1) Relaunch Alpha
2) Hire/Steal back engineers to do so
3) Port everything high-end to Alpha
4) Dump Intel and get in bed with AMD
5) Advertise, Advertise, Advertise
6) Bring in new teams to market products like VMS and Tru64 and HP-UX.

--
OpenVMS - The classics never go out of style.

John Smith wrote:
> 1) Relaunch Alpha
> 2) Hire/Steal back engineers to do so
> 3) Port everything high-end to Alpha
> 4) Dump Intel and get in bed with AMD
> 5) Advertise, Advertise, Advertise
> 6) Bring in new teams to market products like VMS and Tru64 and HP-UX.

1.  No:  that ceased to be a feasible option over 2 years ago, even if 
all the engineers could have been retrieved.  You just can't add 2 years 
(almost 4, now) to *any* processor's schedule and have it remain a 
leader (just look at how Itanic fared:  if McKinley had been released 
even as late as mid-2000, it would have been really exciting) - and with 
no work having been done on post-EV8 Alpha designs since 2001, no 
fast-pathing of their development could occur either
2.  Some, but not for the purpose of (1).
3.  No.
4.  At least somewhat, though AMD probably couldn't meet HP's needs 
without help from Intel.
5.  Yes.
6.  Yes.

7.  You forgot the minor matter of product development (or more to the 
point HP's lack of same over the past few years).  Storage, OSs, ... 
Just as Alpha is no longer competitive, or even revivable, after too 
lengthy a period of neglect, so other HP core technologies are highly 
threatened.

Carly may actually have succeeded in sinking HP beyond recovery, at 
least anything resembling the force that HP used to represent in the 
industry.

- bill

John Smith wrote:
> 
> 1) Relaunch Alpha

Out of curiosity, if relaunch were to happen, would relaunching Alpha
have any serious advantage over relaunching PaRisc ?

> 4) Dump Intel and get in bed with AMD

No need to "dump Intel". HP should simply have its own strategy instead
of following what Microsoft and Intel. It doesn't mean that HP needs to
stop buying from Intel if Intel has superior products.

On Thu, 10 Feb 2005 20:19:31 -0500, Bill Todd <billtodd@metrocast.net>  
wrote:

> John Smith wrote:
>> 1) Relaunch Alpha
>> 2) Hire/Steal back engineers to do so
>> 3) Port everything high-end to Alpha
>> 4) Dump Intel and get in bed with AMD
>> 5) Advertise, Advertise, Advertise
>> 6) Bring in new teams to market products like VMS and Tru64 and HP-UX.
>
> 1.  No:  that ceased to be a feasible option over 2 years ago, even if  
> all the engineers could have been retrieved.  You just can't add 2 years  
> (almost 4, now) to *any* processor's schedule and have it remain a  
> leader (just look at how Itanic fared:  if McKinley had been released  
> even as late as mid-2000, it would have been really exciting) - and with  
> no work having been done on post-EV8 Alpha designs since 2001, no  
> fast-pathing of their development could occur either
> 2.  Some, but not for the purpose of (1).
> 3.  No.
> 4.  At least somewhat, though AMD probably couldn't meet HP's needs  
> without help from Intel.
> 5.  Yes.
> 6.  Yes.
>
> 7.  You forgot the minor matter of product development (or more to the  
> point HP's lack of same over the past few years).  Storage, OSs, ...  
> Just as Alpha is no longer competitive, or even revivable, after too  
> lengthy a period of neglect, so other HP core technologies are highly  
> threatened.
>
> Carly may actually have succeeded in sinking HP beyond recovery, at  
> least anything resembling the force that HP used to represent in the  
> industry.
>
> - bill

Looking at it from the point of view of finding a vehicle to deliver VMS,  
that
is reasonably competitive (doesn't have to be the fastest) and finding a  
partner
to procure that vehicle who you believe will continue developing that  
vehicle and
has the where-with-all to accomplish it, leads you to selecting two, three  
suppliers.
You port VMS to the Power chip and Opteron/Pentium.  The ongoing cost of  
sustaining
Alpha is prohibitive, particularly since you can't amortise it over a  
large run.

This also gives flexibility, as we used to have, by demanding second  
sources.  The real
value in the Digital component of HP is VMS, Alpha has lost all its value.

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

Tom Linden wrote:
>  The real
> value in the Digital component of HP is VMS, Alpha has lost all its value.

There is one aspect of Alpha: installed base and available software,
much of which is no longer developped and thus not be available on
Alpha's successor. Continuing the Alpha processor ratins that software
whereas moving to a new platform requires VMS start from scratch with
software availability.

Note that this is also a reason HP-UX customers are still buying PaRisc
based machines instead of that unwanted IA64 thing. But VMS is in a much
worse shape than HP-UX in terms of application ecosystem.

Bill Todd wrote:
> John Smith wrote:
> 
>> 1) Relaunch Alpha
>> 2) Hire/Steal back engineers to do so
>> 3) Port everything high-end to Alpha
>> 4) Dump Intel and get in bed with AMD
>> 5) Advertise, Advertise, Advertise
>> 6) Bring in new teams to market products like VMS and Tru64 and HP-UX.
> 
> 
> 1.  No:  that ceased to be a feasible option over 2 years ago, even if 
> all the engineers could have been retrieved.  You just can't add 2 years 
> (almost 4, now) to *any* processor's schedule and have it remain a 
> leader (just look at how Itanic fared:  if McKinley had been released 
> even as late as mid-2000, it would have been really exciting) - and with 
> no work having been done on post-EV8 Alpha designs since 2001, no 
> fast-pathing of their development could occur either
> 2.  Some, but not for the purpose of (1).
> 3.  No.
> 4.  At least somewhat, though AMD probably couldn't meet HP's needs 
> without help from Intel.
> 5.  Yes.
> 6.  Yes.
> 
> 7.  You forgot the minor matter of product development (or more to the 
> point HP's lack of same over the past few years).  Storage, OSs, ... 
> Just as Alpha is no longer competitive, or even revivable, after too 
> lengthy a period of neglect, so other HP core technologies are highly 
> threatened.
> 
> Carly may actually have succeeded in sinking HP beyond recovery, at 
> least anything resembling the force that HP used to represent in the 
> industry.

OK - it's not entirely fair to be so dismissive without coming up with a 
counter-proposal (unless there's just no hope at all, which is not yet 
clear).  So here's my take (not that anyone in any position to act is 
likely to be listening to it):

HP will never regain the proprietary hardware advantages that PA-RISC 
and Alpha gave it:  as I noted above, far too much time has gone by for 
either architecture to regain a leadership position in the industry. 
EV7, for example, has roughly half the SPECint and SPECfp performance of 
the current industry leaders, and even if aggressively upgraded to EV79 
would significantly trail what the competition has today, let alone what 
they'll have when such an upgrade could ship.  EV8 could of course do 
better, but would still likely trail a bit by the time it could ship.

The only way to regain a leadership position in processors would be to 
come up with something new (at least as new as Cell, for example - 
though it's not clear that Cell would do the job).  And it's arguable 
that by the time something new could be developed (end of the decade at 
the earliest) a leadership position in processors may not count for that 
much:  the industry is already having difficulty coming up with useful 
processor-specific things to do with the die area on its processor 
chips, and the 'memory wall' continues to dilute the value of higher 
processor performance in many applications.

HP could still provide Alpha with the limited upgrades which *are* 
feasible, to support its remaining customer base (such as it may be). 
And in any event it should promise to sell Alphas (at something 
resembling reasonable prices) as long as customers wish to buy them. 
But that's a relatively short-term fix:  Alpha's OSs need something more 
permanent in order to be perceived as long-term viable rather than on 
the way out.

And of course that does nothing to address the HP-UX customer base, 
which (especially now) is likely considerably larger than Alpha's.

Lousy decision though it was, by now depending on Itanic as the 
longer-term hardware platform for all these OSs may be a fait accompli. 
  With its significant power reductions Montecito has become at least 
reasonable to consider for this purpose:  not what things could have 
been, but perhaps the best that can be done as things now are.

There remains the concern about just how committed Intel may be to 
future Itanic development, however:  at a minimum, Itanic processor 
pricing seems likely to remain high and its performance no better than 
x86-64 equivalents over the next few years.  The cheapest insurance (and 
with side-benefits as well) against problems here would be to port VMS 
and Tru64 to x86-64 (whether the latter is also ported to Itanic or 
not); this is not as viable an option for HP-UX, since it's a big-endian 
environment, so Itanic may be the only realistic answer there.

Alternatives?

1.  Plow a fair amount of money into temporarily extending the life of 
both Alpha and PA-RISC while developing a new processor to replace them 
both.  But what will make that processor (developed at impressive cost, 
regardless of how good it might be) better than its market competition? 
  Even Alpha had some difficulty staying convincingly ahead of x86 
performance by the late '90s, for example - and by the time a new 
architecture could ship, it seems likely that existing ones will have 
incorporated most of the significant features it could boast of (since 
the percentage of die space dedicated to processor cores seems to be 
shrinking fast, incorporating everything, including the kitchen sink, 
into them becomes increasingly easy).

2.  Port the OSs somewhere else.  This might be fairly easy for VMS and 
even Tru64, though perhaps harder for HP-UX (since Itanic was designed 
with its needs in mind).  POWER is certainly impressive, but owned by a 
direct competitor.  x86-64 is attractive, but would still entail porting 
expenses and delays (Itanic could help the OSs ride out the delay, or 
continuations of Alpha and PA-RISC could - but, as noted above, there's 
also the endian question for HP-UX).  SPARC is also competitor-related, 
and while it may be a good platform for vendors who are already on it it 
isn't as attractive for new vendors to move to.  Partnering with AMD 
would be contrarian and exciting, but given the direction Intel seems to 
be moving (in terms of incorporating significant elements of the system 
architecture into the chips using at least part of the expertise that 
the Alpha team brought with them, come 2007 or so) might not offer any 
real advantage over Intel's 2007 and later x86 products (i.e., I think 
that working with *both* Intel and AMD and using the best each has to 
offer makes sense).

3.  Get out of the proprietary OS market.  That was the Carly & Curly 
vision, and they managed to sabotage the viability of the other options 
pretty thoroughly.  But I submit that this approach effectively cedes 
enterprise business almost entirely (save possibly for storage and, if 
HP managed to execute a full about-face in its attitudes toward quality, 
some services):  HP still has major strengths in system expertise that 
it could leverage profitably, and I think it should do so.


OK:  so HP should return to aggressively pursuing its proprietary 
advantages in system hardware and OSs (though without abandoning Linux, 
and even supplying and supporting Windows where that makes sense).  It 
should port VMS to x86-64 both as insurance and to increase its exposure 
and applicability.  It should port Tru64 to *both* Itanic and x86-64 for 
the same reasons, plus as clear evidence of its change in direction.  It 
should devote significant resources not only to marketing but to 
developing its proprietary OSs, including reasonable efforts at 
convergence - again, both for competitive reasons and to provide clear 
evidence of its new direction.  Part of this should include significant 
innovation at the file-system level (as Sun has done with its new ZFS 
file system) - and not only because that's my primary area of personal 
interest.

It should extend its OS expertise into PC systems, as DEC at least 
started to do - e.g., make clustering PCs something other than a joke, 
and extend the file-system work I mentioned above to include them.  The 
company that controls the data controls the world, which is one reason 
why Microsoft is shooting for more lock-in with its new file system for 
Longhorn (not that those shots have necessarily come close to the target 
yet):  while any PC file system extensions HP might create could (and 
probably should) be available for anyone to use, the integration with 
HP's higher-end systems would constitute a significant proprietary 
advantage.

