Is Alpha going away?
Michael Richards
it supposed to be dying? That's what someone told me but I haven't really
been able to find any indication of this. Since you folks probably deal with
this platform on a day to day basis I'm wondering on what your opinions are
on the safety of investing in a server farm of Alpha machines? Is the Alpha
processor really dying or is it still quietly going strong?
-Michael
Terry C. Shannon
"Michael Richards" <mi...@interchange.ca> wrote in message
news:v51x7.40695$5h5.18...@news3.rdc2.on.home.com...
Compaq recently began shipping 1001MHz EV68 processors for GS-Series (and
soon ES-Series) systems.
A 1.2-1.4GHz spin of EV68 is under development.
The Alpha EV7 processor is up and running, as are GS-Series Marvel system
prototypes. Marvel should ship in late 2002 or early 2003.
The Alpha roadmap is good to go through EV79, which means that new Alpha
systems will ship through at least 2004.
If you rummage around on Compaq's HPTC site you'll find a presentation which
claims that there will be a smooth transition from Alpha to IPF in the
post-McKinley era. Compaq and Intel are collaborating on a joint parallelism
project to help make this so. It's difficult to predict the future, but the
US Government just inked an $857M deal that'll involve a heck of a lot of
AlphaServers over the next few years.
Hoff Hoffman
:I haven't been following what's going on with the Alpha platform, but isn't
Are you investing in Alpha itself, or are you investing in OpenVMS
and particularly in various applications running on OpenVMS, and
centrally in your processes and your data?
Compaq has announced that the development of the Alpha EV7 and EV79
microprocessor generations is underway and that Compaq expects to be
shipping systems and servers based on these microprocessors. Compaq
also announced a port of OpenVMS to servers based on the Intel Itanium
Processor Family, and that future generations of high-end platforms
subsequent to EV7 and EV79 will be based on Itanium Processor Family.
We expect to have mixed-architecture clusters of OpenVMS on Itanium
and OpenVMS on Alpha available -- much like what we had and what we
did when we last ported OpenVMS over to a newer platform.
As for the currently-shipping hardware, we've recently started shipping
systems based on EV68 microprocessors.
Various presentations and related materials are available at:
http://www.openvms.compaq.com/presentations/
http://www.compaq.com/hps/ipf-enterprise/
Mark Gorham's presentation is probably the best presentation for
information on EV7 and EV79 and the OpenVMS Itanium efforts, most
of the others have more technical details of what is going on in
OpenVMS and for the information has been released to date on the
version of OpenVMS that will be running on the Itanium Processor
Family platforms. (That said, the technical update presentation
does have a slide at the front of the presentation that contains
the currently-expected development and release schedules.)
---------------------------- #include <rtfaq.h> -----------------------------
For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com
--------------------------- pure personal opinion ---------------------------
Hoff (Stephen) Hoffman OpenVMS Engineering hoffman#xdelta.zko.dec.com
rasler
Well I work for a reseller in the UK and we are still selling as many as
ever....lots per week! Currently the fashion is for ES40 & DS20 E machines
by the bundell .
I love alpha, customers love alpha, Compaq don't seem to care what customers
want anymore just that they know best !!!
Jools
!
Michael Richards <mi...@interchange.ca> wrote in message
news:v51x7.40695$5h5.18...@news3.rdc2.on.home.com...
Jerry Leslie
: I haven't been following what's going on with the Alpha platform, but
Read the press releases for yourself:
http://www.compaq.com/newsroom/pr/2001/pr2001062501.html
http://www.intel.com/pressroom/archive/releases/20010625corp.htm
: That's what someone told me but I haven't really been able to find any
: indication of this. Since you folks probably deal with this platform on
: a day to day basis I'm wondering on what your opinions are
: on the safety of investing in a server farm of Alpha machines?
:
: Is the Alpha processor really dying or is it still quietly going strong?
HP's PA-RISC and Compaq's ALPHA will be supplanted by Intel's IA64
processors.
--Jerry Leslie (my opinions are strictly my own)
Michael Richards
which
> claims that there will be a smooth transition from Alpha to IPF in the
> post-McKinley era. Compaq and Intel are collaborating on a joint
parallelism
> project to help make this so. It's difficult to predict the future, but
the
> US Government just inked an $857M deal that'll involve a heck of a lot of
> AlphaServers over the next few years.
