Conference 7.286::digital

    Amongst other things, Samsung is one of the worlds largest
    manufacturers of memory components. They also produce process
    controllers and other "embedded" processor systems - which one might
    surmise could be one use for Alphas...

    Samsung also owns a lot of AST (41%?), which will presumably be
    using Alpha chips in servers (or maybe not, once they see how
    fast they are like COMPAQ discovered...).  Samsung is also rumored
    to be building a "low cost" Alpha variant for the NT PC market;
    we keep on building the cutting edge EV6, etc stuff.  Kratz

    Don't forget consumer electronics.  My wife just bought a Samsung
    camera.  They probably do the optics too.
    
    Carl
    

    Samsung is an $80B company. Samsung Electronic is a $20B company that
    is into just about everything. The semiconductor part is the world
    leader in Memory products(by volume). Samsung make computers of all
    sizes under the Samsung name and also own 41% of AST. 
    
    -Bruce

    With Alpha, you can safely forget consumer electronics (too large,
    too expensive, too power hungry, not designed for that market).
    StrongARM is another story, but not via Samsung.  

    good article in 24 June 1996 InformationWeek pg. 34 Samsung Added to
    Alpha Lineup.  Can't find the article on-line to post it.  Worth
    reading.  It indicates that Digital's 64-bit RISC will go into Samsung
    PCs e.g., AST.
    
    This was the most interesting paragraph:
    ... Meanwhile, Shannon (editor of Shannon Knows Digital, a newletter in
    Ashland, Mass.) says, Compaq Computer is evaluating RISC platforms and
    is "again looking at Alpha" after talks with Digital broke off last
    year. ...
    
    I'd sure like to know why Digital and Compaq couldn't come to terms.  
    We let Apple slip away we can't let Compaq.  I'd rather go
    down in a sinking ship with Compaq than alone.  Compaq knows how to
    patch the holes.  Digital just lays offs employees to lighten the load 
    and ask everyone to row harder to land - which is no where in sight.
    
    Last paragraph:
    Sources say Digital's goal is to offer Windows NT-based Alpha desktop
    systems by next year priced around $2,500 to $3,000.  Alpha systems
    today offer performance comparable to Intel Pentium systems start at
    about $4,000.
    
    What's going to allow us to get to 2-3K; why can't we do it today?

    Man oh man.
    How much longer can we expect customers to run after the dangling
    carrot.
    
    Can I believe that Samsung could bring the price of production down?
    yes.
    
    However - a year out is a year out - is another year for applications
    to be build for the Pentium instead of the Alpha.
    
    And - who is to say that the Pentium prices won't be coming down
    in a year.  Based on what we've seen in the industry to date - I think
    that it is a safe bet that the Pentium will always be less expensive
    than the Alpha - if for no other reason than they've got the volume.
    
    It's been stated in here many times before - if Digital wants to sell
    the Alpha - better start reducing prices now - and we better give them
    a reason to prefer it to Pentium.  64 bit is a nice story - but so far
    not really needed.  My bet is that the competition will have it by the
    time it is.

    by not really needed - I mean - in significant volume.
    
    We've also been wondering what Samsung could be doing with
    the chip.  We've been thinking - a really fast TV remote control.
    Man could I flip through the channels with one of those babies.
    Though with that battery back strapped to my back, it might make it
    a little difficult to take full advantage of the couch.

    Bill Gates was on a TV program a couple of weeks ago and he indicated
    that Microsoft is hoping that PCs get to 500-1000 in the next couple of
    years.  Therefore, if PCs are less that $1000 and Alpha is 2-3K we're
    still too expensive.
    
    Why don't we just take the Alpha/Celebris XL and lower the price to
    buy-the-business.  If not, shortly we'll be out-of-business and wishing 
    we had.  Just like we miss the boat with selling MicroVAX chips.
    
