Full port of VMS to Itanium.
John Macallister
Itanium ... starting now." .
That, to me, appears to be a fairly positive commitment to the longer term
future of VMS.
The Powerpoint slides of the announcement are at URL:
http://www.compaq.com/hps/ipf-enterprise/download/webcast.ppt
John
Name: John B. Macallister E-mail: j.macal...@physics.ox.ac.uk
Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UK
Phone: +44-1865-273388 (direct) 273333 (reception) 273418 (Fax)
fabio_...@ep-bc.petrobras.com.br
A lot of companies will not port from Alpha to Itanium.
Bad point of view: I believe a lot of customres will jump to long-time
processors
as SPARC or Power PC.
Good point of view: I can improve my value $$$ in the market because
OpenVMS
professionals will become much more
rares .....
Regards
FC
John Macallister <J.Macal...@physics.ox.ac.uk> em 25/06/2001 10:18:02
Favor responder a John Macallister <J.Macal...@physics.ox.ac.uk>
Assunto: Full port of VMS to Itanium.
john nixon
development group (kicking and screaming) into converting from VAX to
Alpha. Last week, we ordered several Alpha's to replace our VAX 7860s.
The conversion should be completed this year.
What a kick in the pants this is!
I hope I still have a job.
I hope Compaq starts yelling from the rooftops how quickly they will port
VMS to Itanium, and how easy the port from AlphaVMS will be. The
competitors are already spreading the word that Alpha (and thus VMS) is
dead.
<fabio_...@ep-bc.petrobras.com.br> wrote in message
news:OFFB99FC77.F49E9947...@ep-bc.petrobras.com.br...
Bob Koehler
> Bad point of view: I believe a lot of customres will jump to long-time
> processors
> as SPARC or Power PC.
I think both PowerPC and IA-64 will easily outlive SPARC.
----------------------------------------------------------------------
Bob Koehler | Computer Sciences Corporation
NASA GSFC Flight Software | Federal Sector, Civil Group
| please remove ".aspm" when replying
Tom Linden
important,
provide of course that the truck performs to specs. You just traded in an
old on for a newer one, which will at some point be replaced again, and
MAYBE it won't have such severe alignement requirements. That's how you
sell it and keep your job. :->
> -----Original Message-----
> From: john nixon [mailto:jni...@cfl.rr.com]
> Sent: Monday, June 25, 2001 8:54 AM
> To: Info...@Mvb.Saic.Com
> Subject: Re: Full port of VMS to Itanium.
>
>
> I have finally, after 10 years or so of trying, finally prodded our
> development group (kicking and screaming) into converting from VAX to
> Alpha. Last week, we ordered several Alpha's to replace our VAX 7860s.
> The conversion should be completed this year.
>
> What a kick in the pants this is!
>
> I hope I still have a job.
>
> I hope Compaq starts yelling from the rooftops how quickly they will port
> VMS to Itanium, and how easy the port from AlphaVMS will be. The
> competitors are already spreading the word that Alpha (and thus VMS) is
> dead.
>
> <fabio_...@ep-bc.petrobras.com.br> wrote in message
> news:OFFB99FC77.F49E9947...@ep-bc.petrobras.com.br...
> > I believe will happend all that troubles of porting the VAX to Alpha.
> > A lot of companies will not port from Alpha to Itanium.
> >
> > Bad point of view: I believe a lot of customres will jump to long-time
> > processors
> > as SPARC or Power PC.
> >
Fred Kleinsorge
Going from Alpha to IPF (Itanium Processor Family) will be less tramatic. I
believe that if you have gotten to Alpha, odds are it will be a recompile
and go.
Were you buying Alpha? Or were you buying VMS? I think that in the grand
scheme of things, you were buying OpenVMS for the things that it brings to
the table... if we could build a VAX that was as fast as Alpha - you'd still
be on VAX.
While I am, and remain, an Alpha supporter. You now will have a chance to
see VMS on a "open" hardware platform.
fabio_...@ep-bc.petrobras.com.br wrote in message ...
Fred Kleinsorge
>I have finally, after 10 years or so of trying, finally prodded our
>development group (kicking and screaming) into converting from VAX to
>Alpha. Last week, we ordered several Alpha's to replace our VAX 7860s.
>The conversion should be completed this year.
>
>What a kick in the pants this is!
>
>I hope I still have a job.