This is specifically the curious part. Since there is some sort of joint
effort between Intel and Compaq does this mean that the Alpha chip is going
to become an Intel chip ala x86 type of distribution that finds one on every
desktop?
It would seem to me that RISC has been showing better results for a number
of years. I'm not a hardware engineer so this is just my impression. For
this reason it seems that maybe Intel needs some Alpha technology.
Just out of curiosity, how does the performance of the 1ghz alpha chip
compare to a P4 1ghz?
-Michael
Bill Todd
Michael Richards <mi...@interchange.ca> wrote in message
news:v51x7.40695$5h5.18...@news3.rdc2.on.home.com...
Alpha is in the process of dying, but it's too early to know exactly how
fast it will do so. Compaq has declared its intention to stop releasing new
designs as of EV7, scheduled to hit the streets in something over a year
from now (exactly when depends on how aggressively - or the reverse - Compaq
wants to move). It has tried to reassure customers with a commitment to a
later minor release (EV79) that may include a process shrink and minor
tweaks (someone who knows more about such plans than I do could comment, but
since Compaq's 'commitments' are demonstrably worth nothing even EV7's
release can't be counted upon).
If Alpha sales drop through the floor as a result of this decision, then all
bets are off (including Compaq's survival, since Alpha systems contributed
the lion's share of its profits for the past two years). If Alpha sales
don't drop too much, then the planned port of VMS to Itanium may happen,
though won't be fully complete until 2004 (with earlier ISV releases), and
EV7 (if released) seems likely to out-perform Itanium until at least 2005
unless Compaq does nothing to enhance it.
Given the precarious position of Alpha, VMS, Tru64, and Compaq itself,
Compaq's demonstrated willingness to kill products (especially
'non-industry-standard' products) that it has fervently 'committed' to
support, and the highly-questionable competence of Compaq's current
management, investing in *anything* that depends upon Compaq's continued
support (or existence) seems risky (though there may be some reason to
suspect that Tandem division products will survive no matter what).
- bill
Chris Morgan
> Just out of curiosity, how does the performance of the 1ghz alpha chip
> compare to a P4 1ghz?
Irrelevant in all respects. The Pentium4 is designed for higher
clockspeeds than alphas. The key comparison is always between shipping
chips. Alpha currently tops out at 1GHz and Pentium 4 at 2 GHz. The
Pentium 4 is comparable in performance and much, much cheaper.
Just as it's true that you can't say more MHz => better performance,
it's also true that you can't ignore better performance just because
it uses more MHz. As a recent article I saw put it, it's not how fast
the pistons move, it's the performance of the car.
--
Chris Morgan <cm at mihalis.net> http://www.mihalis.net
Temp sig. - Enquire within
Terry C. Shannon
> > If you rummage around on Compaq's HPTC site you'll find a presentation
> which
> > claims that there will be a smooth transition from Alpha to IPF in the
> > post-McKinley era. Compaq and Intel are collaborating on a joint
> parallelism
> > project to help make this so. It's difficult to predict the future, but
> the
> > US Government just inked an $857M deal that'll involve a heck of a lot
of
> > AlphaServers over the next few years.
>
> This is specifically the curious part. Since there is some sort of joint
> effort between Intel and Compaq does this mean that the Alpha chip is
going
> to become an Intel chip ala x86 type of distribution that finds one on
every
> desktop?
I don't think this will be the case. The Alpha superscalar architecture and
the IPF EPIC architecture are dramatically different and I doubt Intel would
forsake EPIC at this stage of the game. What is likely to happen is that
Intel will incorporate Alpha features (those not previously purloined) such
as on-chip switching and glueless SMP, perhaps SMT sometime in the future,
and incorporate these artifacts into future IPF designs.
Now, if Intel had accepted DEC's offer to license the Alpha design way back
in 1990, things might look different today.
>
> It would seem to me that RISC has been showing better results for a number
> of years. I'm not a hardware engineer so this is just my impression. For
> this reason it seems that maybe Intel needs some Alpha technology.
>
They certainly will benefit from the compiler expertise (the Achille's Heel
of IPF) as well as the aforementioned architectural enhancements.