    
    	Regards,
    

 >   Why don't we just take the Alpha/Celebris XL and lower the price to
 >   buy-the-business.  If not, shortly we'll be out-of-business and wishing 
 >   we had.  Just like we miss the boat with selling MicroVAX chips.

Because we can't.  We've got maybe 5% of Intel's manufacturing capacity and 
much higher fixed costs.

Bruce

Re: Note 4681.7 by TENNIS::KAM

    
>    I'd sure like to know why Digital and Compaq couldn't come to terms.  

I think a few things, one is that after looking into bringing RISC based
products to market they decided that the return would not be worth the 
expense.  Another is Intel might have asked them, nicely, to stay with
Pentium (& Pro).
    

>    What's going to allow us to get to 2-3K; why can't we do it today?

As mentioned in a few places, Digital and Mistubishi are developing a low
cost EV56 chip.  This chip should allow us to build Alpha systems that out
perform Intel systems at a similar price. 

-Bruce  

> I'd rather go down in a sinking ship with Compaq than alone.
> Compaq knows how to patch the holes.  Digital just lays offs
> employees to lighten the load and ask everyone to row harder
> to land - which is no where in sight.

  The best part is that one day, land is "North". The next
  day, land is "East". The next day, "South". Then "West".

  This may be due to the fact that we have 200+ navigators,
  or it may be due to the fact that we can't agree on which
  cointinent we'd like to reach.

                                   Atlant

re .13

> As mentioned in a few places, Digital and Mistubishi are developing a low
> cost EV56 chip.  This chip should allow us to build Alpha systems that out
> perform Intel systems at a similar price.

Great - more futures. As I remember the last LCA chips were less than 
stellar performers. Still waiting to get that part number for a system with 
an "expensive" EV56 chip that can outperform 200 MHz Pentium Pro server
systems - you know, the ones that are readily available TODAY. 

tim m.

    RE: .14
    
    Gotta agree with you.
    
    I can go back and count the notes over the past years that say:
    
    "Just wait until we get the $?K low-end machine using the EV?? chip!
    We'll blow their doors off!"
    
    Well, either those machines never saw the light of day or were way too
    late to be meaningful...
    
    				-John

    I also agree with you, the reason I mentioned the to be not yet ready
    for prime time chip was to answer the question from .7 which was
    "What's going to allow us to get to 2-3K; why can't we do it today?".

    We need to somehow get more aggressive with our time to market with new
    systems.  I have no idea what it takes to change our systems from EV5
    to EV56, but the SBU has known about EV56 for over a year and has had
    samples for five or six months. The last FRS date I saw for the 440Mhz 
    EV56 based TurboLaser was August.

    Alpha exist to allow us to sell systems that lead in performance. We
    have a lot of competition, but are in a position to keep the
    performance lead. To do this, we need to get better at bringing our
    products to market in less time. This includes both Chips and System.

    -Bruce

    	I hope to ship the first AlphaStation500/400MHz EV56 machines
    	today or tomorrow.
    
    				R.Dischler
    				AS500 Mgr

    Have you published a Sales Update article?  I don't see how you
    can ship them if nobody knows the part numbers.
    
    				-John

Here are some part numbers:

PB56A-AA - UNIX SBB
PB56A-BA - VMS SBB
PB56A-CA - NT SBB

PB561-AA - 2d Unix
PB561-BA - 2d VMS

PB561-CA - 3d NT


-Chris

Does this mean we're one-upping IBM's AS/400??

Who wants two-onethousandths of a computer?

BobW  8-) (for the sarcasm-impaired)

>================================================================================
>Note 4681.17       What is Samsung really DOING with our chip?          17 of 17
>WRKSYS::DISCHLER "I don't wanna wait in vain"         5 lines  26-JUN-1996 11:36
>--------------------------------------------------------------------------------
>                     -< EV56 machine due this week to FRS >-
>
>        I hope to ship the first AlphaStation500/400MHz EV56 machines
>        today or tomorrow.
>    
>                                R.Dischler
>                                AS500 Mgr

    RE: .19
    
    Thank you!  That is the first time I've actually seen part numbers
    that I can look up in VTX PRICE.
    
