Conference 7.286::digital

From what little I've read, a PowerPC can run DOS/Windows or MAC software.
(with some form of emulation of each, so you don't get anything like full
'native' speed of the risc machine).

Alpha runs neither.

Wish it weren't so.  Just how fast could an ALPHA emulate a 486?

	Joe

    > Just how fast could an ALPHA emulate a 486?
    
    From the tests in various PC magazines, not very fast. Besides, who
    would pay Alpha prices for 486 performance? Alpha PC's are priced at or
    above Pentium PC's. Alpha would have to beat the Pentium's performance
    to be a player. Aplha systems just don't seem to be that fast unless
    they're running native mode applications.
    
    Harry
    

    Most forecasts out of Apple say they plan to ship 750,000 systems in
    the first year of availability....makes Alpha AXP systems of ALL kinds
    from ALL vendors look like background noise 2-3 years after
    availability.  And the Apple numbers do not include IBM and other
    systems sold.
    
    We're outmarketed, outsold and outperformed in this market.  If we look
    at major chip technologies, we're clearly behind Intel, AMD, SPARC,
    Motorola, PowerPC, and others in terms of volumes/market share.  We've
    got a long uphill battle.
    
    Mark
    

    Too many people don't understand PowerPC. The "real" one doesn't exist
    yet nor does an operating system, compilers, or applications. 
    
    Which "PowerPC" is everyone referring to, 601, 604, the PReP reference
    platform? They all have different instruction sets and the PReP spec
    (the latest ACE wave) is still being negotiated. Sun wants S-bus, Apple
    wants NuBus, etc.

    re .5;
    
    That's great news.  Now go out on the street and ask people if they've
    heard of PowerPC.  Then ask if they've heard of ALPHA!
    
    It's amazing how a little good marketing can make up for a certain
    lack of technical prowess.

    Re: .1
    
>Alpha runs neither.
    
    Wrong.  Alpha runs DOS and Windows applications the same way that the
    PowerPC 601 does -- with software emulation, and slowly (but not as
    slow as a PowerPC).
    
>    Just how fast could an ALPHA emulate a 486?
    
    Current systems running Windows NT or OSF/1 with SoftPC, get you
    about 486sx performance.  With the next release of Windows NT, you
    should get 40MHz 486 performance (just a WAG, folks -- your mileage
    _will_ vary!)  There are also other technologies available -- see page
    11 of the April 4 PC Week -- but from what I hear, there are a lot of
    nay-sayers regarding add on hardware accelerators.
    
    
    A couple of data point for native applications:
    
    	- There are over 500 shipping native Alpha AXP Windows NT apps
    	- There are over 1,500 committed apps
    	- Recently I saw a list of apps for the PowerPC; there were 
    	  under 100 apps listed
    	- The slowest Alpha is faster than the fastest PowerPC, and
    	  we're not stopping at that
    
    My personal feel is that the PowerPC will succeed as an heir to the
    Macintosh throne and that IBM will be successful in adding additional
    hodge-podges of incompatible versions of the Power architecture to
    their workstation base and will also be able to sell a version of OS/2
    for it.  All well and good for their installed bases.
    
    To sum it up: (IMHO)
    
    - Alpha will succeed technically.
    
    - PowerPC will succeed because of the marketing pull.
    
    - Intel will succeed because of the installed base.
    
    
    Now, which is most important to your customer?
    
    Aaron

  A PC is a computer you use to run widely available, shrink-wrapped
  personal-productivity applications and to connect to departmental
  and enterprise-wide shared resources.  (wordprocessor, spreadsheet,
  graphics, DTP, education, entertainment, terminal emulation, database,
  laptops, notebooks)

  A Workstations is a scientific and engineering tool that usually
  features one or more software development environments and associated
  tools (VMS, UNIX, NT, language compilers, debuggers, X windows, toolkits,
  APIs, multi-protocol network stacks, client/server tools, relational DBMS,
  etc.)

The total PC market is around 40 million units this year.
The total Workstation market is around 1 million units.

Digital is perceived and perceives itself as a workstation vendor.

