byte ordering, making their port
more difficult and dragging down all their independent hardware software
vendors to boot.
Yeah, they really should have gone with the x86 line. I never cease to
be amazed by your objective and well informed view of the entire computer
industry, really.
Dennis O'Connor
jk0...@medtronic.COM (John W. Komp) writes:
] : In article <>, jk0...@medtronic.COM (John W.
] : |> The Pentium currently only has wins (that I know of) in the
] : |> desktop market while the PPC601 has several wins in the controller
] : |> market besides multiple desktop vendors (and not just Apple and IBM).
]
] : That's because the Pentium targets only the desktop market. If you want to
] : discuss the microcontroller market, why not also count i960 design wins?
]
] the pentium). What I was getting at is that appearance that the PPC
] core is generic enough to be targetted at a variety of applications.
I'm not sure what you mean by "PPC core", but I think that the
IBM 400- and Motorola 500-series PPC parts being prepared for
the embedded market are not the same silicon as the 600 series
parts that are being sold into the "user reprogrammable"
(i.e. PC and worksation) market. This is to be expected : the
embedded market has different needs.
So, if we don't use the word "core" but instead use the word
"architecture", well, then you should probably compare the
PPC desktop+embedded effort to "Intel Architecture" family :
Pentium(tm) chips, 486's, 386's, the new 386 embedded parts,
and the venerable 186 line. There's a lot of embedded x86
already out there, and more on the way.
] My reason for the statement was in counter to those that felt the PPC
] was only on par with the x86
From many points of view, being on a par with x86 shouldn't
be considered a bad thing :-)
] and had was of only limited usefulness
] (basically some Apple products and a few IBM workstations).
Rest assured that people in Intel's embedded processor groups
are not pooh-pohing the PPC potential as a competitor.
--
Dennis O'Connor doco...@sedona.intel.com
Intel i960(R) Microprocessor Division Solely responsible for what I do.
"When you discover a fluke, first determine if a whale is attached to it."
Tony Hamilton
> Seems to me that Intel is *forcing* users to pay for backwards-compatibility.
Is this an attempt at being sarcastic just to spite Joe, or are you serious?
In any case, it isn't a correct analogy at all.
Joe, and I completely agree with him, is talking about things which are
statistically significant. I think there are something like 120 million or
more x86-based systems in use out there, and I'm not sure what such numbers
represent (all markets, just business, or what), but I guess there are
something like 50+ million Windows 3.1 users out there. That means everyone
else at least is using _some_ 16-bit or worse applications, and I know that
it is likely that _most_ Windows users are as well. So, I'd say from 80 to
90% of all x86 systems are still running 16-bit software.
You can't even come close to claiming numbers like that for Mac users who
run networks, multiple monitors, and so forth and so on.
Why the argument over all this is beyond me. Macs target niche markets, and
thus have to accomodate more functionality at higher prices.
x86 PC's target mass markets and are thus cheaper in base configurations.
Apple commands 10% of the market, while x86 has 85% or so.
To me, this is all fact, and perfectly reasonable. What is confusing is
why Macs should be pushed as mass-market PC's, or x86 systems as high-end
niche-market machines. If either were appropriate for the other market,
they would have been there already. Joe, myself, and many others simply don't
see the PowerMacs as adding any new variables to this whole equation. We
_do_ see 32-bit OS-running, multimedia-ready, plug-n-play and PCI-compliant
Pentium systems making _big_ changes to the equation next year. And therein
is the underpinnings of why Intel thinks it can win over Mac market share.
Really, the arguments on either side aren't much different:
- Intel says buy into the x86 architecture which is opening up all kinds
of new possibilities given what I just mentioned in the above paragraph.
- Apple says buy into PPC, which once lots of native software is out, will
also enter into new realms of possibility.
Either way, with either architecture, there are drawbacks today.
Whoops, think I'll step down from the soap box now. In any case, I had
to respond to Lawson's comment above because, sarcastic or not, analogies
are very dangerous unless constructed properly. I find that Mac enthusiasts
are good at drawing all _kinds_ of wonderful analogies, like between chip
prices and system prices, catchy architecture buzzwords and survival, apples
and oranges, and oh the list goes on...
(hey, Joe isn't the only Intel employee allowed to vent his frustration at
all this silliness, ya know... ;-)
--
TONY HAMILTON ==================================================
Intel Corporation "Have compiler - will travel"
Tony_H_...@ccm.fm.intel.com -or- tham...@pcocd2.intel.com
========================(I speak only for myself ... not Intel)=
Tony Hamilton
> > to conduct a "FUD campaign" on USENET? If so, then you have a greatly
> > exagerrated sense of self-importance...