It should revitalize its storage R&D efforts, since between the DECpaq 
and original HP storage groups it had much of the cream of the industry 
and could get at least some of it involved again if they could be 
convinced it was serious.  It should get its service organization into 
shape rather than seeing just how much of its work can be accomplished 
(at least sort-of) via satellite links to India.

It should do with its PC division what IBM used to:  partner 
intelligently, ship a quality product, and provide one-stop shopping 
(and service) as a convenience at prices just high enough to make a 
modest profit, rather than try to compete tooth and nail with Dell and 
the Asian white-box makers.

If HP actually became an exciting company again the industry would be 
excited (and receptive) too, because it really wants a *credible* 
alternative to IBM and there's no one around providing one (though if HP 
fails to, Sun may pick up that mantle by default).

- bill

Bill Todd wrote:
> John Smith wrote:
> 
>> 1) Relaunch Alpha
>> 2) Hire/Steal back engineers to do so
>> 3) Port everything high-end to Alpha
>> 4) Dump Intel and get in bed with AMD
>> 5) Advertise, Advertise, Advertise
>> 6) Bring in new teams to market products like VMS and Tru64 and HP-UX.
> 
> 
> 1.  No:  that ceased to be a feasible option over 2 years ago, even if 
> all the engineers could have been retrieved.  You just can't add 2 years 
> (almost 4, now) to *any* processor's schedule and have it remain a 
> leader (just look at how Itanic fared:  if McKinley had been released 
> even as late as mid-2000, it would have been really exciting) - and with 
> no work having been done on post-EV8 Alpha designs since 2001, no 
> fast-pathing of their development could occur either

I'd dispute that, from a certain perspective.  If by relaunch one would 
mean jump back into it at the point it was in 2001, no, that wouldn't be 
so good due to lost time.  EV8's time is now and the near future, and 
EV8 doesn't exist, and probably couldn't exist in that time frame.

Not that that would be a total dog since EV7 even now is still a viable 
platform for today's users.  As part of a solution, evolving EV7 a bit 
would help with business today.  But, you're selling to the already 
captive audience for the most part.  Acquiring new customers when the 
competition is Power5+ with a 3+ year lead would be rather tough, from a 
performance perspective.  VMS and T64 would possibly be some advantage. 
  But you have to ask, where are the T64 engineers?

Saying that getting back into the CPU arena is not feasible is the same 
as saying that all the possible CPU manufacturers already exist, and 
someone new, with adequate talent and capitol couldn't become a player. 
  That I won't believe.  What would be required would be spending some 
time to come up with a product in say 3-4 years which would be 
competitive with where Power will be in 3-4 years, and where Alpha 
'could' have possibly been in 3-4 years.

The list of possibilities isn't small.  One could take a look at the 
'cell' concept that IBM, Sony, and Toshiba just announced.  One could 
consider using Power.  The real issue is what market would you be aiming 
at?  If it's low volume, high cost, I think that's a dead end.  Low 
volume, low cost is one possibility.  The best would be high volume, low 
cost.  What that is is the real question.  What I vaguely remember about 
the research I think was called EV10 was along those lines.  A large 
number of low cost processors coupled together to give great 
performance.  Look at the Linux 'clusters' we've read about with 
hundreds or thousands of processors.  Lots of compute cycles, but not 
coupled together very well.

The real key point Bill makes is "no work having been done on post-EV8 
Alpha designs since 2001".

Dave

Dave Froble wrote:

....

> Saying that getting back into the CPU arena is not feasible is the same 
> as saying that all the possible CPU manufacturers already exist, and 
> someone new, with adequate talent and capitol couldn't become a player. 

No, it is not:  it's saying (in part) that HP has much shorter-term 
problems which it must solve than could be addressed by a new 
architecture, else it wouldn't have anything viable left to run on said 
architecture when it finally appeared.

Plus the minor point that any such new architecture has no reason to be 
any better, after all the money spent to develop it, than some existing 
one - unless it also incorporates some significantly better idea than 
existing architectures contain.

>  That I won't believe.

People usually believe what they want to believe.  That has little 
relevance to reality, unless they can convince large numbers of other 
people to share their vision or delusion (as the case may be).

   What would be required would be spending some
> time to come up with a product in say 3-4 years which would be 
> competitive with where Power will be in 3-4 years, and where Alpha 
> 'could' have possibly been in 3-4 years.

What part of the fact that doing something like that, *if* you succeeded 
at all, would take more like 5 - 6 years is difficult for you to 
comprehend?  The Alpha team hit the ground running (as a coherent 
existing development group, no less) at Intel in mid-2001, and - had 
Tukwila not been canceled - would have taken at least 5.5 years to get 
it to market.

- bill

Dave Froble wrote:
> I'd dispute that, from a certain perspective.  If by relaunch one would
> mean jump back into it at the point it was in 2001, no, that wouldn't be
> so good due to lost time.  EV8's time is now and the near future, and
> EV8 doesn't exist, and probably couldn't exist in that time frame.


I used to have the DEC brainwash and think that Alpha was decades ahead
of its time. But over time, I realised that what this was all about was
speed to market. Competitors had features similar to Alpha, but
delivered later.

If you were to reconstitute an Alpha team today, they wouldn't start
where they last left off. They'd start a new project today with features
that Intel, IBM and Sun are starting to work on too. And they'd benefit
from other developments done while they were "away". They may not be
able to copy what IBM did to Power, but they'd get some inspiration of
what works and what doesn't.


If the Alpha instruction set does allow for greater speeds and easier to
develop and faster to market, then the newly reconstituted team would be
able to move to the next generation in time to regain its edge within a
few years.

Lets take an extreme case: lets say they were to revive the VAX chip.
Engineers wouldn't have to go though every iteration that they would
have gone though since 1996. They'd develop a twin core VAX with shared
cache with a 90nm process and instantly bring the VAX into the 21st
century in one generation.

copying :-)  If the Alpha team are able to deliver to market faster,
then within a couple of years Alpha would regain its edge. Meanwhile it
could have just simple speed bumps. 

One has to look at the big picture now. IBM is marginalizing the wintel
market and pushing Power to lower end.  Should HP do the same with
Alpha, it would relegate wintel to cheap commodity desktops and shut
Windows out of true enterprise, leaving Dell with an essential monopoly
(which would probably result in rise in wintel server prices due to less competition).

Bill Todd wrote:
> existing development group, no less) at Intel in mid-2001, and - had
> Tukwila not been canceled - would have taken at least 5.5 years to get
> it to market.

Working with a dog takes longer. Now, even if it took 5.5 years to get a
new up-to-date Alpha to the market, they could still do process shrinks
of EV7 and do some work at the system level to increase performance.

Consider what it has taken 4.5 years between murder of Alpha and first
commercial version of VMS on that IA64 thing. And there isn't much
software available.

So, if you're going to develop a new platform starting on 2006 when HP
announces the widthdrawal of IA64, it would take about 4 years to get
VMS to restart again. If you announce rebirth of Alpha, you are going to
take about as much time but meanwhile, you inherit the complete software
ecosystem which can grow during thsoe year sbecause VMS is already
available on Alpha, so is Linux.

And consider SGI going Alpha, it would allow SGI to migrate immediatly
to Alpha since Alpha already exists.

In article <420C09D0.795A5F42@teksavvy.com>, Jf Mezei <jfmezei.spamnot@teksavvy.com> writes:
> John Smith wrote:
>> 
>> 1) Relaunch Alpha
> 
> Out of curiosity, if relaunch were to happen, would relaunching Alpha
> have any serious advantage over relaunching PaRisc ?

   1) Alpha was always faster than PaRisc
   2) Porting HP-UX to Alpha, running in big-endian mode would be
      easier than porting VMS to PaRisc.

John Doe wrote:
> Bill Todd wrote:
> 
>>existing development group, no less) at Intel in mid-2001, and - had
>>Tukwila not been canceled - would have taken at least 5.5 years to get
>>it to market.
> 
> 
> Working with a dog takes longer.

And starting with an already-existing group takes shorter, so it 
balances out.

  Now, even if it took 5.5 years to get a
> new up-to-date Alpha to the market, they could still do process shrinks
> of EV7 and do some work at the system level to increase performance.

Which I already noted.  The point is that there's no particular reason 
to believe that an up-to-date Alpha in late 2010 would be worth the 
effort it took to reconstitute the resources necessary to develop it and 
then do so.  There's a large difference between having been an industry 
leader for nearly 20 straight years and being a resurrection of a 
product which had been dead for quite some time while the rest of the 
industry caught up with and then surpassed it.

> 
> Consider what it has taken 4.5 years between murder of Alpha and first
> commercial version of VMS on that IA64 thing.

It has been just over 3.5 years.  But it's not clear what relevance 
either figure has to the question of reentering the leading-edge 
processor biz.

  And there isn't much
> software available.

If HP makes it clear that it now takes VMS seriously, there will be - 
long before 2010.

> 
> So, if you're going to develop a new platform starting on 2006 when HP
> announces the widthdrawal of IA64, it would take about 4 years to get
> VMS to restart again. If you announce rebirth of Alpha, you are going to
> take about as much time but meanwhile, you inherit the complete software
> ecosystem which can grow during thsoe year sbecause VMS is already
> available on Alpha, so is Linux.
> 
> And consider SGI going Alpha, it would allow SGI to migrate immediatly
> to Alpha since Alpha already exists.

SGI has no reason to switch from what it's got now, unless Intel 
abandons Itanic.  And if that happens, SGI will either die or switch to 
a faster processor than Alpha is likely to be before - again - late 2010.

- bill

Bill Todd wrote:

> 
> 2.  Port the OSs somewhere else.  This might be fairly easy for VMS and 
> even Tru64, though perhaps harder for HP-UX (since Itanic was designed 
> with its needs in mind).

"fairly easy for VMS"?  Can I take a hit of what you are smoking?  The 
large amounts of BLISS and especially Macro-32 code in OpenVMS make any 
port more difficult than a system like Tru64 with is mostly in C.  We 
have more compilers to find/build.  Happen to have a AMD64 BLISS 
compiler sitting around?

And to the best of my knowledge, I don't know of anything in Itanium 
that was specially developed for HP-UX.  The instruction set is pretty 
unbiased.

Perhaps some of the memory management support... but if OpenVMS could 
retrofit our VAX/Alpha designs to Itanium, I'd suspect that HP-UX could 
handle that transition as well.

HP-UX's big-endian design might be a better fit for POWER5 than for an 
X86 platform.


-- 
John Reagan
HP Pascal/{A|I}MACRO for OpenVMS Project Leader
Hewlett-Packard Company

On Thu, 10 Feb 2005 21:59:23 -0500, JF Mezei  
<jfmezei.spamnot@teksavvy.com> wrote:

> Tom Linden wrote:
>>  The real
>> value in the Digital component of HP is VMS, Alpha has lost all its  
>> value.
>
> There is one aspect of Alpha: installed base and available software,
> much of which is no longer developped and thus not be available on
> Alpha's successor. Continuing the Alpha processor ratins that software
> whereas moving to a new platform requires VMS start from scratch with
> software availability.
>
> Note that this is also a reason HP-UX customers are still buying PaRisc
> based machines instead of that unwanted IA64 thing. But VMS is in a much
> worse shape than HP-UX in terms of application ecosystem.

Face it Alpha is dead.  To make VMS a viable business it _will_ need
to use a supplier for its chips such as Power /and/or Opteron.