>
For the next several years, Alpha is arguably the best HW available.
Remember, we will continue and implement the EV7 family and system
platforms. If you get from VAX to Alpha, the "port" to IPF should, we
believe, be a recompile & relink. It is also a 99.8% probability that there
will be mixed cluster capability.
>I hope Compaq starts yelling from the rooftops how quickly they will port
>VMS to Itanium, and how easy the port from AlphaVMS will be. The
>competitors are already spreading the word that Alpha (and thus VMS) is
>dead.
Actually, the engineers here in Nashua seem, well, to be almost universally
positive about this. At the same time as it was announced, we also were
told our headcount would increase by a number that we will be hard pressed
to find engineers to fill quickly enough.
Having ported VMS once, we know we can do it again. There will be some
challenges ahead, but we don't know of any show-stoppers. We'll know just
how "easy" it is in a little while. The OS has some significant challenges
that most user code doesn't know or care about (like the console program,
etc).
Larry Kilgallen
> Actually, the engineers here in Nashua seem, well, to be almost universally
> positive about this. At the same time as it was announced, we also were
> told our headcount would increase by a number that we will be hard pressed
> to find engineers to fill quickly enough.
I have my doubts about the quality of work from two-headed engineers :-)
Christopher Smith
> Were you buying Alpha? Or were you buying VMS? I think that
> in the grand
> scheme of things, you were buying OpenVMS for the things that
> it brings to
> the table... if we could build a VAX that was as fast as
> Alpha - you'd still
> be on VAX.
That's an interesting remark, but what exactly does Itanic have that Alpha
does not? I assume it should be significantly cheaper... I only hope that
there aren't holes in the design so large that even VMS engineering can't
work around them. I do have lots of respect for the VMS engineering group.
> While I am, and remain, an Alpha supporter. You now will
> have a chance to
> see VMS on a "open" hardware platform.
I'll believe that when I see the platform become open.
Regards,
Chris
Christopher Smith, Perl Developer
Amdocs - Champaign, IL
/usr/bin/perl -e '
print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");
'
Fred Kleinsorge
and we can get you and your 3 heads back ;-)
Larry Kilgallen wrote in message ...
Fred Kleinsorge
>does not?
Long term marketshare.
Look. I am an Alpha supporter. I think that EV68 is the best processor on
the market today. I think EV7 will be a kick-ass chip, and system. The
Itanium has nothing on it. Until recently, there was a strong question of
if they would *ever* get IA64 to work well enough to be interesting. But
the recent benchmark numbers show that they willl be in the ballpark soon.
So, project it out. Sometime in the mid-late 2000's, IA64 is very
competetive with Alpha on performance, but with HP and Microsoft on it, it
sells many times more than Alpha. The price is lower than Alpha. Not only
that, we don't have to build two seperate lines of HW for IA64 and Alpha.
Lots cheaper. Heck, *maybe*, just *maybe* it means that if we do it right,
we could put VMS on any Itanium box that we want to qual it on, including
the "cheap" ones too.
I'm pretty sure we can pull thins thing off. It's just a matter of how
easy/hard it will be.
Christopher Smith wrote in message
<3B55D7F383B0D31197D9...@cmiexch1.cmi.itds.com>...
Christopher Smith
> Long term marketshare.
Good answer.
> Itanium has nothing on it. Until recently, there was a
> strong question of
> if they would *ever* get IA64 to work well enough to be
> interesting. But
> the recent benchmark numbers show that they willl be in the
> ballpark soon.
I wasn't aware that the prognosis had changed, actually, but that's an
interesting note, too.
> So, project it out. Sometime in the mid-late 2000's, IA64 is very
> competetive with Alpha on performance, but with HP and
> Microsoft on it, it
> sells many times more than Alpha. The price is lower than
> Alpha. Not only
> that, we don't have to build two seperate lines of HW for
> IA64 and Alpha.
> Lots cheaper. Heck, *maybe*, just *maybe* it means that if
> we do it right,
> we could put VMS on any Itanium box that we want to qual it
> on, including
> the "cheap" ones too.
Sounds like a chance for VMS to really compete again. That will be welcome
if it turns out that way.
> I'm pretty sure we can pull thins thing off. It's just a
> matter of how
> easy/hard it will be.