> >
>
> Maybe not the companies, but certainly the employees of various companies feel
> compelled to disseminate it.
>
> Lawson
Look, the only FUD campaign I see is the one which goes "Hey! The Intel
architecture is CISC, old, <insert adjective of your choice here> and so
therefore is DOOMED. Buy PPC!"
I have yet to see any genuine FUD come from frequent posters at Intel,
including Joe and Dennis (nothing personal guys, but you _are_ picked on
the most ... :-). They're responding to FUD themselves, and the counter-
responses tend to degenerate into a flurry of non-sequiturs. The first thing
that always happens is that the carefully drawn-out context that someone like
Joe makes is quickly discarded in order to provide for more lively argument.
What's worse is that I see "us" (the Intel net-commoners) frequently
acknowledge valid counter points, admit mistakes, and carefully qualify what
we are saying with reasonable disclaimers. I _don't_ see it at all, with
rare exception, from those on the other side of the debates. Sure, "we" all
get emotional at times, but only after we've carefully laid out our logic
only to be pummeled by reams of FUD-lashing Mac enthusiasts. The fact that
some Mac folks provided good counter-arguments tends to get lost in all the
noise. I for one, when am presented with an argument that I can't counter,
tend to just leave that thread alone. Under normal circumstances, that poster
should have had the final word, and it would be clear that the "Intel guy"
quietly acknowledged defeat. Instead, 10 others chime in to beat the dead
horse with mis-information, and when I or another Intel tries to correct these
horse-beaters, we're seen as "never giving up" and "spreading more FUD". Well,
that's how I often feel.
Okay, so all this was really just meant to say "We're NOT spreading FUD!
we're NOT we're NOT we're NOT!" :-P
Patrick Chase
|> Since the PowerPC carries no such baggage, it is suitable in modified
|> form as an embedded controller. Relatively unsuccessful attempts like
|> the 80376 aside, the x86 family is not.
Major foot removal operation presently underway. As Dennis points out in
another post, the 80186 has been quite successful as a microcontroller, and
there are embedded 386 variants out there.
-----------------------------------------------------------------------------
johnson r scott
some pretty harsh criticism of the 68LC040 chip (the Motorola chip without
a math co-processor) as well as some FUD about the PowerPC not having a
FPU and thus being an inferior chip vis-a-vis the 486 & Pentium that
originated from the Intel crowd. Now we hear that Intel calculated that:
Intel has always
said that they studied the DOS software market place and fewer than
1% of all programs use enough floating point to see any increase in
speed. Hence, a non-floating point version was an obvious product.
Seems to be a double standard here.
S. Johnson
Dennis O'Connor~
pat...@sdd.hp.com (Patrick Chase) writes:
] In article <>, pat...@sdd.hp.com (Patrick Chase) writes:
] |> Since the PowerPC carries no such baggage, it is suitable in modified
] |> form as an embedded controller. Relatively unsuccessful attempts like
] |> the 80376 aside, the x86 family is not.
]
] Major foot removal operation presently underway. As Dennis points out in
] another post, the 80186 has been quite successful as a microcontroller, and
] there are embedded 386 variants out there.
Hey, no one knows everything, so it's no big deal.
( No, wait, I take it back : _almost_ no one knows everthing. :-)
Dennis O'Connor~
srjg...@uxa.cso.uiuc.edu (johnson r scott) writes:
] Not to start a flame or anything, but it seems to me that there were
] some pretty harsh criticism of the 68LC040 chip (the Motorola chip without
] a math co-processor)
Why would anyone from Intel criticize that ? It's just meeting the
customer's needs, just like Intel did with the Intel486SX. Maybe
your misreading why it was mentioned : people get tired of "486SX
bashing", so they mention that competing companies have seen the logic
of it.
] as well as some FUD about the PowerPC not having a
] FPU and thus being an inferior chip vis-a-vis the 486 & Pentium that
] originated from the Intel crowd.
I thought all the 600-series parts _had_ an FPU built in. Who ever
said any different ?
] Intel has always
] said that they studied the DOS software market place and fewer than
] 1% of all programs use enough floating point to see any increase in
] speed. Hence, a non-floating point version was an obvious product.
]
] Seems to be a double standard here.
I think your misinterpretting what was posted. I've never met or
corresponded with an Intel person who didn't see the justifications
for both the 486SX and the 68LC040.
It's called, giving your customer what they think they want.