-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

In article <G+KLwPFoAVFx@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:
>In article <420C09D0.795A5F42@teksavvy.com>, Jf Mezei <jfmezei.spamnot@teksavvy.com> writes:
>> John Smith wrote:
>>> 
>>> 1) Relaunch Alpha
>> 
>> Out of curiosity, if relaunch were to happen, would relaunching Alpha
>> have any serious advantage over relaunching PaRisc ?
>
>   1) Alpha was always faster than PaRisc
>   2) Porting HP-UX to Alpha, running in big-endian mode would be
>      easier than porting VMS to PaRisc.
>
Also how long would it take to port Tandem NSK to PA-RISC ?
How close was the port to Alpha to completion when the decision to move to IA64
was made ?

David Webb
Security team leader
CCSS
Middlesex University

On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk> wrote:

> Also how long would it take to port Tandem NSK to PA-RISC ?
> How close was the port to Alpha to completion when the decision to move  
> to IA64
> was made ?

Maybe the question that should be asked, is how long would it take to
provide migration tools from NSK to VMS?  Or alternatively, to whom can we
sell Non-Stop business unit?

I would give odds that HP will start shedding business units.

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

John Reagan wrote:
> Bill Todd wrote:
> 
>>
>> 2.  Port the OSs somewhere else.  This might be fairly easy for VMS 
>> and even Tru64, though perhaps harder for HP-UX (since Itanic was 
>> designed with its needs in mind).
> 
> 
> "fairly easy for VMS"?  Can I take a hit of what you are smoking?

Ask your co-workers for one:  they're the ones who have maintained that 
the port of VMS to Itanic had the side-benefit of making VMS much more 
portable in general.

Perhaps they were just rationalizing - but that's what they said, and 
rather unambiguously.

- bill

Bill Todd wrote:
> John Reagan wrote:
> 
>> Bill Todd wrote:
>>
>>>
>>> 2.  Port the OSs somewhere else.  This might be fairly easy for VMS 
>>> and even Tru64, though perhaps harder for HP-UX (since Itanic was 
>>> designed with its needs in mind).
>>
>>
>>
>> "fairly easy for VMS"?  Can I take a hit of what you are smoking?
> 
> 
> Ask your co-workers for one:  they're the ones who have maintained that 
> the port of VMS to Itanic had the side-benefit of making VMS much more 
> portable in general.
> 
> Perhaps they were just rationalizing - but that's what they said, and 
> rather unambiguously.
> 
> - bill

True.  Since the Alpha and Itanium versions are indeed built from a 
common source area, most of the source code (whether Macro-32, BLISS, or 
C) is in much better shape.  Much of the platform-specific code has been 
identified and isolated with macros, etc.

Not to disagree with my co-workers who just ported from Alpha to I64, as 
somebody who has ported compilers from VAX to Ultrix to X86 to Alpha to 
Tru64 to Itanium, going from 1 target to 2 finds about 80% of the places 
that are really platform-specific.  It wasn't until the 3 or even 4th 
target that I had the Pascal compiler pretty portable.  Porting it to 
OpenVMS I64 from OpenVMS Alpha took me about 2 days with the RTL taking 
another day.

However, that all assumes the existance of Macro-32, BLISS, and C 
compilers for any future target.  C compilers you can find (although our 
C source uses many Compaq C extensions that might be hard to eliminate). 
  Macro-32 and BLISS compilers are the problem that turns this from 
"fairly easy" to "90% fairly easy; 10% get several buckets of cash and 
come back in a year and hope I haven't pulled out what is left of my hair".

-- 
John Reagan
HP Pascal/{A|I}MACRO for OpenVMS Project Leader
Hewlett-Packard Company

Tom Linden wrote:
> On Thu, 10 Feb 2005 21:59:23 -0500, JF Mezei  
> <jfmezei.spamnot@teksavvy.com> wrote:
> 
>> Tom Linden wrote:
>>
>>>  The real
>>> value in the Digital component of HP is VMS, Alpha has lost all its  
>>> value.
>>
>>
>> There is one aspect of Alpha: installed base and available software,
>> much of which is no longer developped and thus not be available on
>> Alpha's successor. Continuing the Alpha processor ratins that software
>> whereas moving to a new platform requires VMS start from scratch with
>> software availability.

This is an argument for Alpha, and VAX.  It's a limited argument.

>> Note that this is also a reason HP-UX customers are still buying PaRisc
>> based machines instead of that unwanted IA64 thing. But VMS is in a much
>> worse shape than HP-UX in terms of application ecosystem.
> 
> 
> Face it Alpha is dead.  To make VMS a viable business it _will_ need
> to use a supplier for its chips such as Power /and/or Opteron.

Much as I hate to, I have to admit that Alpha is most likely history. 
To what purpose should HP spend the considerable money it would take, 
and for what, just another "me too".

There are 2 issues, near term, and far future.  In the near term, 
continuing to sell Alphas, with whatever enhancements can be done for 
EV7, is one solution.  Using the itanic when it's a fit is also 
reasonable.  For far term, I can see a costly and tremendous effort to 
pull Alpha back to the front of the pack, just to see that new advances 
have made all today's techniques obsolete.  All dressed up, with nowhere 
to go.

Dave

Bill Todd wrote:
> John Reagan wrote:
> 
>> Bill Todd wrote:
>>
>>>
>>> 2.  Port the OSs somewhere else.  This might be fairly easy for VMS 
>>> and even Tru64, though perhaps harder for HP-UX (since Itanic was 
>>> designed with its needs in mind).
>>
>>
>>
>> "fairly easy for VMS"?  Can I take a hit of what you are smoking?
> 
> 
> Ask your co-workers for one:  they're the ones who have maintained that 
> the port of VMS to Itanic had the side-benefit of making VMS much more 
> portable in general.
> 
> Perhaps they were just rationalizing - but that's what they said, and 
> rather unambiguously.
> 
> - bill

The OS itself is one thing.  I think John still has a point with the 
compilers.

Hey, job security John.

John Smith wrote:

> 4) Dump Intel and get in bed with AMD

Well, Intel is a "mission critical" user of VMS and until we hear that
they've dumped VMS, I think we should like having Intel on our side.
Does AMD use VMS? What "mission critical" app's do HP run on VMS?

Tom Linden wrote:
> On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk>
> wrote:
>
>> Also how long would it take to port Tandem NSK to PA-RISC ?
>> How close was the port to Alpha to completion when the decision to
>> move to IA64
>> was made ?
>
> Maybe the question that should be asked, is how long would it take to
> provide migration tools from NSK to VMS?  Or alternatively, to whom
> can we sell Non-Stop business unit?
>
> I would give odds that HP will start shedding business units.


IBM would probably snap up NSK in a heartbeat.

--
OpenVMS - The classics never go out of style.

DL Phillips wrote:
> John Smith wrote:
>
>> 4) Dump Intel and get in bed with AMD
>
> Well, Intel is a "mission critical" user of VMS and until we hear that
> they've dumped VMS, I think we should like having Intel on our side.

Intel ran their Fab's on VAX and Alpha before HP bought Compaq, and they'll
continue to run them on VMS until VMS dies unless something better comes
along. Intel is in the business of making money and won't do anything to
upset that apple cart unnecessarily. Alas the same can't be said for HP.



> Does AMD use VMS?

Don't know.


>What "mission critical" app's do HP run on VMS?
None that I'm aware of....not their accounting, not their order fullfilment.
Perhaps the in-house betting pool on carly(tm)'s departure - but now that's
a legacy system.

--

OpenVMS - The classics never go out of style.

John Smith wrote:
> DL Phillips wrote:
> > John Smith wrote:
> >
> >> 4) Dump Intel and get in bed with AMD
> >
> > Well, Intel is a "mission critical" user of VMS and until we hear
that
> > they've dumped VMS, I think we should like having Intel on our
side.
>
> Intel ran their Fab's on VAX and Alpha before HP bought Compaq, and
they'll
> continue to run them on VMS until VMS dies unless something better
comes
> along. Intel is in the business of making money and won't do anything
to
> upset that apple cart unnecessarily. Alas the same can't be said for
HP.
>
>

Yep. So, why does everyone think that Intel isn't trying as hard as it
can to let VMS succeed on an Intel chip?

>
> > Does AMD use VMS?
>
> Don't know.
>

I'll bet we'd know about it if they did.

>
> >What "mission critical" app's do HP run on VMS?
> None that I'm aware of....not their accounting, not their order
fullfilment.
> Perhaps the in-house betting pool on carly(tm)'s departure - but now
that's
> a legacy system.
>

Yep again. So, why should HP care about VMS (other than as a product)
if they don't even use it? Do you think maybe one reason they picked
HP-UX over the obviously superior Tru64 is the fact that HP's business
runs on HP-UX?

I'd say of all the parties being discussed, only Intel has a real
interest in VMS, and Intel carries a *little* weight in the computer
market. I'm pulling for Intel to help keep VMS alive. Maybe the new HP
will see things differently, but only time will tell.

Sure, I'd like a low end AMD/x86-64 port, but I'm not going to hold my
breath.

-Doug

In article <1108161539.104534.64860@c13g2000cwb.googlegroups.com>, "DL Phillips" <whohe@whoever.com> writes:

> Yep. So, why does everyone think that Intel isn't trying as hard as it
> can to let VMS succeed on an Intel chip?

Not "everybody" feels that way.  Just a few prolific posters in this
newsgroup who feel voicing their opinion is worth driving away those
who seek serious technical discussion.

Considering the killfile entries, the net effect can be described as
their desire to talk to themselves.

DL Phillips wrote:
> John Smith wrote:
>> DL Phillips wrote:
>>> John Smith wrote:
>>>
>>>> 4) Dump Intel and get in bed with AMD
>>>
>>> Well, Intel is a "mission critical" user of VMS and until we hear
>>> that they've dumped VMS, I think we should like having Intel on our
>>> side.
>>
>> Intel ran their Fab's on VAX and Alpha before HP bought Compaq, and
>> they'll continue to run them on VMS until VMS dies unless something
>> better comes along. Intel is in the business of making money and
>> won't do anything to upset that apple cart unnecessarily. Alas the
>> same can't be said for HP.
>>
>>
>
> Yep. So, why does everyone think that Intel isn't trying as hard as it
> can to let VMS succeed on an Intel chip?


Intel, to a certain extent, does not care which os runs on its chips.
Intel doesn't own VMS.
The owner of VMS doesn't advertise it.


>>> Does AMD use VMS?
>>
>> Don't know.
>>
>
> I'll bet we'd know about it if they did.
>
>>
>>> What "mission critical" app's do HP run on VMS?
>> None that I'm aware of....not their accounting, not their order
>> fullfilment. Perhaps the in-house betting pool on carly(tm)'s
>> departure - but now that's a legacy system.
>>
>
> Yep again. So, why should HP care about VMS (other than as a product)
> if they don't even use it? Do you think maybe one reason they picked
> HP-UX over the obviously superior Tru64 is the fact that HP's business
> runs on HP-UX?


Nah. HP was better at selling their shitty flavor of unix than DECpaq was at
selling their superior flavor of unix, so there were fewer feathers ruffled
by killing Tru64.

Remember that to people like curly, carly(tm), and Wayman, the numbers are
everything and everything else is nothing. That's just the way it is at
40,000 ft.


> I'd say of all the parties being discussed, only Intel has a real
> interest in VMS, and Intel carries a *little* weight in the computer
> market. I'm pulling for Intel to help keep VMS alive. Maybe the new HP
> will see things differently, but only time will tell.


OpenVMS - Capellas sold it down the river once....looks like he might get
to do it again.



> Sure, I'd like a low end AMD/x86-64 port, but I'm not going to hold my
> breath.

I'd hate to have to do CPR on you for that length of time  ;-)

--
OpenVMS - The classics never go out of style.