I guess the best thing to hope for now is that the Alpha has a maximum
amount of impact on future IA64 designs. My next question is, what is there
to stop that? I assume that IA64, as it stands, has a lot of political
clout inside of Intel. It must to have been so terribly bad before, and
still get development money pumped in. :)
Robert Deininger
I like reading optimism from Nashua.
I already look at the Compaq jobs on the openVMS portal. There were still
only 7 listed. Will that job list recover some of it's former glory?
I'll keep asking, without expecting an answer. Were you folks consulted
about this VMS porting project ahead of time?
--
Robert Deininger
rdein...@mindspring.com
JF Mezei
> Today's announcement says that there will be a "Full port of ... VMS .. to
> Itanium ... starting now." .
>
> That, to me, appears to be a fairly positive commitment to the longer term
> future of VMS.
1- Alpha is dead as of TODAY. Who in their right mind will invest in a million
dollar wildfire system today knowing the architecture is dead ?
2-With nobody buying Alpha systems, nobody will be buying VMS systems. VMS
revenus will dry up REAL FAST. Remember, it will take 4 (and knowing intel,
probably 8) years before IA64 has the bit in it to support VMS and more
importantly NSK. Meanwhile, NT will happily chug along on IA64.
3- This means that Compaq is wasting money porting VMS to IA64 (the port is
probably just to save face). By the time VMS is ready on IA64, its market will
be greatly diminshed.
4-The only way Compaq can offload their current stock of Alpha is to lower the
prices WAY DOWN. I guess thsi is great news for hobyists.
In one feel swooop, Compaq has managed to kill the revival effort that had
been made for VMS in the past year. The uncertainty over VMS that will last
many years will make sure VMS returns into major downsizing mode. I personally
do not beleive that VMS will survive. (Sure, Compaq will continue to support
it, but it will have fewer and fewer customers).
Robert Deininger
<3B55D7F383B0D31197D9...@cmiexch1.cmi.itds.com>,
Christopher Smith <csm...@amdocs.com> wrote:
> That's an interesting remark, but what exactly does Itanic have that Alpha
> does not?
Competent management and marketing. But alpha has had the best engineers.
The best we can hope for is that intel uses the new people and property
wisely, rather than burying them.
--
Robert Deininger
rdein...@mindspring.com
Larry Kilgallen
> 1- Alpha is dead as of TODAY. Who in their right mind will invest in a million
> dollar wildfire system today knowing the architecture is dead ?
Somebody who wants performance _today_ and has kept their source code
for a future port.
fabio_...@ep-bc.petrobras.com.br
sure
if will be 80 or 160).
I believe the company will wait...
Everybody should wait ! !
Reg
FC
"Fred Kleinsorge" <klein...@star.zko.dec.com> em 25/06/2001 14:05:36
Favor responder a "Fred Kleinsorge" <klein...@star.zko.dec.com>
Assunto: Re: Full port of VMS to Itanium.
I believe that the VAX port probably was the most tramatic for everyone.
Going from Alpha to IPF (Itanium Processor Family) will be less tramatic.
I
believe that if you have gotten to Alpha, odds are it will be a recompile
and go.
Were you buying Alpha? Or were you buying VMS? I think that in the grand
scheme of things, you were buying OpenVMS for the things that it brings to
the table... if we could build a VAX that was as fast as Alpha - you'd
still
be on VAX.
While I am, and remain, an Alpha supporter. You now will have a chance to
see VMS on a "open" hardware platform.
fabio_...@ep-bc.petrobras.com.br wrote in message ...
>I believe will happend all that troubles of porting the VAX to Alpha.
>A lot of companies will not port from Alpha to Itanium.
>
>Bad point of view: I believe a lot of customres will jump to long-time
>processors
> as SPARC or Power PC.
>
>Good point of view: I can improve my value $$$ in the market because
>OpenVMS
> professionals will become much more
>rares .....
>
>
>Regards
>
>FC
>
>
>
>
>John Macallister <J.Macal...@physics.ox.ac.uk> em 25/06/2001 10:18:02
>
>Favor responder a John Macallister <J.Macal...@physics.ox.ac.uk>
>
>
>
> Info...@Mvb.Saic.Com
>
>
>
>Assunto: Full port of VMS to Itanium.
>
>
>Today's announcement says that there will be a "Full port of ... VMS .. to
>Itanium ... starting now." .
>
>
>That, to me, appears to be a fairly positive commitment to the longer term
>future of VMS.