Today there could well be over 500 native applications for Alpha
compared to less than 100 for Power Macintosh.  It doesn't matter. 
Alpha is not "positioned" as a PC.  In a year Alpha will still be
struggling to get applications.  PowerPC will be mainstream.

Don't get me wrong, Alpha is important to Digital and our installed base,
but unless we start seeing ourselves as a PC company, we will continue
to be relegated to a niche role in the computer industry.

Notice how much attention the "IBM PC Company" is getting
inside and outside IBM?  IBM intends to no longer be dependent
on the mainframe/minicomputer business for most of its revenue
in 5 years.

Something to think about.

- Peter

    re: .4 and the existence of PowerPC ...
    
    PowerPC does indeed exist. I've already seen about 200 PowerMac's
    installed in the last two weeks at my customer.  The number of people
    who have come by to see them is incredible, so I expect even more to
    show up as soon as Apple can ramp production.
    
    I can't recall how long it took between the introduction of Alpha as a
    chip and when actual systems started taking off, but I don't recall it
    being anywhere near as fast as PowerPC.
    
    Geoff

	Yea, but you can't and never will be able to put Windows NT on
	an Apple PowerPC Mac. Apple doesn't want to deal with the PreP
	standard.

	DEC Marketing and the SLT is to blame on the Alpha not getting
	the visibility and market share that PowerPC is getting. We
	need to turn this around. I sure hope Enrico can do something.

						mike

>    Digital is perceived and perceives itself as a workstation vendor.
    
    Maybe you see Digital that way, but I don't believe that is or should
    be the common view anymore.
    
    Consider Sable -- more power than an 8 processor SUN UNIX server with
    pricing and I/O and software of a PC.
    
    Or consider Avanti and its follow-ons.
    
    The grey line between workstations and PCs has been crossed.  They are
    now one and the same.  It no longer matters what hardware you're
    running on to determine if it's a workstation or PC.  I've got a 60 MHz
    Pentium on my desk with NT, C/C++, X-windows, and TCP/IP on it.  A
    workstation, by your criteria, right?  Oops, I neglected to tell you
    that I also have Microsoft Excel, PowerPoint, and Word on it.  A PC,
    right?  Did I mention that it's also an FTP, DECnet, and LAN Manager
    server? 
    
    I've got customers buying Jensens just for that reason -- to get rid of
    the Silicon Graphics workstation on their desks that sits right beside
    their DOS/Windows PC.  Now they have one desktop system.
    
    I'm glad that Digital was the one to destroy that line between
    workstations and PCs.  I hope that I'm still around when we capitalize
    on it.  
    
    We need to be out there making it very hard for people to justify
    buying HP, SUN, IBM, and SG workstations.  I'm doing my part,  Mr.
    Lucente's parting shot notwithstanding.
    
    Aaron   

    
        RE .10
    
    
        So what was Mr. Lucente's parting shot, for those who 
        didn't hear it?
    

    PowerPC (to me) is little more than a followon to the 68000 in concept.
    Remember all the platform architectures that utilized the 68000 (Sun,
    Apollo, Amiga, Mac, NeXt).
    
    Now, when someone tells me that the PowerPC is real because Apple has
    one and that there are 100 applications (all on Mac?), I say that's
    great for Apple but all the money IBM is spending on marketing (the
    perception of) a system is just blowing hot air. People don't build
    operating sytems and applications for chips. For now, there is only one
    system that has native code for a PowerPC.

>            So what was Mr. Lucente's parting shot, for those who
>            didn't hear it?
    
    I expect that I'll lose my job tomorrow.
    
    Cheers,
    Aaron

>    Maybe you see Digital that way, but I don't believe that is or should
>    be the common view anymore.

Glad to hear it!   [I'm on your side].

So where can I pick up a fully configured AlphaPC with Microsoft Office
for <$3000?  Will it work with my HP Laserwriter or Laserjet?  What
about those low cost disk and tape drives advertized in the PC rags?
What do I need to use Novell?

Can I drop by Sears or Lechmere or CompUSA to pick it up tomorrow?