DL Phillips wrote:
> John Smith wrote:

....

>>Intel ran their Fab's on VAX and Alpha before HP bought Compaq, and
> 
> they'll
> 
>>continue to run them on VMS until VMS dies unless something better
> 
> comes
> 
>>along. Intel is in the business of making money and won't do anything
> 
> to
> 
>>upset that apple cart unnecessarily. Alas the same can't be said for
> 
> HP.
> 
>>
> 
> Yep. So, why does everyone think that Intel isn't trying as hard as it
> can to let VMS succeed on an Intel chip?

Intel couldn't care less about VMS for VMS's sake:  their fabs will run 
just fine on VMS as it is today (and quite likely as it was years ago).

They'd be happy to see VMS boost Itanic sales, of course - but don't 
seem sufficiently interested in that potential to push HP in that direction.

....

> I'd say of all the parties being discussed, only Intel has a real
> interest in VMS

And you'd be wrong.  As I noted above, Intel doesn't need anything from 
VMS that it doesn't already have (absent increased Itanic sales, and it 
doesn't seem to think that's a possibility worth pursuing).  Whereas HP 
has a significant interest in VMS's potential to help HP's bottom line 
but has been - very persistently - too stupid to recognize it.

- bill

Tom Linden wrote:
> I would give odds that HP will start shedding business units.


Some articles mentioned the new guy just getting rid of everything
inherited from Compaq.

Dave Froble wrote:
> Much as I hate to, I have to admit that Alpha is most likely history.
> To what purpose should HP spend the considerable money it would take,
> and for what, just another "me too".

Competitive advantage.

HP has learned that there is no money to be made in commodity. IBM knew
that a long time ago and has already taken action.

So if HP is to succeed, will it turn itself into a low cost Dell
competitor, or will it try to find some true technological edge to
differentiate itself from the others ?

Alpha may have been asleep for a couple of years, but so has PaRisc.
Nevertheless, resurecting Alpha would still be far easier and cheaper
and less troublesome for customers than going to a totally new platform.

JF Mezei wrote:
> Dave Froble wrote:
> 
>>Much as I hate to, I have to admit that Alpha is most likely history.
>>To what purpose should HP spend the considerable money it would take,
>>and for what, just another "me too".
> 
> 
> Competitive advantage.

I've made repeated suggestions that the only way any competitive 
advantage would accrue to a resurrected Alpha platform would be for it 
to incorporate something superior to the competition, and so far no one 
(including you) has come up with any such feature.

Alpha *would have* led in SMT technology had EV8 been completed, but by 
the time any SMT Alpha could get out now that technology will be old hat 
(either already embraced by all or bypassed by other areas of progress 
such as so much reduction in core sizes that leveraging multiple threads 
per core is no longer particularly useful).  Alpha *does* lead (POWER 
possibly excepted) in integration of on-chip system support, but others 
are already moving in that direction and even Intel's commodity 
processors will have it in a couple of years.  Alpha *used to* lead in 
raw processor performance, but the elephants with less streamlined ISAs 
have learned to dance sufficiently well that any lead in the future 
would not (absent some new secret weapon) be a decisive one.

> 
> HP has learned that there is no money to be made in commodity.

You mean like in printers?

  IBM knew
> that a long time ago and has already taken action.

Commodity products fit into IBM's corporate fold a lot less compatibly 
than they fit into HP's.  Commodity products done *right* could be a 
competitive advantage for HP, as long as they were perceived as 
ancillary to its primary business (i.e., as items provided for customer 
convenience rather than as a core element of its business).

> 
> So if HP is to succeed, will it turn itself into a low cost Dell
> competitor, or will it try to find some true technological edge to
> differentiate itself from the others ?

The question, of course, is where to look for that edge, given limited 
resources.  And since any such edge absolutely *requires* a lot of 
system-level work, adding anything else that isn't absolutely necessary 
would dilute it (it's not as if HP was so healthy that it could throw 
cash around that it didn't really need to).

> 
> Alpha may have been asleep for a couple of years, but so has PaRisc.

And now they're both dead - just haven't yet stopped twitching.

> Nevertheless, resurecting Alpha would still be far easier and cheaper
> and less troublesome for customers than going to a totally new platform.

Having HP go out of business trying to resurrect them would be even more 
troublesome for customers, I suspect.  HP should continue to sell the 
platforms as long as customers want them, but (one more time, absent 
some wonderful idea that would lead to a significantly superior product) 
should look elsewhere for their longer-term future.

Intel and HP wasted vast amounts of cash and resources developing a 
product with no clear advantage over its competition (even though it now 
may be able to survive given that the best competition, POWER excepted, 
has disappeared).  Hopefully HP won't be so stupid that it repeats that 
error.

When you've already *got* a leadership product, throwing it away 
represents gross incompetence.  But so does trying to resurrect it later 
when you have no real reason to think it can regain that leadership.

- bill

John Smith wrote:
> Tom Linden wrote:
> 
>>On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk>
>>wrote:
>>
>>
>>>Also how long would it take to port Tandem NSK to PA-RISC ?
>>>How close was the port to Alpha to completion when the decision to
>>>move to IA64
>>>was made ?
>>
>>Maybe the question that should be asked, is how long would it take to
>>provide migration tools from NSK to VMS?  Or alternatively, to whom
>>can we sell Non-Stop business unit?
>>
>>I would give odds that HP will start shedding business units.
> 
> 
> 
> IBM would probably snap up NSK in a heartbeat.
> 
> --
> OpenVMS - The classics never go out of style.
> 
> 
Just curious, if they'd be so interested in it now, why didn't they 
'snap it up' prior to Compaq doing so?

On Fri, 11 Feb 2005 15:52:14 -0500, Dave Froble <davef@tsoft-inc.com>  
wrote:

> Tom Linden wrote:
>> On Thu, 10 Feb 2005 21:59:23 -0500, JF Mezei   
>> <jfmezei.spamnot@teksavvy.com> wrote:
>>
>>> Tom Linden wrote:
>>>
>>>>  The real
>>>> value in the Digital component of HP is VMS, Alpha has lost all its   
>>>> value.
>>>
>>>
>>> There is one aspect of Alpha: installed base and available software,
>>> much of which is no longer developped and thus not be available on
>>> Alpha's successor. Continuing the Alpha processor ratins that software
>>> whereas moving to a new platform requires VMS start from scratch with
>>> software availability.
>
> This is an argument for Alpha, and VAX.  It's a limited argument.
>
>>> Note that this is also a reason HP-UX customers are still buying PaRisc
>>> based machines instead of that unwanted IA64 thing. But VMS is in a  
>>> much
>>> worse shape than HP-UX in terms of application ecosystem.
>>   Face it Alpha is dead.  To make VMS a viable business it _will_ need
>> to use a supplier for its chips such as Power /and/or Opteron.
>
> Much as I hate to, I have to admit that Alpha is most likely history. To  
> what purpose should HP spend the considerable money it would take, and  
> for what, just another "me too".
>
> There are 2 issues, near term, and far future.  In the near term,  
> continuing to sell Alphas, with whatever enhancements can be done for  
> EV7, is one solution.  Using the itanic when it's a fit is also  
> reasonable.  For far term, I can see a costly and tremendous effort to  
> pull Alpha back to the front of the pack, just to see that new advances  
> have made all today's techniques obsolete.  All dressed up, with nowhere  
> to go.
>
The cost argument, BTW, also applies to Itanic, its ongoing development  
costs
can not be amortised over a large run, as had been envisaged and was indeed
part of the reason for embarking on the program, at least that is how I
interpreted it.


> Dave



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com> wrote:

> Tom Linden wrote:
>> On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk>
>> wrote:
>>
>>> Also how long would it take to port Tandem NSK to PA-RISC ?
>>> How close was the port to Alpha to completion when the decision to
>>> move to IA64
>>> was made ?
>>
>> Maybe the question that should be asked, is how long would it take to
>> provide migration tools from NSK to VMS?  Or alternatively, to whom
>> can we sell Non-Stop business unit?
>>
>> I would give odds that HP will start shedding business units.
>
>
> IBM would probably snap up NSK in a heartbeat.
Probably not for the product, but the customer.  BTW, what are their  
revenues?
>
> --
> OpenVMS - The classics never go out of style.
>
>



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

JF Mezei wrote:
> Dave Froble wrote:
> 
>>Much as I hate to, I have to admit that Alpha is most likely history.
>>To what purpose should HP spend the considerable money it would take,
>>and for what, just another "me too".
> 
> 
> Competitive advantage.
> 
> HP has learned that there is no money to be made in commodity. IBM knew
> that a long time ago and has already taken action.
> 
> So if HP is to succeed, will it turn itself into a low cost Dell
> competitor, or will it try to find some true technological edge to
> differentiate itself from the others ?
> 
> Alpha may have been asleep for a couple of years, but so has PaRisc.
> Nevertheless, resurecting Alpha would still be far easier and cheaper
> and less troublesome for customers than going to a totally new platform.

The itanic, whether you and I and anyone else likes it, is here and 
running VMS.

How long do you think it would take to put together a project to again 
work on Alpha?  Hours?  Days?  Weeks?  Months?  The last option listed 
is most likely.  That's before they start to do anything.  It's not just 
the group(s) working on 'real' stuff, DEC had plenty of research going 
on.  You may have noticed my commenting on 'EV10'?  That was pure 
research into an idea, not anything that would be a product.  Compaq did 
a real good job of slashing both R&D, with emphysis on the 'R'.

So, Ok, you do get both some R and some D going again.  How long before 
you get anything out of 'D'?  How long before 'D' even decides what it's 
going to be working on?  Don't even ask about timing for the 'R' part.

The real issue in all this is what's going to come out of all the 'R' 
some others were smart enough to keep doing?  A week ago many had no 
knowledge of what IBM, Sony, and Toshiba were up to.  It's too early to 
tell if they're on to something, but, some new technologies are going to 
emerge.  Blowing money on the last century's technology isn't a real 
good idea.

Dave

In article <5vWdnXURA-ko1ZDfRVn-qg@metrocastcablevision.com>,
	Bill Todd <billtodd@metrocast.net> writes:
> JF Mezei wrote:
>> Dave Froble wrote:
>> 
>>>Much as I hate to, I have to admit that Alpha is most likely history.
>>>To what purpose should HP spend the considerable money it would take,
>>>and for what, just another "me too".
>> 
>> 
>> Competitive advantage.
> 
> I've made repeated suggestions that the only way any competitive 
> advantage would accrue to a resurrected Alpha platform would be for it 
> to incorporate something superior to the competition, and so far no one 
> (including you) has come up with any such feature.

It does have a competitive advantage.  It can run VMS.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill@cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>   

Tom Linden wrote:
> On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com> wrote:

....

>> IBM would probably snap up NSK in a heartbeat.
> 
> Probably not for the product, but the customer.  BTW, what are their  
> revenues?

The same Compaq slide that stated VMS annual system revenues as $4 
billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2 
billion.  And in rather marked contrast to VMS's and Tru64's, NSK's were 
reported as holding approximately steady the following fall.

Which may say nothing about what they are now, of course.

- bill

Bill Todd wrote:
> Alpha *would have* led in SMT technology had EV8 been completed, but by
> the time any SMT Alpha could get out now that technology will be old hat

Yes. But the point is that a project started today would go beyond SMT,
just like other chip makers are already looking beyond SMT. There are
parts of EV7 and EV8 that are still usable, and there is experience and
inspiration from Power and Sparc and perhasp even the 8086 that can be
used to start a new Alpha today aleready ahead of where it was at the
time Compaq slowed and then stopped work on Alpha.

