Article 150988 of comp.os.vms:
jfmezei@istar.ca writes:
>
>rob@polaris.east.dialog.com wrote:
>
>> A rhetorical question that begs to be answered. I believe Digital
>> will gain market share in a big way. Wonder if Compaq would
>> OEM Digital PCs in exchange for slapping together a few Alpha
>> PCs to lend a tier-one precense :-).
>
>Digital lacks the chromosome necessary to sell and market stuff. Digital came
>out with a killer chip a few years ago and couldn't decide on a name for years.
>The Alpha is just another fast chip today that is not popular for desktops. The
>other guys will catch up with Alpha and the slight edge Alpha may keep will
>not be worth the problems of a different architecture, even with FX!32 for J Q
>Public's desktop.
Ah, but the Alpha isn't just another fast chip. Must be at least
2 times as fast as the competition to take away market share. This
time next spring (April Comdex by chance) Digital will be shipping
the 21264 in volume. Rumor here, rumor there has this CPU at 30+
SpecInt95 and 50+ SpecFp95. With numbers like that Digital will
finally be able to turn heads. I say finally because if you go
back in history (November 1992) to the original Alpha 21064 rollout
you see something very interesting take place. Sun (arguably a
much better marketing effort from them at times) pre-empted the
Alpha rollout by talking up the SuperSparc. Now as we know, the
SuperSparc in many ways in various forms couldn't hold a candle to
the Alpha but the Alpha (21064) wasn't over 2 times as fast. Digital's
rollout of the 21064 had the same minimal effect on the rest. SGI
with MIPS was in no danger, IBM RS/6000 was a solid performer, etc.
Fast forward to April 1995. The TurboLaser rollout (21164) was
a great success. It was different (VLM) and is capturing server
share. The workstations are a different story, they came later
in the summer 1995, by that time Sun only had to talk a UltraSparc
song and dance for a few short months until it shipped in volume
about 5 months later. MIPS with R10000 is in no danger, HP has
a rocket in the PA8000 so 21164 isn't ripping workstation share.
(Maybe from HP, since the PA 8000 is hung up for at least another
month so their next hot workstation really isn't there yet).
Things are suddenly much different, at 500 MHz (a 70% performance
improvement in a year) the 21164 is much more robust and has
pulled ahead of the pack (on integer anyhow). A low cost Alpha
will be showing up at year-end that is a rocket. The PA 8000 can't
compete in price, nor can the R10000. The P6 (Pentium Pro) though
lagging is still price king but is teetering on that magical 2 times
barrier. The 21164 @ 500 MHz will do 15.4 int95 and 21.1 fp95. A
fantastic number cruncher compared to a Pro, but alas it seems the
21164 running NT is relegated to the engineers and scientists but
that is okay. A very nice slice if Digital pulls it off.
Here is where it gets interesting and Samsung must think so too . . .
The 21264 will have at least a 9 month lead on everyone's next
generation introduction. Sun will not be able to do a 12 month
Ultra III song and dance. Intel can't do a Merced dance for 9 months.
MIPS *still* isn't shipping that vaunted 275 MHz R10000. The PA 8000
shrink will pull it ahead of the 21164 but less than 1/2 a 21264 in
performance, they too are waiting on the P7. Digital will have the
VAST majority of industry focus in 1997. That may be the last great
window of performance **dominance** (not just 40% faster) Digital
sees this century.
Here is my idea. Mr. Palmer, you need to gather your brightest and
go on a 10 day retreat, bring along your inside and OUTSIDE marketing
folks and say:
"Look, we are going to spend $20 million, $40 million of
that 1 billion in cash, WHATEVER IT TAKES (right!!). The board
may sack my butt but I don't give a rats lunch. We are going
to let the world know we have a CPU and boxes to go with it
that are 3 times faster than any other currently shipping
platform. I want TV commercials, I want Final Four,
I want Wall Street Journal, I want TNN, the Disney Channel,
Discover. And folks we are going to do something unique
in the Industry. We are going to give away computers to
our harshest critics. That's right. I want to give away
at least $1 million dollars worth of fully loaded Alpha PCs.
Alta Vista is a great idea. This will more than pay for
itself, we will turn our critics into better marketing than
you can buy. We will call it the Million Dollar Giveaway.
Hey, you know I was flicking around the other night
and saw Metallica do a free rock concert for a guy that
won. I want to do a spin on that, we load up the Digital
trucks and go out across the country. I will pop out of
one where we do a big-big picnic for the small town and
sing at the local saloon. Ted should pick it up on
CNN, maybe we can get Jane to pop out of one, it will be real
news, heck maybe Iron Butterfly is availble to sing with us.