Dan Hildebrand
Steve Kanefsky <kane...@halcyon.halcyon.com> wrote:
>
>What if I don't want a parallel port or serial ports or a mouse port or
>SIMM slots or an overdrive socket or ISA slots, etc. etc.? Should each of
>these be optional equipment and only provided if specified, or is it
>beneficial to have a certain minimum level of functionality? In today's
>world, it's beneficial to have a certain minimum level of capability to
>communicate with other computers. Ethernet is now cheap enough that it
>can serve as this minimum level.
Personally, I'd love to see the proprietary "high-speed" serial ports that
some modem vendors are creating for >115 Kbaud be replaced with Ethernet
ports instead. Rather than trying to push RS232, with it's interrupt per
character (or per 14 characters with a 16550), it would be much nicer to
use Ethernet with an interrupt per packet, providing more than enough
throughput and a much lower CPU load. With the new bus-mastering AMD
79C960/79C965/79C970 single chip Ethernet for ISA/VLBus/PCI out now, and
already appearing on the motherboards of some PC vendors, modem vendors
could provide a serial port and an Ethernet port, and not bother with the
parallel port that some are doing right now. With the newer OS's going in
the direction of providing TCP/IP as a standard OS component (MS Windows
4.0 (AKA Chicago), WNT, OS/2, most UNIX's, etc), the modem wouldn't even
require custom drivers.
--
Dan Hildebrand email: da...@qnx.com
QNX Software Systems, Ltd. QUICS: danh (613) 591-0934 (data)
(613) 591-0931 x204 (voice) mail: 175 Terence Matthews
(613) 591-3579 (fax) Kanata, Ontario, Canada K2M 1W8
Chris Umbricht, M.D.
jben...@frx146.intel.com writes:
>There isn't any real, significant reason
>the PPC was chosen. If you look at all performance accounts, they
>are similar. There are some supposed advantages in bit-wise instructions
>that do make emulation a little bit better on the PPC, but on many
>other accounts, they are the same animal.
>
>However, you can talk to any marketting individual for Intel processors
>that you want, and you will find that Apple never really considered
>using the x86, although they were wooed (and still are being wooed)
>pretty hard. It's pretty easy to turn something down that you aren't
>really considering anyway.
>
>Most people *I* have come across either at Intel or who use a PC, wish
>Apple would port the darned thing and behave like the excellent OS
company
>that they are.
No kidding. If Apple had gone with '86 architecture, Intel would indeed
be sitting pretty, right? Now, how likely would it be that we would be
seeing this accelerated price erosion and development cycle of the P5 and
successors? Not bloody likely, as some might put it.
While the first generation 601 is indeed in the same ballpark as far as
performance goes, the price sure isn't... and I seriously doubt the
development cycle of the P5 descendants will be able to keep even close
the PPC. Just look at the Specint (!) of the 125MHz 601 or the 100MHz
604, which are in production now. The 150MHz 604 blows anything away I've
seen from Intel so far.
So tell me again (with a straight face) you can't think of a reason Apple
chose to go with the PPC chips...
Chris Umbricht, M.D.
vl...@panini.cs.jhu.edu writes:
>In case you were completely unaware of it, apple did sell a line of
mac's
>based on a '68040 SX', a '040 w/o math co-processor. VERY much like the
>486sx, if you ask me (or anyone else).
I am aware of it, and I think it has caused a lot more grief than
benefits. Defending a decision does'nt mean defending all of them. IMHO,
Apple has made more than their fair share of stupid decisions. But PPC
was not one of them.
BTW, I'd appreciate having the context presented when being quoted.
Chris Umbricht, M.D.
dle...@iws732.intel.com writes:
> Your understanding of the microprocessor business appears to be
minimal.
>Joe Bennett is right- the PC market is customer driven like all others.
The
>Intel486SX(tm) is not "crippled," and it was created to provide a
low-cost solu-
>tion for customers. To understand this better, learn about the
relationships
>between defect density, die area, and yield. Since you probably don't
have
I agree my understanding of the mp business is probably minimal compared
to yours, but it may well be less biased, or at least less historically
impaired by selective losses...
Unless my recollection is faulty, the initial 486sx was indeed crippled,
whether by design or as discovered during testing, who knows? So it had
the same die size.
Seeing how the market responded to lower cost chips, it probably made
sense to design sx-type chips for the reasons you so eloquently state.
Nobody is saying Intel marketing isn't smart (but the reviews are mixed
as to chip architecture...;-) But I think it would have been in everyone
else's interest to crank out more 486dx's. It doubt that would have been
more expensive than the overhead of duplicate designing, manufacturing
and marketing, but it would have put pressure on the profit margins of
the full chips, which I suspect is why it was not done.
But I still think that was in Intel's interest, not the user's. After all
this helped create the hoards of PC's with abysmal FP performance,
ensuring nobody would program FP functions when it could be avoided in
any way, reinforcing the circular argument that FP performance is
unimportant since no programs use it...