> > HP has learned that there is no money to be made in commodity.
> 
> You mean like in printers?

Printers are far from commodity. They are highly proprietary with much
done to prevent one from even refilling ink genuine cartridges. Heck HP
even now prevents ink from one continent from being used on printers
from other continents.

> The question, of course, is where to look for that edge, given limited
> resources.  And since any such edge absolutely *requires* a lot of
> system-level work, adding anything else that isn't absolutely necessary
> would dilute it (it's not as if HP was so healthy that it could throw
> cash around that it didn't really need to).

Consider the 3 billion they threw at Intel for no specific return. Seems
to me that HP has had plenty of ink money to throw at Carly's pet projects.


> Having HP go out of business trying to resurrect them would be even more
> troublesome for customers, I suspect.  HP should continue to sell the
> platforms as long as customers want them, but (one more time, absent
> some wonderful idea that would lead to a significantly superior product)
> should look elsewhere for their longer-term future.


When you look at HP's enterprise business, what is more important ? The
ecosystem of the installed base and exsiting software, or switching chip
in the hopes to that the new chip will be better ?


> Intel and HP wasted vast amounts of cash and resources developing a
> product with no clear advantage over its competition

Yes. But the problem with IA64 was its original mamdate which lead to
too radical a change as well as too may requirements which resulted in
IA64 being a failure because it tried to be everything to everyone.

Just because IA64 is a commercial failure doesn't mean that with a
leaner and meaner chip such as Alpha, HP couldn't be succesful.

> represents gross incompetence.  But so does trying to resurrect it later
> when you have no real reason to think it can regain that leadership.

So, the minute you agree that IA64 is a goner, what should HP replace it
with ? Is it only option the 8086 ? 

If Intel and AMD can scale the 8086 to match Power and Sparc's
enterprise capabilities, then it may work. But that may take more time
than ramping Alpha or PaRisc back up.

Bill Todd wrote:

....

> The same Compaq slide that stated VMS annual system revenues as $4 
> billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2 
> billion.  And in rather marked contrast to VMS's and Tru64's, NSK's were 
> reported as holding approximately steady the following fall.
> 
> Which may say nothing about what they are now, of course.
> 
> - bill

Whoops - I'm really not sure whether the above figure applied solely to 
NSK or to all Tandem-related revenue, which may still have included 
their 'Integrity' Unix servers back then (nor have I any idea what the 
latter may have amounted to).

- bill

Bill Gunshannon wrote:
> In article <5vWdnXURA-ko1ZDfRVn-qg@metrocastcablevision.com>,
> Bill Todd <billtodd@metrocast.net> writes:
>> JF Mezei wrote:
>>> Dave Froble wrote:
>>>
>>>> Much as I hate to, I have to admit that Alpha is most likely
>>>> history. To what purpose should HP spend the considerable money it
>>>> would take, and for what, just another "me too".
>>>
>>>
>>> Competitive advantage.
>>
>> I've made repeated suggestions that the only way any competitive
>> advantage would accrue to a resurrected Alpha platform would be for
>> it to incorporate something superior to the competition, and so far
>> no one (including you) has come up with any such feature.
>
> It does have a competitive advantage.  It can run VMS.
>
> bill


So what whomever winds up with VMS (not necessarily HP after the new CEO is
brought in) ought to do is port VMS to everything and then let SI's and
VAR's sell it on whatever they want and have a 1st level/2nd-level support
arrangement - SI/VAR handles 1st level and can sell a 2nd level support
contract direct with VMS Engineering.

--
OpenVMS - The classics never go out of style.

Dave Froble wrote:
> John Smith wrote:
>> Tom Linden wrote:
>>
>>> On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk>
>>> wrote:
>>>
>>>
>>>> Also how long would it take to port Tandem NSK to PA-RISC ?
>>>> How close was the port to Alpha to completion when the decision to
>>>> move to IA64
>>>> was made ?
>>>
>>> Maybe the question that should be asked, is how long would it take
>>> to provide migration tools from NSK to VMS?  Or alternatively, to
>>> whom can we sell Non-Stop business unit?
>>>
>>> I would give odds that HP will start shedding business units.
>>
>>
>>
>> IBM would probably snap up NSK in a heartbeat.
>>
>> --
>> OpenVMS - The classics never go out of style.
>>
>>
> Just curious, if they'd be so interested in it now, why didn't they
> 'snap it up' prior to Compaq doing so?


They would have had to buy the whole company.

Now at HP, the Board is having second/third/fourth thoughts about whether
they should just be a printing company. With that, everything else is
potentially for sale and probably piecemeal if they thought it would get
them more money.

--
OpenVMS - The classics never go out of style.

Bill Todd wrote:
> Tom Linden wrote:
>> On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com>
>> wrote:
>
> ...
>
>>> IBM would probably snap up NSK in a heartbeat.
>>
>> Probably not for the product, but the customer.  BTW, what are their
>> revenues?
>
> The same Compaq slide that stated VMS annual system revenues as $4
> billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2
> billion.  And in rather marked contrast to VMS's and Tru64's, NSK's
> were reported as holding approximately steady the following fall.
>
> Which may say nothing about what they are now, of course.
>
> - bill


NSK, while not a big line item in the HP sales forecast, has loyal
customers....not too many NSK installations are switching to Windows for
it's reliability, nor to Linux for its transaction processing. Those
platforms are not perceived to be threats to NSK, either in reality or in
perception.

HP has done nothing to position VMS in the same light even though in almost
every way it is as robust as NSK when properly configured, and under certain
circumstance more robust.

--
OpenVMS - The classics never go out of style.

On Fri, 11 Feb 2005 21:01:10 -0500, John Smith <a@nonymous.com> wrote:

> Bill Todd wrote:
>> Tom Linden wrote:
>>> On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com>
>>> wrote:
>>
>> ...
>>
>>>> IBM would probably snap up NSK in a heartbeat.
>>>
>>> Probably not for the product, but the customer.  BTW, what are their
>>> revenues?
>>
>> The same Compaq slide that stated VMS annual system revenues as $4
>> billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2
>> billion.  And in rather marked contrast to VMS's and Tru64's, NSK's
>> were reported as holding approximately steady the following fall.
>>
>> Which may say nothing about what they are now, of course.
>>
>> - bill
>
>
> NSK, while not a big line item in the HP sales forecast, has loyal
> customers....not too many NSK installations are switching to Windows for
> it's reliability, nor to Linux for its transaction processing. Those
> platforms are not perceived to be threats to NSK, either in reality or in
> perception.
Stratus has done a fault-tolerant version for Windows, so that may change.
But, how many 9's can you get from a VMS cluster?
>
> HP has done nothing to position VMS in the same light even though in  
> almost
> every way it is as robust as NSK when properly configured, and under  
> certain
> circumstance more robust.
>
> --
> OpenVMS - The classics never go out of style.
>
>



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

JF Mezei wrote:
> Dave Froble wrote:
>> Just curious, if they'd be so interested in it now, why didn't they
>> 'snap it up' prior to Compaq doing so?
>
> Bob Palmer started negotiations with Pfeiffer 3 years prior to the
> actual transaction. That would put it in a 1995/1996 time frame. IBM
> was
> just recovering from its near bankrupcy at that time and Gerstner's
> was
> just starting to see success in reshaping mentality and corporate
> culture at IBM.  Gerstner was not found of takeovers. Integrating
> Digital into IBM at the time would not have been good for IBM and IBM
> shareholders.
>
> IBM was later offered to buy Compaq and when Gerstner looked at the
> numbers, he immediatly dismissed the idea laughing at the investment
> bankers who were selling this as the next best thing since sliced
> bear.
> Such a transaction would have obliterated IBM,s profits with 50
> billion
> over 5 years and put IBM i serious loss situation.
>
> History books will call it the curse of VMS. Digital went down because
> of it. Compaq went down because it. HP is going down.

Wrong. DEC and Compaq when down because of what they didn't do with it -
create demand for it.

--
OpenVMS - The classics never go out of style.

Tom Linden wrote:

> The cost argument, BTW, also applies to Itanic, its ongoing development  
> costs
> can not be amortised over a large run, as had been envisaged and was indeed
> part of the reason for embarking on the program, at least that is how I
> interpreted it.

Yeah, but that's Intel's problem.  They wanted it.  Now they got it.  As 
long as they make then, why not use them?

Dave

Tom Linden wrote:
> On Fri, 11 Feb 2005 21:01:10 -0500, John Smith <a@nonymous.com> wrote:
>
>> Bill Todd wrote:
>>> Tom Linden wrote:
>>>> On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com>
>>>> wrote:
>>>
>>> ...
>>>
>>>>> IBM would probably snap up NSK in a heartbeat.
>>>>
>>>> Probably not for the product, but the customer.  BTW, what are
>>>> their revenues?
>>>
>>> The same Compaq slide that stated VMS annual system revenues as $4
>>> billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2
>>> billion.  And in rather marked contrast to VMS's and Tru64's, NSK's
>>> were reported as holding approximately steady the following fall.
>>>
>>> Which may say nothing about what they are now, of course.
>>>
>>> - bill
>>
>>
>> NSK, while not a big line item in the HP sales forecast, has loyal
>> customers....not too many NSK installations are switching to Windows
>> for it's reliability, nor to Linux for its transaction processing.
>> Those platforms are not perceived to be threats to NSK, either in
>> reality or in perception.
> Stratus has done a fault-tolerant version for Windows, so that may
> change. But, how many 9's can you get from a VMS cluster?


How many 9's do you want?
The military has 'Dial-a-Yield' for their nukes.

Just don't forget that the number of real 9's also depends on a lot of other
factors outside the buildings.


--
OpenVMS - The classics never go out of style.

Dave Froble wrote:
> Just curious, if they'd be so interested in it now, why didn't they
> 'snap it up' prior to Compaq doing so?

Bob Palmer started negotiations with Pfeiffer 3 years prior to the
actual transaction. That would put it in a 1995/1996 time frame. IBM was
just recovering from its near bankrupcy at that time and Gerstner's was
just starting to see success in reshaping mentality and corporate
culture at IBM.  Gerstner was not found of takeovers. Integrating
Digital into IBM at the time would not have been good for IBM and IBM shareholders.

IBM was later offered to buy Compaq and when Gerstner looked at the
numbers, he immediatly dismissed the idea laughing at the investment
bankers who were selling this as the next best thing since sliced bear.
Such a transaction would have obliterated IBM,s profits with 50 billion
over 5 years and put IBM i serious loss situation.

History books will call it the curse of VMS. Digital went down because
of it. Compaq went down because it. HP is going down.

John Smith wrote:
> Dave Froble wrote:
> 
>>John Smith wrote:
>>
>>>Tom Linden wrote:
>>>
>>>
>>>>On Fri, 11 Feb 2005 17:17:34 +0000 (UTC), <david20@alpha2.mdx.ac.uk>
>>>>wrote:
>>>>
>>>>
>>>>
>>>>>Also how long would it take to port Tandem NSK to PA-RISC ?
>>>>>How close was the port to Alpha to completion when the decision to
>>>>>move to IA64
>>>>>was made ?
>>>>
>>>>Maybe the question that should be asked, is how long would it take
>>>>to provide migration tools from NSK to VMS?  Or alternatively, to
>>>>whom can we sell Non-Stop business unit?
>>>>
>>>>I would give odds that HP will start shedding business units.
>>>
>>>
>>>
>>>IBM would probably snap up NSK in a heartbeat.
>>>
>>>--
>>>OpenVMS - The classics never go out of style.
>>>
>>>
>>
>>Just curious, if they'd be so interested in it now, why didn't they
>>'snap it up' prior to Compaq doing so?
> 
> 
> 
> They would have had to buy the whole company.