Did you see the latest issue of AlphaUser magazine that reviewed
Database apps?  I'm trying to decide between MS-Access, FoxPro,
and RDB.  Is RDB cross platform compatible?  How many "Alphas" did
it get?

- Peter

    re: .12  You can't separate systems from chips ...
    
    >Now, when someone tells me that the PowerPC is real because Apple has
    >one and that there are 100 applications (all on Mac?), I say that's
    >great for Apple but all the money IBM is spending on marketing (the
    >perception of) a system is just blowing hot air. People don't build
    >operating sytems and applications for chips. For now, there is only one
    >system that has native code for a PowerPC.

    You're missing a real key point here: Volume is the name of the game
    in making microprocessors. Apple, with it's projected 750,000 units,
    represents a CPU chip volume equivalent to *seventy-five* percent of
    today's workstation CPU chip market. And that volume represents a
    very stable base for Apple and Motorola to recoup their engineering
    investment and production ramp-up costs. Better yet, it gives Moto
    the cash flow to invest in broadening the PowerPC product line and
    targeting other market niches, like auto electronics and consumer
    electronics.  Digital doesn't have the volume to compete with just
    internally-generated demand. We *have* to establish Alpha as a
    product in itself, and sell it to the outside world.
    
    Geoff

    My point was not to cringe in the face of mega marketing of the PowerPC
    chp. Apple only has 10% of the PC market (which I admit is more than
    others) I don't slight Mac either as I have 4 of them!
    
    While I agree that our chip business may need some catching up, our
    system business does have a pretty good story to tell relative to the
    PowerPC systems.

    I agree that volume is key. That volume is not going to come by selling
    fully-configured Alpha PCs for $4995. We need to get the prices lower
    than that of 486/33 systems. That's right. Lower. People buy value to
    them, they do not buy technical excellence.
    
    Volume comes by selling cheaply, and selling to the mass market.
    The "workstation" market is not it. The PC market is just one of them.
    
    Alpha is fast enough to supply full-speed JPEG (?) compression and/or
    decompression to digital media. So they should be in VCRs.
    
    Alpha is sophisticated enough to do this compression/decompression
    while doing other functions.
    
    Alpha is fast enough to keep up with virtual reality games. So sell
    them to game companies.
    
    Alpha is fast enough and sophisticated enough to do real-time analysis
    of automobile engine functions. So sell them to auto manufacturers as
    the onboard computer.
    
    (this is getting repetitive. See note 3006.0)
    
    							Thanks
    							Jon

    <from livewire>
	

    My chip is faster than yours 

    Digital Equipment's development of the Alpha AXP microprocessor appears
    based on the thesis that the demand for computing power has a way of
    rising to tax the available hardware - so it is best to always have
    just a little more power available.  The Alpha processes integers 64
    bits at a time, which means it can communicate with a lot more RAM -
    more than seems necessary.  But analysts believe that present midrange
    servers will soon be hitting the limits of their 32-bit processors, and
    entertainment computing could very soon be demanding that sort of
    power. Digital has it now; it is just waiting for the customers to
    catch up.

    Forbes. English. 1994-04-25. Size: pp 162-164



>          . Digital has it now; it is just waiting for the customers to
>    catch up.                   ====================================
						^
			                        |

			            That's the problem!
			

RE: .18

	The "problem" is that we are not creating a market for the Alpha
	and it's capabilities. Customers would catch up if they thought
	they needed too. Create a market, make them think they NEED 64
	bits ("Gee Stan, why aren't you at 64bits like the rest of us?"), 
	and maybe, just maybe, we can make some cash this year.


	Digitals Stealth Marketing at work. (I don't totally blame the
	marketing orgs, the SLT is to blame here equally)

								mike

    Creating the market:  I've seen TV ads for what appears to be
    a new game system -- the Jaguar.  Their slogan:  
    
    		"64 bits -- Do the math."
    
    Do game systems need 64 bits?  Unlikely.  However, it's entirely
    possible that a market could be created that is initially based
    on perception; only later will the real need for 64 bits become
    apparent, and by that time the players will already be in position.
    