Clint Olsen
Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> wrote:
>
>No kidding. If Apple had gone with '86 architecture, Intel would indeed
>be sitting pretty, right? Now, how likely would it be that we would be
>seeing this accelerated price erosion and development cycle of the P5 and
>successors? Not bloody likely, as some might put it.
>While the first generation 601 is indeed in the same ballpark as far as
>performance goes, the price sure isn't... and I seriously doubt the
>development cycle of the P5 descendants will be able to keep even close
>the PPC. Just look at the Specint (!) of the 125MHz 601 or the 100MHz
>604, which are in production now. The 150MHz 604 blows anything away I've
>seen from Intel so far.
If Intel hyped as much about it's future products as Moto and IBM
do, you'd be impressed too.
What's with you people touting stuff that may not even get here by
the end of the year?
It's one thing to discuss plans and strategies for future products,
but it's getting a little boring hearing, "Just wait until 6XX!!
Then you'll see!"
Give it a rest.
-Clint
Burkhard Neidecker-Lutz
>
> < discussion about yield deleted>
Mmh. So why doesn't die area translate to chip price ?
Chip est. Price est SPECint 92
---------------------- ---------- --------------
Mips R4200PC-80 $ 70 55
Alpha AXP 21068-66MHz $ 186
PPC 603-66 $ 200 60
PPC 603-80 $ 250
Mips R4600-100 $ 300
Alpha AXP 21066-166MHz $ 347
Alpha AXP 21064-150MHz $ 404
Alpha AXP 21064-166MHz $ 404
SuperSPARC-40 $ 450
PPC 601-66 $ 470
Alpha AXP 21064-200MHz $ 544 130
PPC 601-80 $ 625 80
P5-60 (Pentium) $ 670
Mips R4400PC-150 $ 700
P5-66 (Pentium) $ 750 70
Alpha AXP 21064A-233MHz $ 788
P54C-90 (Pentium) $ 850
Mips R4400SC-150 $ 850 95
P54C-100 (Pentium) $ 1000 100
SuperSPARC-60 $ 1000 85
Alpha AXP 21064A-275MHz $ 1083 170
Burkhard Neidecker-Lutz
Distributed Multimedia Group, CEC Karlsruhe
Advanced Technology Group, Digital Equipment Corporation
nei...@nestvx.enet.dec.com
Hyong Chol Lee
>In article <2t2h6t...@jhunix.hcf.jhu.edu>,
>Chris Umbricht, M.D. <cb...@jhuigf.med.jhu.edu> wrote:
>>
<stuff deleted>
>>Just look at the Specint (!) of the 125MHz 601 or the 100MHz
>>604, which are in production now. The 150MHz 604 blows anything away I've
>>seen from Intel so far.
>^^^^^
>If Intel hyped as much about it's future products as Moto and IBM
>do, you'd be impressed too.
But they do. We've already heard about P6 and the possible rampup to a 200 MHz
Pentium.
>What's with you people touting stuff that may not even get here by
>the end of the year?
>It's one thing to discuss plans and strategies for future products,
>but it's getting a little boring hearing, "Just wait until 6XX!!
>Then you'll see!"
>-Clint
This is a good point. It'd be nice if folks could stay with the 80 MHz 601 and
the 90 MHz Pentium performance numbers.
--
- hyong
hyo...@cco.caltech.edu
"You do not like them. So you say. Try them! Try them! And you may.
Try them and you may, I say."
David Smith
>In article <1994Jun4.195548.24620@dal1> sea...@ac.dal.ca writes:
>>In article <2sq729$e...@inews.intel.com>, jben...@frx146.intel.com (Joe Bennett) writes:
>>> In article <1994Jun04.0...@zeus.aix.calpoly.edu> bp...@harp.aix.calpoly.edu (Piner) writes:
>>>>In article <2sks2c$c...@inews.intel.com>,
>>>>Joe Bennett <jben...@frx146.intel.com> wrote:
>>>
>>Multiple monitor support is not some far-out thing that only a few users
>>might possibly use. Take PowerBooks, for instance. You complain about having
>
>1) Very few people have notebooks and desktop machines.
Huh, many people who have notebook machines have, or would like to get a
"real" monitor (and keyboard too" for when they are at home or in the
office. Also for presentations. You can plug a nice big monitor, or a
projection screen into a powerbook, MUCH better than either moving the
files or lugging a desktop machine around, or hoping that one will be
avalible.
> eyed as Mac people do over the built in Ethernet, when I have yet
> to find the Ethernet connection in my house that warrants the use of
> it.