Terry C. Shannon
"Bill Todd" <bill...@metrocast.net> wrote in message
news:CE3x7.885388$ai2.68...@bin2.nnrp.aus1.giganews.com...
A commitment and a dollar seventy five will get you a cup of coffee at a
Fivebucks(tm) franchise, but that's about it.
That said, I'd be very surprised to see EV7 get scuttled. Plenty of
resources are being pumped into EV7/Marvel. The "commitments" include EV78
and the EV79 shrink; if post-McKinley IPF parts are as "fashionably late" as
Itanium proved to be there is no technical reason why CPQ couldn't do an
EV10 or EV11 shrink.
>
> If Alpha sales drop through the floor as a result of this decision, then
all
> bets are off (including Compaq's survival, since Alpha systems contributed
> the lion's share of its profits for the past two years). If Alpha sales
> don't drop too much, then the planned port of VMS to Itanium may happen,
> though won't be fully complete until 2004 (with earlier ISV releases), and
> EV7 (if released) seems likely to out-perform Itanium until at least 2005
> unless Compaq does nothing to enhance it.
>
> Given the precarious position of Alpha, VMS, Tru64, and Compaq itself,
> Compaq's demonstrated willingness to kill products (especially
> 'non-industry-standard' products) that it has fervently 'committed' to
> support, and the highly-questionable competence of Compaq's current
> management, investing in *anything* that depends upon Compaq's continued
> support (or existence) seems risky (though there may be some reason to
> suspect that Tandem division products will survive no matter what).
>
The Himalaya NSK franchise is indeed safe. If Compaq (or HPaq) scuttle EV7,
the survival of the firm is indeed in grave doubt. Just look at where
Compaq's margins come from: it's not from Wintel gear!
Pentyern
while I followed the discussion about dying or not dying of the Alpha
platform, I ask myself how Samsug is involved in this prozess. Aren't they
involved in the developing process or are they just manufacturing
alpha-prozessors?
Terry C. Shannon
"Pentyern" <pent...@gmx.net> wrote in message
news:9q3f84$euc$1...@tyfon.itea.ntnu.no...
Samsung has an architectural license, thus they have the right to fab
existing Alpha designs and to develop their own variants (providing that
said variants are compatible with the Alpha architecture). I don't believe
that Samsung will exercise this right, they'll just keep fabbing for the
next several years.
Note that Samsung/Compaq spinoff Alpha Processor Incorporated renamed itself
API NetWorks (gotta love that InterCapping) and is now focused more on
network and transport stuff than on the Alpha chip itself.
Mike Foley
"Terry C. Shannon" <terrys...@mediaone.net> wrote in message>
> Note that Samsung/Compaq spinoff Alpha Processor Incorporated renamed
itself
> API NetWorks (gotta love that InterCapping) and is now focused more on
> network and transport stuff than on the Alpha chip itself.
Their primary focus now is building HyperTransport(tm) chips. The Alpha
server business
is pretty much done. Samsung, from what I have heard, is still selling
and planning to
sell Alpha CPU's to Compaq.
Just about all of us Alpha server folks were laid off in August.
mike
Former
technical marketing engineer at API
Currently an
unemployed student
Terry C. Shannon
"Mike Foley" <mikie...@yahoo.com> wrote in message
news:ttghkat...@corp.supernews.com...
>
> "Terry C. Shannon" <terrys...@mediaone.net> wrote in message>
> > Note that Samsung/Compaq spinoff Alpha Processor Incorporated renamed
> itself
> > API NetWorks (gotta love that InterCapping) and is now focused more on
> > network and transport stuff than on the Alpha chip itself.
>
> Their primary focus now is building HyperTransport(tm) chips. The
Alpha
> server business
> is pretty much done. Samsung, from what I have heard, is still selling
> and planning to
> sell Alpha CPU's to Compaq.
>
> Just about all of us Alpha server folks were laid off in August.
>
Sorry to hear that. IMHO API failed to leverage a business tactic called
"marketing." (Sounds familiar, doesn't it?). What's more, API was beholden
to Samsung's outrageously aggressive Alpha roadmaps, which appeared to be
made of whole cloth right from the get-go.