    

Will the market ever want the AXP??  Kind of doubt it.  When I talk to
people that left digital via TFSO or voluntarily and they purchased a 
PC or PC printer - guess what brand they didn't purchase?  Digital.

It's amazing that individuals that have been at DEC for 5-15+ 
years when purchasing equipment for personal use do not purchase
DEC.  If you can't get the employees to use/purchase the very stuff
they're familiar with I doubt that you have much of a chance with
someone external.

I think DEC lost the window of opportunity.  There was a guy in Japan
awhile back that was working with Nintendo to use the AXP in their
next generation of products.  When the interest got too high the
management of DEC backed of.  The individual has since left DEC but
nintendo is now using MIPS I believe.  Ford has chosen the PowerPC.
DEC doesn't have anyone.  We're destined to be another Sun Microsystem
in size because we lack a strategy. 10B today 5B tomorrow.

re Note 3022.21 by TENNIS::KAM:

> It's amazing that individuals that have been at DEC for 5-15+ 
> years when purchasing equipment for personal use do not purchase
> DEC.  If you can't get the employees to use/purchase the very stuff
> they're familiar with I doubt that you have much of a chance with
> someone external.
  
        People who have been with DEC for 5-15+ years have been using
        timeshared VAX systems and perhaps VAX or DECstation
        workstations.  A few have been using PCs we bought out from
        other vendors.  During that time frame we have convinced
        ourselves over and over again that Digital can't make it in
        PCs.  Very few have been using our current crop of personal
        systems.

        So I am not surprised that they don't turn to DEC for PCs.

        Bob

    RE: last couple 
    
    I've been with Digital for 16+ years, and I still have my own Rainbow
    at home. At the time that I bought it through EPP, it was a good deal.
    IMO, when EPP prices for current PC's become more reasonable, I might
    consider buying one. In the past, whenever my son had a homework
    assignment that required the use of a PC, I would drive him to the
    local library so that he could use 1 of a handful of Apple/Macs
    available.
    
    Recent advertisements for other company PC's have been tempting, but I
    have a hard time spending (less) money for someone elses.
    
    kjg

    since about 1976-80 I've used a modem and had access to all the
    resources you'd ever want or need so why buy a PC???

It isn't the employee's fault they don't buy Digital, it's Digital's fault.

The PC product prices are too high (even with employee discount) and the
availability (ship date) of current products doesn't compare to what *anyone*
can get from external sources.

And you are right, if Digital employees won't/don't buy Digital, why should
our customers?

	Joe

04/27/94--Computer Industry News - FOR INTERNAL USE ONLY
(c) Dow Jones News Service
              
INTEL IS EMBARKING ON BIG CAMPAIGN TO BUOY PENTIUM ---
CHIP MAKER PLANS $150 MILLION MARKETING EFFORT AS ANALYSTS CITE WEAK SALES

             With initial sales of Intel Corp.'s Pentium microprocessor line
          disappointing some market watchers, the chip giant plans to boost
          the product with the most ambitious marketing campaign in the
          company's history.
             The high-profile Pentium processor is the linchpin of Intel's
          strategy to maintain its extraordinarily profitable control over the
          chips that power personal computers. Intel's plan is to spend $150
          million on marketing to move a major chunk of the PC market to the
          proprietary Pentium technology by the end of the year. It has set a
          target of selling six to seven million Pentium chips this year.

    Why do you draw the conclusion that it will sell simply because they
    have announced intent to spend $150M on advertising? It hasn't really
    sold yet and how much has already been spent on marketing?
    
    Maybe we need a new benchmark.
    

    Hide in a corner and watch... I'd bet that Intel is planning a _very_
    memorable advertising campaign. I'd also bet that a year from now
    we'll be lamenting what a huge success the Pentium is (especially as
    compared to the awesome Alpha that only Digital employees are aware
    of). :-(
    
    Harry
    

    I believe the only reasons Pentiums aren't selling well now is that the
    prices keep dropping, and consumers are waiting to get a better buy. 
    Hence, there will be a huge pent up demand for Pentium (pun intended). 
    I know I plan to buy one, but I believe that I'll be able to save about
    $500 if I wait another six months.