You have a laserprinter? Most now support Ethernet, which provides MUCH
faster printing than Appletalk. As to if Mac's NEED networking at all,
it seems that the world is going to networks, education and busnesses
already have almost universial networking, and ISDN, the "info highway"
or whatever is promising to bring it into the home. Since the ethernet
is just the drivers it isn't much to include it, (the connectors are most
of the money once the thing is built).
>4) These may be things that some Mac users really like. I have never
> argued that they shouldn't like it. However, if Apple truly wants
> to lure customers from the PC to the Mac, switching the processor
> like they did isn't the right strategy. If they would only look at
> the rest of the computer industry, they might realize, "Oh, so THIS
> is why we are only at 11% of the market!"
Huh, I belive that Apple is the THIRD LARGEST SHIPPER OF COMPUTERS. %11
sounds like a small amount but it is only one company... This could go
to the debate of Cloning Mac's though.
>Until somebody in the Mac arena can argue that it comes for free, I have
>to continue to discount hand waving arguments like you make. Since you
>do NOT know how much some of the built-in-only-used-by-the-minority
>functions actually cost, how do you know the 6100 (sans monitor and
>keyboard) wouldn't cost, say, $1300, instead of $1700, and thus giving
>Apple lovers something worth actually arguing about in a cost/performance
>debate?
Your right, although there are Apple engineers on here that might know.
Since most of teh stuff that is built in is not large hardware, it is
development costs, that any R&D manufacture faces, that we are paying
for. So how many people buy IBM (not clone) PC's? I applaud that Apple
has not cloned the Mac. Yes it would likely be more accepted, but look
at IBM, losing millions of $'s a year, and in jepordy of going out of
busness, because of compations and bad deals with clone makers who were
able to make 'em cheaper because of both lack of R&D expenses in general
and, in some cases, "substandard" parts.
--
---- David A. Smith | The opinions above are mine and do not reflect
- smi...@cs.csee.usf.edu | those of the University of South Florida.
- das...@suntan.eng.usf.edu |
| PGP/RIPEM key avalible via finger.
103t_e...@west.cscwc.pima.edu
>>
>
> I got my info by trying them out. I have never trusted PC mag, to be honest.
> Also, if you run windows on either machine, neither machine will reveal its
> true performance capabilities (emulation or no, ms windows will ream performance
> quite badly, which is why I run Linux and X instead, it is faster if the system
> has enough memory)
>
But, no UNIX (including Linux) is available on the PowerMacs yet. How are you
able to compare them?
Lawson
Mark Decker (guest from Wm Jewell College)
> We _do_ see 32-bit OS-running, multimedia-ready, plug-n-play and PCI-compliant
>Pentium systems making _big_ changes to the equation next year. And therein
>is the underpinnings of why Intel thinks it can win over Mac market share.
>Really, the arguments on either side aren't much different:
>- Intel says buy into the x86 architecture which is opening up all kinds
> of new possibilities given what I just mentioned in the above paragraph.
>- Apple says buy into PPC, which once lots of native software is out, will
> also enter into new realms of possibility.
Yeah, but the 'new possiblities' that you mentioned in the above paragraph
are NOT 'new'. The Macintosh already has them. So the difference seems
to be that Intel is promising future PCs to have Mac features. And
Apple (which already has these features) is the one 'actually' promising
new possibilities.
1) 32-bit OS (apple has)
2) multi-media ready (apple has)
3) plug-n-play (apple has)
4) PCI, (apple doesn't have,... but like PCs it will have it in the future)
But unlike PCs,... when they switch over on the Macs, vendors will
know that they can count on it, in ALL new systems.
mkdk
David Lerner
I hope I did not imply in my earlier post that chip price (by this I am assuming
you mean the price which the product is sold at) is determined by die area. If I did,
sorry. What is _generally_ true is that die manufacturing cost is proportional to die
size, all other things being equal, i.e. number of mask layers, geometry, and other
assorted characteristics of the process. As an earlier posted pointed out, this
does not include package cost, which for the same package type is roughly independent
of die size. (to sort of answer that earlier question about SX vs. DX package costs,
you can look inside PCs and find IntelDX2(tm) parts in plastic packages and
Intel486SX(tm) parts inside of ceramic packages).
Product price is not determined by manufacturing cost, however. It all comes down
to supply and demand. If DEC can get $1083 for an Alpha AXP 21064A-275MHz, and cannot
meet all the demand there is for that part, they'd be foolish to sell it for less.
Same goes for PPC, Pentium(tm), and baseball tickets. In all honesty, however, I'm an
engineer, not a finance or marketing guy, so I may not know what I'm talking about.