Hear they got a killer song called Inagadaltavista, that will
be our theme song.
Any objections???"
"But Bob"
"Yes, Dean"
"Bob, you can't just give away computers"
"Why not Dean?"
"Uh. . . I don't know, never mind."
"Thanks for the input Dean, go grill some burgers Dean and
get me another beer while you're up and when you get back
drop and give me a hundred."
>
> Who knows. Imagine there
>> are more than a few shouting matches going on inside Digital
>> right now.
>
>Look at Apple. How come it did not choose ALPHA instead of the PowerPC chip ?
Because they made a big mistake. The PowerPC is trailing the pack
and won't hold a candle to the Alpha this century. The 21264 will
be at least 3 times faster than the 620 which is due in a server
in 1997. 620 isn't a desktop CPU. The 21264 is that and many
other things. PowerPC camp is in a very awkward position. Little
did you know but shortly after you posted this, Motorla's stock dropped
like a rock. Part of that is attributable to Apple's demise and lack
of pushing product out the door. Look at the PowerPC in the RS/6000
world, very little of it since it really isn't there yet. Check out
the closing comment on the back page in the lower-right hand corner of
the latest ComputerWorld. IBMs Web Server talks about a 604 upgrade
for RS/6000 servers appearing in 1995. It is mid-1996 and still isn't
shipping. That is rather painfully laughable.
>Does anyone know if Digital even tried to sell ALPHA to APPLE ?
Who cares? So they can have their hopes dashed? Poor PowerPC
camp. Apple is in the midst of a terrible death spiral. A very
heavy whisper could blow Apple away right now. Why death spiral?
Their best and brightest are bailing out. Microsoft picking 'em up
right and left. According to "The Dilbert Principle", by Scott Adams,
your best and brightest are always the first to go in a dying
organization. Duh. They be the ones that can figure out the
organization is dying.
>
>>
>> Why do desktop PCs? One stop shopping? So? Put your own
>> label on somebody elses PC and let them have the headaches. Add
>> value ala StorageWorks. Must have been painful for Digital to
>> cut their hard drive manufacturing but it was a very wise decision.
>
>I was under the impression that DEC's manufacturing plants for PCs were very
>efficient. If that is the case, then DEC would be better off building PCs for
>others and let the others put their labels and do their own marketing.
Sanyo entered the PC business. Time to get out. Look just what
happened today. HP closed their disk-drive business. That paragraph
above makes me look rather psychic. Don't tell anyone, but me and
Dionne Warwick are like 2 fingers on a hand. We're so close.. hold
on, I'm getting another reading... OOoohhoo . uuhhh >>> Digital
should take a big chance regarding PCs, get out of the desktop now. 2
years from now HP will ditch their PC business when margins drop to
$1.99 per machine. <<< Drop me a private e-mail and I will let you know
where to send the $300 dollars for that reading.
>
>>
>> I drifted past the main point. Suppose Digital could/will manufacture
>> a PC that significantly outperforms a Pentium Pro? The 21164PC will
>> be just that. But wait, it gets better. Before Intel ships their
>> first P7 (many months before) Digital will be shipping a 21264
>> that slices and dices any and all comers.
>
>For as long as DEC prices its ALPHAs higher than Pentiums,
>for as long as DEC remain incapable of shedding its "DEC IS EXPENSIVE" image,
>for as long as DEC's top guys are incapable of long term commitments, you cannot
>expect Digital to become popular.
When Bob and the gang come back from that week-end, it will be
as if they are new men/women/people. They will be so fired up that
legend will grow out of that meeting. It will take on an aura never
before seen inside Digital. Stories and rumors will fly that will
elevate everyone. Dean will quit in embarassment.
>
>Digital had a good but short lived advertising campaign. But results didn't
>come it quickly enough and the guy at the top fired 7000 along witha few other
>VPs because he said it would not work. I call that short term thinking.
Yeah, like you know all the ins and outs of running a multinational.
As IF Bob fired 7000 because he was having a bad hair day. Sometimes
you have to make very hard and very radical decisions. The recent
downsizing at other major corporations quickly fade but Digital's are
seared in your mind. Wait, Dionne is sending me more ...
"Apple will shed several divisions in an effort to stay afloat.
By early 1998, Apple will be akin to NeXt COmpUtEr."
Make that $600 bucks.
>
>
>> Won't Intel have to do their own version of FX!32? Can't imagine
>> them setting aside P6 real estate on a P7 just to maintain some
>
>Yeah, and guess what, it will probably be just as efficient as FX!32.
Sure with several years of translation experience for Digital
and none for Intel. Intel's will be just as good. Maybe. But
even if it is. Intel will have to sell everyone on recompiling.