    				-John

    >With Alpha, you can safely forget consumer electronics (too large,
    >too expensive, too power hungry, not designed for that market).
    
    It would be great in a toaster or a crockpot, though - with an Alpha
    embedded controller you don't need a heating element.  And, in a Mr.
    Coffee, it would provide that fastest Java brewer while providing the
    fastest execution of Java (assuming the coffee maker is plugged into
    the internet).  
    
    Of course, we would have to publish the refereed tpC benchmark results
    (transactions per cup, of course!).

    ... but since there are shops where they measure Engineering
    Productivity by the KLoC/D/MgCQ� ratio, in today's rapid-prototyping
    world where much code is strongly interpreted, the presence of an Alpha
    in the JavaCODE brewer could strongly potentiate the numinator.  
    
    Not to mention simply brewing the JavaLIQUID more efficiently, thus
    potentiating the denominator!!
    
    Heck, if I could work on such a brewer and that ratio strictly
    determined salary, I might be economically forced to learn Java,
    myself!!  :-)
    
    
                                                                  
    -----
    � Thousands of Lines of Code per Day per milligram of Caffeine Quaffed

    Drdan:
    
    I heard u installed an Alpha chip in your Norelco...just make sure u
    don't nip your nose and by all means, watch out when doing those
    sideburns!
    
    Ron

RE: LCA performance

Note we sold 100Mhz EV4 Pelican systems that had *less* performance than 
the 166Mhz LCA4 processor, later boosted to 233Mhz.  IMHO, the problem with 
the LCA chip was due to political fighting between the workstation and the 
terminal group over Multia.

RE: Compaq and Alpha

There was a recent article in PCweek or InfoWorld where Compaq was said to 
have officially cancelled the team evaluating Alpha and other RISC 
processor.  The reason given was too much overall program costs.

    Good point re 100MHz Alpha desktop systems. Performance isn't
    everything for everybody.
    
    In the UK, Multia was sold exclusively via X-terminal distributors, but
    its real addressable market was the corporate IT folks who realised
    that standard PC = IT anarchy = serious cost of ownership disaster.
    (The Multia Enterprise Client name should have been a hint there). What
    do X-terminal distributors know about PC cost of ownership ? And
    although the ABU folks may have understood the PC cost of ownership
    disaster, they weren't motivated (=rewarded) for selling Multia.
    
    A plan to fail, whether it had Intel Inside or Alpha Instead ?
    
    regards
    john

    I heard via a relative who works for Motorola in Austin, Texas that
    Samsung is building a huge semiconductor fab in Texas.  What does that
    mean?
    

That Mr. Palmer needs a new vacation spot?

    Good article in SunExpert Magazine June 1996 pg. 46, The current crop 
    of systems - We evaluated 64-bit systems from all the major vendors.
    
    What's compared:
     MIPS Technologies Inc. R10000
     Sun Microsystems Inc. UltraSPARC-I/II
     HAL Computer Systems Inc. SPARC64
     Digital Equipment Corp. Alpha AXP
     Apple Computer/IBM/Motorola PowerPC 620
     Hewlett-Packard Co. PA8000
    
    Pretty favorable comparison amongst all vendors.  Also, another article
    on pg. 60, Inside the RISC Processor - An overview of CPU design
    choices and their implications.
    
    In HP Professional June 1996 there's a bunch of information on the new
    PA8000 system what are/will be available and the upgrades from the
    current systems to the PA8000.  The C-, K-, and J-class systems are
    upgradeable to PA8000.  The C-class, C110, is about $7,000.  According
    to the article there is already an overwhelming demand for the
    upgrades.  Once HP has this we won't know what hit us.