�    I believe the only reasons Pentiums aren't selling well now is that the
�    prices keep dropping, and consumers are waiting to get a better buy. 


could it be that the recent 486 chips that were sold in systems in the retail
market aren't ready to upgrade yet?  Let's face it, does the average consumer
upgrade their systems as often as business?

Michael

    Intel originally said 1 million Pentiums in 1994.  Then they raised
    it to 6 million.  "Poor" sales is relative to their new number,
    not their old.  We've sold @20,000 Alpha AXP's, by comparison.
    When Alpha was announced, we said we would sell 10 Alphas for
    every Pentium.  About a year ago, we said 1:1.  We're getting
    closer to sensing reality, I guess.  kb

re: .31

I don't claim to know what the actual number is,
but we've sold many more Alphas than 20K.
Please be careful with your numbers.

    Well, a recent memo broke down Alpha systems shipments and was 20k.
    The Boston Globe also had the 20k figure not long ago (probably
    from the same memo ;-) and said breakeven was 90k.
    
    Why are you so sure "we've sold many more"?  kb

Relax. Just a thought for some marketing wizard to score off a coup of the
decade.

SunSoft, based on its marketing savvy, has thought it fit to port Solaris on
PowerPC. Technically speaking, someone can get hold of PowerPC-based hardware
to run Solaris on it and be in the market. I mean, that is how you deal with
Acers of the world.

Now Solaris is a creditable OS from a worthy competitor. I fail to
understand what is still missing in Alpha such that SunSoft has still not
considered porting Solaris on it. Obviously they know that Alpha has found a
place in Guiness Book of world records and all other technical stuff. 

Now back to Sun Microsystems. SPARC seems to be running out of steam. Sun was
heard of developing some chip alongwith TI but obviously  there are a lot of
problems. Modern chips are a very high-cost game. Digital (still a $14B company
vs Sun $4B) has already spent hundreds of million on Alpha.

The whole point I am trying to make is that there must be something missing -
technical or non-technical, that is coming in the way of Alpha being widely
accepted by the third-party computer vendors. I have mentioned Sun just to
represent some big fish in the computers business.

On the marketing front, I have seen so many articles on CPUs where Alpha gets
relegated to a somewhat obscure place - Fortune giving Pentium on the cover,
Byte running a series on PowerPC etc.

Now response to some of the replies -

Reply .4:

>>    Too many people don't understand PowerPC. The "real" one doesn't exist
>>    yet nor does an operating system, compilers, or applications. 

The initial versions of PowerPC seem to be selling well. What will happen when
the real-ones, whatever they maybe, appear in the market ?

Reply .6:

>>    	- The slowest Alpha is faster than the fastest PowerPC, and
>>    	  we're not stopping at that

Yep. That is all the more reason to wonder why computer vendors still are not
adopting Alpha. Do not get me wrong. We all wish Alpha to succeede but I am
wondering why SunSoft does not want Solaris to run the fastest on Alpha ?
    
Regards,
Gupta

    >The initial versions of PowerPC seem to be selling well. What will
    >happen when the real-ones, whatever they maybe, appear in the market ?
    
    What information is there to substantiate this? Are we referring to the
    PowerMac? Imbedded systems? PowerPC (as a chip) appears to be getting a
    lot of press but that is no indication of sales of systems.

    Megahertz isn't everything.  Intel's Pentium runs DOS applications
    faster than Alpha.

    re: .35  and substantiating information ...
    
    The proof of the pudding will be in the next SIA quarterly report and
    the Microprocessor report. 
    
    Geoff

    re: .37
    
    I'm not trying to create a rathole on this really. What I'm trying to
    do is put realism to the misbelief that PowerPC *systems* (not chips)
    have taken over the world.
    