I'm assuming that Compaq did buy the whole company.  What part of "prior 
to Compaq doing so" didn't you understand?

> Now at HP, the Board is having second/third/fourth thoughts about whether
> they should just be a printing company. With that, everything else is
> potentially for sale and probably piecemeal if they thought it would get
> them more money.
> 
> --
> OpenVMS - The classics never go out of style.
> 
> 

Dave Froble wrote:
 
> Yeah, but that's Intel's problem.  They wanted it.  Now they got it.  As
> long as they make then, why not use them?

It takes a few years to port to a new platform. If Intel has begn to
scale down IA64, reduce its target market niche, and you know that
revenus are not impressive and rumour tha the only reason Intel is
keepiong IA64 alive is contractual obligatiosn with HP, one can
reasonably guess that as soon as the contract ends, so does IA64.

So, when rumour sof IA64's demise become credible, you need to start
looking at alternatives.

Dave Froble wrote:
> John Smith wrote:
>> Dave Froble wrote:
>>
>>> John Smith wrote:
>>>
>>>> Tom Linden wrote:
>>>>
>>>>
>>>>> On Fri, 11 Feb 2005 17:17:34 +0000 (UTC),
>>>>> <david20@alpha2.mdx.ac.uk> wrote:
>>>>>
>>>>>
>>>>>
>>>>>> Also how long would it take to port Tandem NSK to PA-RISC ?
>>>>>> How close was the port to Alpha to completion when the decision
>>>>>> to move to IA64
>>>>>> was made ?
>>>>>
>>>>> Maybe the question that should be asked, is how long would it take
>>>>> to provide migration tools from NSK to VMS?  Or alternatively, to
>>>>> whom can we sell Non-Stop business unit?
>>>>>
>>>>> I would give odds that HP will start shedding business units.
>>>>
>>>>
>>>>
>>>> IBM would probably snap up NSK in a heartbeat.
>>>>
>>>> --
>>>> OpenVMS - The classics never go out of style.
>>>>
>>>>
>>>
>>> Just curious, if they'd be so interested in it now, why didn't they
>>> 'snap it up' prior to Compaq doing so?
>>
>>
>>
>> They would have had to buy the whole company.
>
> I'm assuming that Compaq did buy the whole company.  What part of
> "prior to Compaq doing so" didn't you understand?


I mis-read your post because I was thinking about Gerstner declining a shot
a Compaq as a whole.

As Tom Linden said, I think IBM would take Tandem because it would then
pretty much give IBM the entire datacenter at many banks and insurance
companies -- Z/OS mainframe, AIX, NSK, and the licence to print money
without competition therein. Plus some other goodies like Atalla.
--

OpenVMS - The classics never go out of style.

John Smith wrote:
> As Tom Linden said, I think IBM would take Tandem because it would then
> pretty much give IBM the entire datacenter at many banks and insurance
> companies -- Z/OS mainframe, AIX, NSK, and the licence to print money
> without competition therein. Plus some other goodies like Atalla.

Tandem also brings with it lots of bragging rights. Non-stop fault
tolerant, as well as the glamour of having the stock exchanges and some
telecom applications. But it is a small business in the grand scheme of things.

JF Mezei wrote:
> Dave Froble wrote:
> 
>>Just curious, if they'd be so interested in it now, why didn't they
>>'snap it up' prior to Compaq doing so?
> 
> 
> Bob Palmer started negotiations with Pfeiffer 3 years prior to the
> actual transaction.

Ah ....   The topic was Tandem, not DEC.

JF Mezei wrote:
> Dave Froble wrote:
>  
> 
>>Yeah, but that's Intel's problem.  They wanted it.  Now they got it.  As
>>long as they make then, why not use them?
> 
> 
> It takes a few years to port to a new platform. If Intel has begn to
> scale down IA64, reduce its target market niche, and you know that
> revenus are not impressive and rumour tha the only reason Intel is
> keepiong IA64 alive is contractual obligatiosn with HP, one can
> reasonably guess that as soon as the contract ends, so does IA64.
> 
> So, when rumour sof IA64's demise become credible, you need to start
> looking at alternatives.

I don't buy those arguments.

If Intel wanted to discontinue IA-64, then they would.  Contracts be 
damned.  Don't forget, as Rob and others were always telling us, Intel 
has plenty of cash.  Just how much would they have to wave under HP's 
nose to do away with any contracts, should such exist.

Intel is continuing IA-64 for the same reasons as before, EGO!  They 
want to play with the big boys.  Sure, it's fine to have the volume of 
the desktop, but the prestige is at the top, and they covet that.  What 
they should have done is take over Alpha, have their low volume high 
cost, and high performance, CPU, and then dance with AMD in the x86 
space.  If they would have done that, they wouldn't have had to worry 
about undercutting IA-64 with x86-64.

Hey, we still live in interesting times, but for now I think that if 
IA-64 was going to be a casuality, it would have already happened.  They 
are now working hard to minimize AMD, and IA-64 survived the panic.

Dave Froble wrote:
> If Intel wanted to discontinue IA-64, then they would.  Contracts be
> damned.  Don't forget, as Rob and others were always telling us, Intel
> has plenty of cash.  Just how much would they have to wave under HP's
> nose to do away with any contracts, should such exist.

You forgot:
Andy Grove is gone, Ottelini is in. Much easier for him to correct
mistakes of the past.

Margins have narrowed, and Intel has indicated it would look carefully
at less profitable products.

> Intel is continuing IA-64 for the same reasons as before, EGO!

EGO is no longer an issue. They changed CEO. And Intel is smart enough
to know you just can't abruptly cabncel a chip until you have a
replacement. For now, they continue with the IA64 plan of record until
the 8086 gets the features that make IA64 unnecessary. 2007 is the key
year here.

Intel learned from the Alpha murder debacle. They aren't going to
formally announce it end of line until it know it has a working replacement.

On Fri, 11 Feb 2005 22:15:09 -0500, John Smith <a@nonymous.com> wrote:

> Tom Linden wrote:
>> On Fri, 11 Feb 2005 21:01:10 -0500, John Smith <a@nonymous.com> wrote:
>>
>>> Bill Todd wrote:
>>>> Tom Linden wrote:
>>>>> On Fri, 11 Feb 2005 17:12:40 -0500, John Smith <a@nonymous.com>
>>>>> wrote:
>>>>
>>>> ...
>>>>
>>>>>> IBM would probably snap up NSK in a heartbeat.
>>>>>
>>>>> Probably not for the product, but the customer.  BTW, what are
>>>>> their revenues?
>>>>
>>>> The same Compaq slide that stated VMS annual system revenues as $4
>>>> billion and Tru64's as $3 billion in March, 2001, pegged NSK's at $2
>>>> billion.  And in rather marked contrast to VMS's and Tru64's, NSK's
>>>> were reported as holding approximately steady the following fall.
>>>>
>>>> Which may say nothing about what they are now, of course.
>>>>
>>>> - bill
>>>
>>>
>>> NSK, while not a big line item in the HP sales forecast, has loyal
>>> customers....not too many NSK installations are switching to Windows
>>> for it's reliability, nor to Linux for its transaction processing.
>>> Those platforms are not perceived to be threats to NSK, either in
>>> reality or in perception.
>> Stratus has done a fault-tolerant version for Windows, so that may
>> change. But, how many 9's can you get from a VMS cluster?
>
>
> How many 9's do you want?
> The military has 'Dial-a-Yield' for their nukes.
>
> Just don't forget that the number of real 9's also depends on a lot of  
> other
> factors outside the buildings.

Understood, but I think you get my point.
>
>
> --
> OpenVMS - The classics never go out of style.
>
>



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

On Fri, 11 Feb 2005 22:40:08 -0500, John Smith <a@nonymous.com> wrote:

> Dave Froble wrote:
>> John Smith wrote:
>>> Dave Froble wrote:
>>>
>>>> John Smith wrote:
>>>>
>>>>> Tom Linden wrote:
>>>>>
>>>>>
>>>>>> On Fri, 11 Feb 2005 17:17:34 +0000 (UTC),
>>>>>> <david20@alpha2.mdx.ac.uk> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Also how long would it take to port Tandem NSK to PA-RISC ?
>>>>>>> How close was the port to Alpha to completion when the decision
>>>>>>> to move to IA64
>>>>>>> was made ?
>>>>>>
>>>>>> Maybe the question that should be asked, is how long would it take
>>>>>> to provide migration tools from NSK to VMS?  Or alternatively, to
>>>>>> whom can we sell Non-Stop business unit?
>>>>>>
>>>>>> I would give odds that HP will start shedding business units.
>>>>>
>>>>>
>>>>>
>>>>> IBM would probably snap up NSK in a heartbeat.
>>>>>
>>>>> --
>>>>> OpenVMS - The classics never go out of style.
>>>>>
>>>>>
>>>>
>>>> Just curious, if they'd be so interested in it now, why didn't they
>>>> 'snap it up' prior to Compaq doing so?
>>>
>>>
>>>
>>> They would have had to buy the whole company.
>>
>> I'm assuming that Compaq did buy the whole company.  What part of
>> "prior to Compaq doing so" didn't you understand?
>
>
> I mis-read your post because I was thinking about Gerstner declining a  
> shot
> a Compaq as a whole.
>
> As Tom Linden said, I think IBM would take Tandem because it would then
> pretty much give IBM the entire datacenter at many banks and insurance
> companies -- Z/OS mainframe, AIX, NSK, and the licence to print money
> without competition therein. Plus some other goodies like Atalla.

That's not quite what I said.  I said that if IBM were to buy Tandem it  
would
be for the customer base, not the product or technology.  Remember back in  
85
IBM got very cozy with Stratus, but they never bought them, and stopped  
selling
the product after some years, even after very impressive wins, like Visa.
> --
>
> OpenVMS - The classics never go out of style.
>
>



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

"John Smith" <a@nonymous.com> writes:

> Dave Froble wrote:

>> I'm assuming that Compaq did buy the whole company.  What part of
>> "prior to Compaq doing so" didn't you understand?

> I mis-read your post because I was thinking about Gerstner declining
> a shot a Compaq as a whole.

> As Tom Linden said, I think IBM would take Tandem because it would
> then pretty much give IBM the entire datacenter at many banks and
> insurance companies -- Z/OS mainframe, AIX, NSK, and the licence to
> print money without competition therein. Plus some other goodies
> like Atalla.

IBM had already eaten Stratus, so grabbing Tandem as well would probably
have woken up the FTC, and it would have got several big customers
very unhappy.

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

John Reagan <john.reagan@hp.com> writes:

>> 2.  Port the OSs somewhere else.  This might be fairly easy for VMS
>> and even Tru64, though perhaps harder for HP-UX (since Itanic was
>> designed with its needs in mind).

> "fairly easy for VMS"?  Can I take a hit of what you are smoking?
> The large amounts of BLISS and especially Macro-32 code in OpenVMS
> make any port more difficult than a system like Tru64 with is mostly
> in C.  We have more compilers to find/build.  Happen to have a AMD64
> BLISS compiler sitting around?

You do in fact. Well, OK it needs an update to the code generator to
do the 64-bit ISNs.

> And to the best of my knowledge, I don't know of anything in Itanium
> that was specially developed for HP-UX.  The instruction set is pretty
> unbiased.

The itanic was designed to be the follow on to the 32 bit PAs. Slippage
resulted in PA3 as we know it now shipping.