At THAT point it will make more sense to do 21264 as it will
significantly outperform the P7 vapor at the end of 1997 (21264 at
600 MHz then).
Let's say it doesn't -- no mind. Digital and Samsung and partners
will have a good solid 9 months of uncontested desktop and server
dominance. Samsung knows that. Surely Samsung could have decided
to OEM PowerPC? What a waste that would be ;-). Where is PowerPC
going to gain market at? That is a very troubling question for AIM.
ABSOLUTELY no where to go. Drop by comp.sys.powerpc, the saddened
MACaholics post:
"When Exponential comes out with their big bad CPU then
things will happen."
Oh yeah, what will they put them in, MAC clones? Boy oh boy they
should really move off the shelf. PowerPC missed the NT boat, sorry
boys.
>
>
>> The REALLY big difference here is that Digital has an
>> Intel killer about to be revealed, surely Intel is not worried but
>> Digital has a much better chance than the PowerPC camp EVER did.
>
>And when ALPHA was originally unveiled, it was also supposed to kill all other
>chips. I have news for you: IT DIDN'T. AND IT WON'T.
Nope, never believed it. The early numbers didn't support it. Still
don't. Come back and haunt me if the 21264 numbers also don't kill
all comers, meaning this:
"Digital will be shipping a desktop box that is 2 times
as fast as any other shipping box at 1/2 the price OR
3 times as fast as any shipping box at the same price point.
All this based on a snap shot on April 15, 1997."
See you then.
>
>ALPHA will survive. DEC will survive. But it will not become a mass production
>item.
PC Week has an article mentioning Digital is gearing up their
semiconductor works. They may be dreaming. But someone is projecting
they will be pushing product out the door. Samsung seems to think
so too. 1 week after reaching agreement, Samsung announced plans for
a desktop Alpha PC. Exciting stuff. $2500 to $3000 machine that
does 15+ SpecInt95 and 21+ SpecFp95 by late in the year. That is
a very nice box for number crunchers (engineers, scientists and
mathematicians) and knocks the legs out of other RISC players.
Desktop dominance? No way. A VERY nice niche.
>
>Look at UNIX: does Digital have the big market share it used to have back in
>the days of the PDP ? NOP. UNIXes on other platforms such as SUN and HP have
>had much more success. Why then does eveyone expect that when NT becomes
>extremely popular that Digital will capture the market ?
Sun has had their day in the Sun but clouds are approaching and
Digital is going to rain on their parade. Let me see, Ultra II
at 200 MHz is 7.1 and 11.1 int and fp respectively. How does
that stack up to a $3000 21164PC at 500 MHz? It doesn't.
NT on Alpha PC at $3000, frightens all competitors. They can
do a big marketing dance but will not be able to deflect this time.
GUARANTEED. Application availability and price is FINALLY going
to do things up right, ask Samsung.
>
>Digital is not the only one to have sold its soul to Bill Gates. Guess what,
>the NT clustering MS will produce will combine much more than DEC's clustering
>technology. You'll note that DEC now (or soon) will offer clustering for NT
>with eventually, a tool to upgrade to the REAL MS clustering product (which
>will also incorporate some Tandem technologies and I beleive HP as well).
Yes, I know about their NT clustering strategy and Digital and HP
and many others are contributing to Microsoft's cluster product
for NT code-named Wolfpack. Digital has nothing to be hesitant
about. They know they offer the highest performing platform
for NT. Digital can only benefit from NT's success. Ask yourself
who suffers the most if NT succeeds greatly? Scotty "Do NT When
Hell Freezes" McNeally. Myopic fellow ain't he? Sun can't bid on
several government contracts that require NT. Smart move Scotty.
>
>Sorry to sound so much doom and gloom, but one should not have wild dreams
>about Digital regaining leadership positions. To do so would require money and
>long term commitments. VPs have been playing musical chairs for some time, DEC
>has lost a lot of employees and is in no shape to be agressive in an extremely
>competitive market where even healthy companies are having problems making a
>buck.
Digital has shed divisions and dead wood and has collaterally damaged
things that didn't need shedding. Some brooms sweep too clean.
But your whole "doom and gloom" attitude isn't surprising. Many
have bought into the idea that Digital can't do anything right. Did
you know that Digital has the number one UNIX server shipment rate?
Need to sign up for Inform magazine. Sure it is propaganda but I
find facts in Sun publications too. Facts can be surprisingly reliable
at times. Digital is pushing product through the channel, more than
ever. Let those that are very good at selling, sell. Let those that
are very good at making disk drives make disk drives. Let those
that are very good at making desktop PCs make desktop PCs. When
margins are killing you, find a new schtik.
Rob