    The reason I happen to be so interested in PowerPC these days is that
    IBM just won a major contract with my customer where the PowerPC
    (RS/6000 Mod 250) was bid and there's an issue whether there is any
    software that runs natively on the PowerPC (vs. emulated). Further,
    there are questions that deal with whether PowerPC is a system or
    simply a chip within another system.

    I'm on the front lines..the bottom line is we'd sell five times more
    AXPs if:
    
    1.  NT had 386/486 emualtion
    2.  NT (3.5/Daytona) was optimized for RISC multiprocessors (namely Alpha)
        (READ: greater that 2)
    
    I have to tell customers every day that 386/486 emulation doesn't exist
    on the AXP.
    
    

    RE: .-1
    
    1. The beta versions of NT for AXP did not include DOS emulation, but
       final release most definitely will run DOS and WIN 16 applications.
    
    2. Why should Microsoft optimize the NT kernel for 2% of the market?

    >1. The beta versions of NT for AXP did not include DOS emulation, but
    >   final release most definitely will run DOS and WIN 16 applications.
    
    You're right, NT will run DOS and Win16 apps, but _only_ if those
    applications do not use 386/486/Pentium instructions.  Some of the
    newer Win16 applications which are on the market use these instructions
    and _will_not_ run on Windows NT for Alpha AXP.
    

re .39

> [if]  1.  NT had 386/486 emulation

and .40

> 1.  The beta versions of NT for AXP did not include DOS emulation,
>      but final release most definitely will run DOS and WIN 16         
>      applications.

Unfortunately, these are not the same thing.

NT 3.1 on Alpha will run DOS and WIN 16 applications iff 
those applications are 286-level applications.  You can't
run apps which need 386/486 capabilities.  I'm not enough
of an Intel architecture whiz to describe in more detail what
that really means, but I do know that quite a few Windows
apps cannot be run on Alpha, precisely because of the current
lack of 386/486 emulation.

    re:.40
    I'm definitely NOT an Intel wizz, but I do know a few things about the
    386 vs. the 286.
    
    The 286 was a 16-bit embedded controller chip that got turned into a
    computer by default.  The 386 was a 32-bit computer chip with backward
    compatibility to the 286.  Meaning?
    
    The 286 uses segmented-mode (8086) addresses, or a different
    "protected" mode of its own with some serious flaws.  The 386 adds a
    true 32-bit address mode, including a 48-bit (32:32 segmented) mode.
    This lets programmers leave segments for the worms.  The 286 has 16MB
    (24bits) of address space, the 386 goes to 48 bits.
    
    The 386 handles interrupts better, so it multitasks cleanly, while the
    286 has, I think one set of vectors shared among all processes.
    
    MS-Windows runs in only Standard mode on the 286, but in Enhanced mode
    on the 386, and many new applications run only Enhanced.
    
    The 386 has additional instructions, etc....

re Note 3022.42 by TRURL::porter:

> NT 3.1 on Alpha will run DOS and WIN 16 applications iff 
> those applications are 286-level applications.  You can't
> run apps which need 386/486 capabilities.  

        To illustrate this another way:  NT 3.1 on Alpha emulates a
        VAXmate or an IBM AT.  Remember how long ago those were
        obsolete?  Software developers know that people who are
        spending real money on software today have long ago moved to
        the newer processors.

        Bob

    A couple of years ago the local community college I attend installed a
    whole room full of DECstation 5000's.  The room rarely has students in
    it.  The ones that do go in there are running a PC emulation.  Most
    just LOVE the VI editor. (sarcasm intended)
    
    There is another room, now, of 486 based machines with 2 AXP150
    servers running NT.  The printers are HP.  The network, which runs
    flawlessly, is a DEChub.  There is a lot of DEC hardware there.  There
    is no DEC software.  There are no DEC tools to build software on the NT
    platform.  We have this OSF/1 operating system but the user interface
    is totally obsolete.
    
    Having a chip that runs fast is not enough to make it sell.  People in
    the commodity consumption end of the business (the users) are not
    buying the chip, they are buying utility.  People writing software for
    those commodity machines need compilers and other tools.  We don't have
    any.  Until we do, the Alpha will not sell.  
    