> Perhaps some of the memory management support... but if OpenVMS
> could retrofit our VAX/Alpha designs to Itanium, I'd suspect that
> HP-UX could handle that transition as well.

> HP-UX's big-endian design might be a better fit for POWER5 than for
> an X86 platform.

The ISVs an the change to little endian are the big humps. And
the PC braindeath of course.

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

prep@prep.synonet.com wrote:

> IBM had already eaten Stratus, so grabbing Tandem as well would probably
> have woken up the FTC, and it would have got several big customers
> very unhappy.

Not entirely sure of that. Tandem may be considered too mission critical
for the USA to be allowed to fall into unreliable hands. HP's turmoils
and Wintel centricity may be convincing enough that only IBM is serous
enough to be given the responsability to maintain Tandem support/development.

In article <110qlbjact654b5@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes:
>JF Mezei wrote:
>> Dave Froble wrote:
>> 
>>>Much as I hate to, I have to admit that Alpha is most likely history.
>>>To what purpose should HP spend the considerable money it would take,
>>>and for what, just another "me too".
>> 
>> 
>> Competitive advantage.
>> 
>> HP has learned that there is no money to be made in commodity. IBM knew
>> that a long time ago and has already taken action.
>> 
>> So if HP is to succeed, will it turn itself into a low cost Dell
>> competitor, or will it try to find some true technological edge to
>> differentiate itself from the others ?
>> 
>> Alpha may have been asleep for a couple of years, but so has PaRisc.
>> Nevertheless, resurecting Alpha would still be far easier and cheaper
>> and less troublesome for customers than going to a totally new platform.
>
>The itanic, whether you and I and anyone else likes it, is here and 
>running VMS.
>

But what is the long term future for IA64 (assuming Intel and HP don't drop it)
?

Will it have to be redesigned to become more of a RISC chip and drop EPIC ?
If it keeps EPIC will the numbers sold justify the work required on the
compilers ?

David Webb
Security team leader
CCSS
Middlesex University




>How long do you think it would take to put together a project to again 
>work on Alpha?  Hours?  Days?  Weeks?  Months?  The last option listed 
>is most likely.  That's before they start to do anything.  It's not just 
>the group(s) working on 'real' stuff, DEC had plenty of research going 
>on.  You may have noticed my commenting on 'EV10'?  That was pure 
>research into an idea, not anything that would be a product.  Compaq did 
>a real good job of slashing both R&D, with emphysis on the 'R'.
>
>So, Ok, you do get both some R and some D going again.  How long before 
>you get anything out of 'D'?  How long before 'D' even decides what it's 
>going to be working on?  Don't even ask about timing for the 'R' part.
>
>The real issue in all this is what's going to come out of all the 'R' 
>some others were smart enough to keep doing?  A week ago many had no 
>knowledge of what IBM, Sony, and Toshiba were up to.  It's too early to 
>tell if they're on to something, but, some new technologies are going to 
>emerge.  Blowing money on the last century's technology isn't a real 
>good idea.
>
>Dave

david20@alpha2.mdx.ac.uk wrote:

> But what is the long term future for IA64 (assuming Intel and HP don't drop it)
> ?
> 
> Will it have to be redesigned to become more of a RISC chip and drop EPIC ?
> If it keeps EPIC will the numbers sold justify the work required on the
> compilers ?
> 
> David Webb
> Security team leader
> CCSS
> Middlesex University

IA-64 is now solely in the hands of Intel.  What will they do?  Who 
knows, which is probably one of the biggest reasons DEC/Compaq/HP should 
not have killed Alpha.  Alpha was in their hands, and they were not at 
risk (or is that RISC) of anyone else pulling the rug out from under them.

The itanic is here now, and from most reports it works reasonably well. 
  The keys here are, #1, for how long will it be here, and #2, is it a 
satisfactory CPU.

For #1, that's up to Intel.  It's out of HP's hands.  To counter this, 
they could continue to have IBM fab the EV7/EV7z and possibly some minor 
mods, process shrinks, and such indefinitely, or as long as customers 
are buying them.  Sell VMS on both platforms.  Keep their options open. 
  Hey, IA-64 could still catch on and be profitable for Intel.

As for #2, most VMS users aren't using VMS because it's blazingly faster 
than anything else available.  If this was ever true, it hasn't been for 
a long time, and there are still VMS users.  There are many reasons 
users stay with VMS, and this is a good market for HP.  So, even if the 
itanic never competes with Power and whatever, it may be adequate for 
VMS users.  Alpha EV7 will probably be adequate for most VMS users for a 
long time.  I'm talking 5-10 years.  The systems can get better even if 
using the same CPU.

There are at least 2 inexpensive methods for VMS to remain a viable OS 
in the near future.  The question is whether HP choses to implement and 
maintain either or both of the methods.

Dave

david20@alpha2.mdx.ac.uk wrote:
> Will it have to be redesigned to become more of a RISC chip and drop EPIC ?
> If it keeps EPIC will the numbers sold justify the work required on the
> compilers ?

I think that the original argument was that it was cheaper to write
compilers than to design chips. Shift intelligence from the chip to the
compiler and you should, in theory, have cheaper chip to develop.

The thing is that the compiler ecosystem on IA64 seems extremely
confusing. How many C compilers will exist on IA64 once NSK runs on it ?
 Whose compilers does SGI use on IA64-Linux ? How well tuned is the GNU
compiler ? Does it use an Intel back end to do all the EPIC stuff, or
does it have its own ? 

Heck, even on VMS, it seems hard to figure our exactly where the IA64
compilers are coming from.

My take is that if Intel can justify sinking more money into Itanic, it
will add new risc-like features but not remove the original EPIC core.
They'll find areas where trhe compilers can't do as good a job as Alpha,
and they'll tweak the chip to reduce this disadvantage.

My guess is that it won't get that far.

In article <110v9epootsmpfb@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes:

> The itanic is here now, and from most reports it works reasonably well. 
>   The keys here are, #1, for how long will it be here, and #2, is it a 
> satisfactory CPU.
> 
> For #1, that's up to Intel.  It's out of HP's hands.  To counter this, 
> they could continue to have IBM fab the EV7/EV7z and possibly some minor 
> mods, process shrinks, and such indefinitely, or as long as customers 
> are buying them.  Sell VMS on both platforms.  Keep their options open. 

They will do that if customer keep buying Alpha in sufficient quantity.

Larry Kilgallen wrote:
> In article <110v9epootsmpfb@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes:
> 
> 
>>The itanic is here now, and from most reports it works reasonably well. 
>>  The keys here are, #1, for how long will it be here, and #2, is it a 
>>satisfactory CPU.
>>
>>For #1, that's up to Intel.  It's out of HP's hands.  To counter this, 
>>they could continue to have IBM fab the EV7/EV7z and possibly some minor 
>>mods, process shrinks, and such indefinitely, or as long as customers 
>>are buying them.  Sell VMS on both platforms.  Keep their options open. 
> 
> 
> They will do that if customer keep buying Alpha in sufficient quantity.

Unfortunately, there are almost certainly nowhere nearly enough 
customers so dependent on the Alpha platform that they would commit even 
further to it in sufficient numbers absent *any* indication of interest 
(let alone real 'commitment') in it by its owner.  So the ball is, as it 
has always been, in said owner's court, and there's little evidence that 
I'm aware of that said owner will do anything but continue to ignore it.

- bill

Larry Kilgallen wrote:
> They will do that if customer keep buying Alpha in sufficient quantity.

I am not so sure anymore.

The road map has had slides changed from "at least until 2006" to "until 2006".

If HP really does intent to justify the IA64 decision, they will
cannabalise Alpha and PaRisc and force customes onto IA64 to raise IA64
sales. They are perfectly capable of making such a move. (big mistake,
of course).

If the real strategy is to find a way out of IA64, the yes, they will
continue both PaRisc and Alpha according to customer demand. 2004 sent
the stage for this with plenty of hints that IA64 wasn't viable in long
term. 2005 will be an interesting year, especially in light of both
Intel and HP getting new CEOs.

Larry Kilgallen wrote:
> In article <110v9epootsmpfb@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes:
> 
> 
>>The itanic is here now, and from most reports it works reasonably well. 
>>  The keys here are, #1, for how long will it be here, and #2, is it a 
>>satisfactory CPU.
>>
>>For #1, that's up to Intel.  It's out of HP's hands.  To counter this, 
>>they could continue to have IBM fab the EV7/EV7z and possibly some minor 
>>mods, process shrinks, and such indefinitely, or as long as customers 
>>are buying them.  Sell VMS on both platforms.  Keep their options open. 
> 
> 
> They will do that if customer keep buying Alpha in sufficient quantity.

I'll wait to see the attitude of the next HP CEO before I totally buy 
into that.  Carly's "burnt our bridges behind us" attitude did not let 
me feel that continuing sales of Alpha would extend it's life.

I hope you are right.

Dave

Dave Froble wrote:
> I'll wait to see the attitude of the next HP CEO before I totally buy
> into that.  Carly's "burnt our bridges behind us" attitude did not let
> me feel that continuing sales of Alpha would extend it's life.


Somewhere inside HP is someone who really knows the true feelings of
paying customers towards that IA64 thing. Something not filtered by
marketing. 

A good CEO will find those persons and get the real story about how
custoemrs really feel. If he/she can't find it, he/she will hire and
INDEPENDANT organisation to poll customers and NON customers and
especially FORMER customers. That is what Gerstner did at IBM when he
realised that all of the customer data inside IBM was slanted by people
who wanted to provide good news in order to get ahead in their carreers.

JF Mezei wrote:
> Dave Froble wrote:
>> I'll wait to see the attitude of the next HP CEO before I totally buy
>> into that.  Carly's "burnt our bridges behind us" attitude did not
>> let me feel that continuing sales of Alpha would extend it's life.
>
>
> Somewhere inside HP is someone who really knows the true feelings of
> paying customers towards that IA64 thing. Something not filtered by
> marketing.
>
> A good CEO will find those persons and get the real story about how
> custoemrs really feel. If he/she can't find it, he/she will hire and
> INDEPENDANT organisation to poll customers and NON customers and
> especially FORMER customers. That is what Gerstner did at IBM when he
> realised that all of the customer data inside IBM was slanted by
> people who wanted to provide good news in order to get ahead in their
> carreers.


Can anyone here really define how the $3B that HP said they have committed
to Itanic over the next few years (Jan. 19th or thereabouts announcement)
will actually be spent?

Say HP decides to get out of the IA64 game - could they take $1B of that
money, relaunch Alpha - speed bumps for the short-term and new real
engineering effort (EV8, etc..) for the mid-term (next 8 years), with
research into something better beyond that?

In other words, to re-examine some of Bill's prior arguments, could Alpha -
sufficiently funded today - be a better bet than Itanic on a performance
level? It would also isolate HP from capricious action by Intel.

If Intel decides to axe Itanic, HP's business is flushed down the tubes -
that would be how many times each of the enterprise customer base would have
had to do a migration in how many years?

--
OpenVMS - The classics never go out of style.