    Microsoft's compilers support Intel based PC's.  Borland's compilers
    support Intel instruction sets.  Their code, should it actually run on
    the ALPHA/NT, will certainly not be optimized for the Alpha instruction
    set.  I think we need to supply competitive tools at competitive
    prices that make our product look as good as it really is.

    I have some data from the south Queenferry AXP plant that favourably
    compares the current Alpha chips. We have more power and we will be
    able to harness this for standard DOS/Windows applications even better
    when the new version of SoftPC for Alpha is loaded. 
    
    On native applications for 32bit operating systems we are faster than
    PowerPC's today. On 64bit OSF we run rings around it. We have already
    demoed a 375mhz chip based Alpha PC at Cebit this year. PowerPC is
    almost 2 years away from that type of uumph!
    
    Lets not forget, the current risc chips are really only good today for servers
    or power users. We will do most of our business on Netware servers and
    unix workstations. Apple needs to make this current chip work or else
    they are history. We already have it working on 4 operating systems and
    just need to market a proven technology.
    
    I should know.... I worked for Apple for 3 years!!
    
    Best regards
    
    Damien Peat @ DBO

    I call attention to PC Week of May 2, 1994, page 75.
    
    "IMS' 3250 RISC chip emulates both PC, Macintosh CPUs"
    
    This article is about a new chip -- VAPORWARE that may ship i a few
    months -- that has microcode to emulate both the 68040 and i486, and
    runs about as fast as a real 68040 or i486.  It also has a native RISC
    mode. IMS is also trying to market their on-chip emulation to other
    RISC vendors.
    
    If it works (and the writer, Bill Machrone, is taking a wait-and-see
    attitude), then software emulation will become obsolete.  And IMS could
    become a major foundryless chip house.

05/04/94--Computer Industry News - FOR INTERNAL USE ONLY
(c) Dow Jones News Service


MOTOROLA SETS CHIP PRICES

             SCHAUMBERG, Ill. -- Motorola Inc. said it plans to make its
          high-performance PowerPC 603 microprocessors available in quantities
          of 20,000 at $160 each for the 66 megahertz version and $199 for the
          80 megahertz chip.
             The announcement was the first pricing for the PowerPC 603 chip,
          which is designed to operate portable personal computers. Motorola
          said it expects to begin volume production in the third quarter.

    Speaking of what we're up against price wise...
    
    WSJ, 5/2/94:
    "Intel Gives Customers Word of Price Cuts for Its Chips"
    SAN FRANCISCO- Intel Corp is preparing customers for yet another
    round of microprocessor price cuts in the second half of the year,
    stepping up its efforts to move the market to its Pentium processor.
    The entry level Pentium chip, for example, will fall from $370 in
    the fourth quarter from $675, said one customer who had been briefed
    by Intel. ...
    Intel faces stepped up competition from Advanced Micro Devices Inc
    and a partnetship between Cyrix Corp and International Business
    Machines Corp in the market for Intel's earlier 486 chip.  For that
    reason, Intel is hoping to use lower pricing to make Pentium chips
    the new mainstream engine for desktop personal computers.
    PC Week magazine yesterday reported that the price for a PC using the
    60 megahertz pentium would drop to about $2000 in the fourth quarter
    from $2600 at the moment.  It said the newer 90 megahertz Pentium
    will drop to $625 from $830 over the same period, while PCs that use
    the chip will drop to $2500 from $3000."
    
    copied without permision; you have my permission to forward this to
    anyone that's awake in SCO.

    	"Information Week" in its May 9, 1994 edition reports that IBM,
    albeit reluctantly, will offer WNT as an option on PowerPC machines
    later this year.  It is apparently taking this step beacause of
    customer demand and the fact that its own new OS software will be late.
    The article makes absolutely no mention of the Alpha chip despite
    the early and loud announcements of Alpha support for WNT.  If
    the PowerPC/WNT platform is able to hold its own against WNT running
    on Alpha (I'm talking sales, not performance}, then we have really
    blown it, in my opinion.

    Probably a dumb question; but...
    
    If IBM/Motorola/Apple are teamed up on the PowerPC, is anything
    stopping Digital/Intel from teaming up on a similar product?
    
    I don't doubt that my question is overly simplistic, but I thought I'd
    ask...if it displays me as an idot, please be gentle in your reply ;-)
    
    
    Steve

    
    Prolly Intel, whats in it for them to compete with themselves??

Reply .50:

>>    	"Information Week" in its May 9, 1994 edition reports that IBM,
>>    albeit reluctantly, will offer WNT as an option on PowerPC machines

>>    The article makes absolutely no mention of the Alpha chip despite
>>    the early and loud announcements of Alpha support for WNT.  If
>>    the PowerPC/WNT platform is able to hold its own against WNT running
>>    on Alpha (I'm talking sales, not performance}, then we have really
>>    blown it, in my opinion.

Why these people from the press are more enthusiastic about what IBM plans to
do 1 year down the line and do not bother to even mention that Digital
accomplished that 2 years ago ?
 
Can somone through some light on this strange psychology of these press people ?

Reply .46: 

>>    PowerPC's today. On 64bit OSF we run rings around it. We have already
>>    demoed a 375mhz chip based Alpha PC at Cebit this year. PowerPC is
>>    almost 2 years away from that type of uumph!

Why even such performance figures, which are bound to improve with time, do not
seem to impress enough that other computer vendors embrace Alpha ?

>>    they are history. We already have it working on 4 operating systems and
>>    just need to market a proven technology.
  
Maybe someone from Alpha Marketing can comment on this.

Regards,
Gupta  

    Re .-1
    
    There is a mention of Alpha in this months Acorn User. It was in a
    review of Acorn's new RISC PC. This has a slot for a second processor,
    and one of the possibilities mentioned an Alpha running WNT.
    
    It also says that Apple are having to change their ads since they were
    implying that PowerPC was the market leader for RISC based PCs, when
    currently more ARM based systems have been sold.
    
    Peter

    
    	Oh hooray!  Alpha being used as a co-processor on an Acorn eh?
    
    	Why are we not promoting the Alpha PC with all we have?  Since it
    will beat the pants off the competion, is it going to be another of
    DIGITAL's best kept secrets?.
    
    				Malcolm.

    The June issue of IEEE Computer has an article comparing  the
    technical aspects of the PowerPC 601 and the Alpha (AXP) 21064.
    
    	FYI
    
    		Yaacov

    And, and, and...?  How'd we do????
    

    >                          <<< Note 3022.58 by DRDAN::KALIKOW "World-Wide Web: Postmodem Culture" >>>
    > 
    > And, and, and...?  How'd we do????
    
    More than anything it is a TECHNICAL comparison of the architectures.
    They did however run SPEC benchmarks against the chips. As opposed to
    "marketing" articles which tend to have biased comparisons, this
    article has comparisons between (Chip/Mhz)
    21064/150,21064/200,21066/166,21064A/275(est) as opposed to 
    601/66,601+/100 and 604/100. They also have the Availability data for
    each chip so that you can see what "we have now" and what "they have
    later".
    As I said above, this part is not the emphasis of their article but
    the closing paragraph in this section is worth noting:
    "The table shows that the 21064 appears to be ahead in the performance
    race, as measured by the SPEC benchmarks. It also shows that while the
    PowerPC has a slower clock period, the performance results are closer
    than the clock period alone would suggest. This is consistent with the
    'more work per clock period' philosophy used in the PowerPC designs"
    
    An interesting observation is made in the Article Summary:
    "The PowerPC gains performance by design cleverness; the Alpha gains
    performance by design simplicty, This trade-off is a classic one, and
    the fact that both philosophies lead to viable processors, is proabably
    an indication that either choice is satisfactory as long as the
    implementation is done well".
    My personal extrapolation from this is that scaling (up in clock &
    processor speed, down in size and heat) should be less limited using a
    simple design than using a "clever" design.
    
    Enough said, this is getting a tad too technical and proabably
    shouldn't be in this notesfile.