John Smith wrote:
> JF Mezei wrote:
> 
>>Dave Froble wrote:
>>
>>>I'll wait to see the attitude of the next HP CEO before I totally buy
>>>into that.  Carly's "burnt our bridges behind us" attitude did not
>>>let me feel that continuing sales of Alpha would extend it's life.
>>
>>
>>Somewhere inside HP is someone who really knows the true feelings of
>>paying customers towards that IA64 thing. Something not filtered by
>>marketing.
>>
>>A good CEO will find those persons and get the real story about how
>>custoemrs really feel. If he/she can't find it, he/she will hire and
>>INDEPENDANT organisation to poll customers and NON customers and
>>especially FORMER customers. That is what Gerstner did at IBM when he
>>realised that all of the customer data inside IBM was slanted by
>>people who wanted to provide good news in order to get ahead in their
>>carreers.
> 
> 
> 
> Can anyone here really define how the $3B that HP said they have committed
> to Itanic over the next few years (Jan. 19th or thereabouts announcement)
> will actually be spent?
> 
> Say HP decides to get out of the IA64 game - could they take $1B of that
> money, relaunch Alpha - speed bumps for the short-term and new real
> engineering effort (EV8, etc..) for the mid-term (next 8 years), with
> research into something better beyond that?
> 
> In other words, to re-examine some of Bill's prior arguments, could Alpha -
> sufficiently funded today - be a better bet than Itanic on a performance
> level? It would also isolate HP from capricious action by Intel.
> 
> If Intel decides to axe Itanic, HP's business is flushed down the tubes -
> that would be how many times each of the enterprise customer base would have
> had to do a migration in how many years?

Another side of this issue is that every Alpha sale in preference to 
itanic is another reason for Intel to give up on the POS.  So, 
continuing to sell Alphas might be seen as a bad thing to someone who 
wants to sell as many itanics as possible.

Dave

John Smith wrote:

> If Intel decides to axe Itanic, HP's business is flushed down the tubes -
> that would be how many times each of the enterprise customer base would have
> had to do a migration in how many years?

At this point in time, so few commercial customers have actually
migrated all their business to IA64 things that killing IA64 today would
be welcome by most customers since it would remove the cumbersome burden
and expenses of that migration.

If they're going to spend the money to migrate to another platform, they
want one that is viable, that doesn't have rumours of its impending
doom, and whose vendor is trying to expand market instead of narrowing
its market niche every second month.

John Smith wrote:

....

> In other words, to re-examine some of Bill's prior arguments, could Alpha -
> sufficiently funded today - be a better bet than Itanic on a performance
> level?

Not unless Intel stops aggressively developing Itanic, or unless someone 
comes up with some great new secret performance feature for Alpha that 
Itanic for some reason won't also benefit from (and even if such a 
feature existed it couldn't really be expected to ship before 2010-11).

Alpha design stopped almost 4 years ago, so there's no way Alpha could 
catch up now without a ground-up redesign - i.e., the stuff in the pipe 
4 years ago would have been great to have had starting last year, but 3 
years from now - even in the then-current processes - will be about as 
exciting in terms of performance as UltraSPARC was a year ago (the new 
dual-core SPARCs appear to be doing considerably better now, by the way, 
and SPARC64 even more so).

> It would also isolate HP from capricious action by Intel.

Wouldn't be much help for HP-UX, whose customers would likely appreciate 
a forced migration to a trailing-edge Alpha platform about as much as 
VMS customers would have appreciated a forced migration to SPARC in 2003.

For HP to think that resurrecting Alpha was worthwhile, it would not 
only have to be *very* worried about Itanic having any future at all (at 
worst, HP could likely pay Intel to maintain at least modest development 
efforts that would keep it viable until the end of the decade) but also 
have to value VMS and Tru64 and their customer bases.  Do you really see 
any indication that it does?

> 
> If Intel decides to axe Itanic, HP's business is flushed down the tubes -
> that would be how many times each of the enterprise customer base would have
> had to do a migration in how many years?

I'm not all that sure that HP cares very much about the enterprise 
customer base any more anyway.  Its actions certainly seem to be 
assuming that maintaining significant R&D there isn't a priority.

As I said not long ago, Carly and Curly may have destroyed most of HP's 
enterprise product lines beyond recovery (HP-UX still appears to be 
viable, but certainly doesn't look enough like a core part of HP's 
future any more that one should consider it safe).  With no indication 
that HP even takes VMS seriously, getting it to consider reviving Alpha 
seems like the longest of long shots.

- bill

Bill Todd wrote:
> Not unless Intel stops aggressively developing Itanic, or unless someone
> comes up with some great new secret performance feature for Alpha

Does Montecito offer a totally different core, or just packages two
cores on one chip with necessary support infrastructure around the 2
cores ?

Does it have any built-in hardware to spread the load betwene the 2
cores, or is that left entirely to compilers ?

Beyond Montecito, what does Intel have planned for IA64 in terms of new
cores ? Isn't it just speed bumps, more cache for the next couple of
years ?

has Intel even announced intentions to have a single coherant and shared
cache between cores ?

BTW, AMD announced fairly substantial price reductions on its 64 bit
8086 chips yesterday. That will put more pressure on Intel to become
more efficient and cut unprofitable product lines, since its cash cow,
the 8086 will have to lower its margins to compete against AMD.

JF Mezei wrote:
> Bill Todd wrote:
> 
>>Not unless Intel stops aggressively developing Itanic, or unless someone
>>comes up with some great new secret performance feature for Alpha
> 
> 
> Does Montecito offer a totally different core,

No.

  or just packages two
> cores on one chip with necessary support infrastructure around the 2
> cores ?

No.

While the general design of the Montecito core dates right back to 
McKinley, there have been some significant new technologies applied to 
it.  First, performance per Watt has been increased by something close 
to a factor of 3 over Madison by selectively shutting down inactive 
parts of the chip on a nanosecond-by-nanosecond basis and other tweaks 
I'm less familiar with - the net effect being that a dual-core Montecito 
chip should draw a bit less power than a 1.5 GHz Madison 6M (and 
significantly less than a 1.6 GHz Madison 9M) while each core 
significantly out-performs a Madison 9M.  Second, the chip responds 
dynamically to workload and adjusts the clock rate to keep the power 
consumption near the maximum (unless, of course, the chip is idle) - 
e.g., a nominal 2 GHz Montecito should run very compute-intensive loads 
at that rate but is expected to run at about 2.2 GHz with a less 
compute-intensive TPC-C workload.  Third, each core supports two threads 
(in a somewhat more primitive manner than fine-grained SMT, but it 
should provide a noticeable boost to throughput in multi-threaded 
workloads).  There are other less significant tweaks as well.

In sum, Montecito should be the first Itanic that will be really 
respectable:  it should overtake POWER5 in smaller system configurations 
even on commercial workloads (though POWER5+ will be out by then and 
should maintain a lead, and I see no reason to expect Itanic's scaling 
problems to go away so POWER should retain a *significant* lead in 
larger systems), and will finally be competitive in terms of performance 
per Watt (quite possibly better than EV8 would have been in the same 
process, though well behind a 90 nm. EV8's performance core-for-core in 
a commercial workload and, of course, without EV8's excellent scaling 
characteristics).

> 
> Does it have any built-in hardware to spread the load betwene the 2
> cores, or is that left entirely to compilers ?

It's a conventional MP with 2 threads per core.  Do you know of *any* 
architecture which supports some kind of hardware-automated load-spreading?

> 
> Beyond Montecito, what does Intel have planned for IA64 in terms of new
> cores ? Isn't it just speed bumps, more cache for the next couple of
> years ?

Since the Alpha team's core got axed for Tukwila, my impression is that 
there's no new Itanic core publicly planned:  Tukwila in 2007 will have 
another spin on the current core, plus EV7-style on-chip support for 
memory and routing (which may partially or completely eliminate the 
scaling problems which Itanic has had compared with POWER, EV7, and even 
SPARC).

> 
> has Intel even announced intentions to have a single coherant and shared
> cache between cores ?

Not that I know of, but that's really not very important.

> 
> BTW, AMD announced fairly substantial price reductions on its 64 bit
> 8086 chips yesterday. That will put more pressure on Intel to become
> more efficient and cut unprofitable product lines, since its cash cow,
> the 8086 will have to lower its margins to compete against AMD.

The situation with Itanic right now has some similarities to that of 
Alpha at the time of the Alphacide.  The next generation (Montecito) is 
getting ready to ship and will offer significant improvements over the 
current one, and the generation after that (Tukwila) is well on the way 
to completion and promises another very significant round of improvements.

So Itanic has the next several years pretty well covered without needing 
all that much more development effort (and cash), just as was true for 
Alpha in 2001 and why killing EV8 was such a false 'economy'.  Intel has 
little reason to kill a product which is - finally - becoming something 
which people might actually want to buy.  It may well *never* get back 
the cash (let alone the opportunity cost) which it has sunk into it, but 
that's irrelevant to continuing to produce and develop it now:  as long 
as it can generate positive cash flow (which won't happen if it's 
visibly being abandoned), it will make sense to continue it, and Intel 
can afford to take a wait-and-see attitude toward plowing new mountains 
of cash into the kind of frenetic (one might almost say desperate) 
triple-path development efforts which have characterized Itanic until now.

In other words, by all appearances Itanic will overall apparently have 
been a lousy investment but the results aren't worthless - sort of like 
purchasing a piece of land in the expectation that its value would 
triple and eventually finding that it will never be worth as much as you 
paid for it (and will in fact be even more of a loss after you consider 
years' worth of taxes on it), but is still definitely worth *something*. 
  And the decision to kill Alpha was beyond question incredibly stupid 
for many reasons, but after all the pain at least its designated 
successor looks as if it may finally become reasonably usable.

HP and Itanic may be joined at the hip.  If HP goes under (or at least 
ceases to be a player in the enterprise space), Intel may not have 
enough of an Itanic market to break even.  Intel is pretty much 
dependent upon others to make or break Itanic in the marketplace - and 
by this point may be sufficiently disgusted that it doesn't care all 
that much which happens but will just try to be prepared for either 
eventuality.  Then again, it's not all that clear that HP cares too much 
what happens in that space either:  I certainly don't see enough 
long-term investing there to be convincing.

Pretty ironic given how intensely committed both companies seemed just 
months ago:  finally getting over the hump, and no energy left to move on.

- bill

Bill Todd <billtodd@metrocast.net> writes:

> HP and Itanic may be joined at the hip.  If HP goes under (or at
> least ceases to be a player in the enterprise space), Intel may not
> have enough of an Itanic market to break even.  Intel is pretty much
> dependent upon others to make or break Itanic in the marketplace -
> and by this point may be sufficiently disgusted that it doesn't care
> all that much which happens but will just try to be prepared for
> either eventuality.  Then again, it's not all that clear that HP
> cares too much what happens in that space either: I certainly don't
> see enough long-term investing there to be convincing.

If hp drops non-printer stuff, the could they hand the top end
stuff over to intel? They already have half the compilers, so
add VMS, T64's corpse, hpux and the storage stuff. Now that would
make bill's day :)

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

prep@prep.synonet.com wrote:
> Bill Todd <billtodd@metrocast.net> writes:
>
>> HP and Itanic may be joined at the hip.  If HP goes under (or at
>> least ceases to be a player in the enterprise space), Intel may not
>> have enough of an Itanic market to break even.  Intel is pretty much
>> dependent upon others to make or break Itanic in the marketplace -
>> and by this point may be sufficiently disgusted that it doesn't care
>> all that much which happens but will just try to be prepared for
>> either eventuality.  Then again, it's not all that clear that HP
>> cares too much what happens in that space either: I certainly don't
>> see enough long-term investing there to be convincing.
>
> If hp drops non-printer stuff, the could they hand the top end
> stuff over to intel? They already have half the compilers, so
> add VMS, T64's corpse, hpux and the storage stuff. Now that would
> make bill's day :)


Then you'd hear the cries of "Intel, why are you competing on our turf - we
buy chips from you" coming from Dell, IBM, SGI, etc....

Might just be enough to drive the bulk of x86 business into the hands of
AMD. Intel won't let that happen. They'd kill VMS et. al. before that.
--
OpenVMS - The classics never go out of style.