>
Fred Kleinsorge
Robert Deininger wrote in message ...
>I'll keep asking, without expecting an answer. Were you folks consulted
>about this VMS porting project ahead of time?
>
Only a handful of very senior people, some of whom were *highly* technical,
were involved in the decision. Oddly enough, the entire thing was started
by a handful of our most senior technical HW people doing a analysis
projecting things out through the next 5-10 years. It wasn't management
pushing it down.
But very few of the folks in the trenches, are told of this type thing -
mostly for legal reasons.
We're digesting this, just like you are. The good news is that we are all
pretty much looking at this as something pretty fun/cool to do.
John Vottero
news:3B378774...@videotron.ca...
> John Macallister wrote:
> > Today's announcement says that there will be a "Full port of ... VMS ..
to
> > Itanium ... starting now." .
> >
> > That, to me, appears to be a fairly positive commitment to the longer
term
> > future of VMS.
>
> 1- Alpha is dead as of TODAY. Who in their right mind will invest in a
million
> dollar wildfire system today knowing the architecture is dead ?
>
What architecture would you suggest they buy? IA64 isn't available. SPARC
looks like it's going to be the odd man out. HP has announced the end of
life for PA-RISC.
If you buy an Alpha, in 2004 you can bring an IA64 machine into the cluster,
recompile and go. That sounds pretty good to me.
> 2-With nobody buying Alpha systems, nobody will be buying VMS systems. VMS
> revenus will dry up REAL FAST. Remember, it will take 4 (and knowing
intel,
> probably 8) years before IA64 has the bit in it to support VMS and more
> importantly NSK. Meanwhile, NT will happily chug along on IA64.
>
People have been claiming that VMS revenues were going to dry up "REAL FAST"
for years, it hasn't happened and isn't about to.
> 3- This means that Compaq is wasting money porting VMS to IA64 (the port
is
> probably just to save face). By the time VMS is ready on IA64, its market
will
> be greatly diminshed.
>
Are you suggesting that Compaq should abandon VMS? What do you want them to
do? They could spend Billions creating EV8, EV9 etc or spend Millions
porting VMS to IA64.
> 4-The only way Compaq can offload their current stock of Alpha is to lower
the
> prices WAY DOWN. I guess thsi is great news for hobyists.
>
> In one feel swooop, Compaq has managed to kill the revival effort that had
> been made for VMS in the past year. The uncertainty over VMS that will
last
> many years will make sure VMS returns into major downsizing mode. I
personally
> do not beleive that VMS will survive. (Sure, Compaq will continue to
support
> it, but it will have fewer and fewer customers).
Where's the uncertainty? Yesterday I was wondering if Alpha would be able
to survive and what would happen to VMS if it didn't, today I know what the
future of VMS is. And it looks brighter than it did yesterday.
JF Mezei
> Were you buying Alpha? Or were you buying VMS?
Combination of both which yielded the quality, stability and configuration
options that gave VMS an edge.
Consider who is left as VMS customers today. Consider stuff such as Wildfires,
and the special ALpha processors that allow wildfires, and the integration
that allows Galaxy. These were made possible because Alpha was mature enough
to acquire the features to do this. IA674 is an untested kid probably still
with math errors such as what Pentium had.
Andrew Harrisson, is Sun hiring ?
JF Mezei
> platforms. If you get from VAX to Alpha, the "port" to IPF should, we
> believe, be a recompile & relink. It is also a 99.8% probability that there
> will be mixed cluster capability.
It isn't just a question of recompiling a product. It is also a question of
recertifying it, and duplicating/updating the support stuff such as
documentation, training support personel, updating customer database to know
whether they run on the dead Alpha or the intel crap, providing multiple media
and making sure the customers get the right ones, and for every path you
issue, you have to do all that extra work too. It was bad enough that Digital
failed in its transition of VAX->ALPHA (the fact that so many vaxes remain in
production 10 years after is a testament to that failure, with demand for new
VAXes continuing), but now you have a third platform to support on an anemic
operating system whose customers will be split on 3 platforms.
> Actually, the engineers here in Nashua seem, well, to be almost universally
> positive about this. At the same time as it was announced, we also were
> told our headcount would increase by a number that we will be hard pressed
> to find engineers to fill quickly enough.
You could hire all the Alpha engineers. I bet and I hope that many of them
will refuse to work for Intel.
Hunter Goatley
>I believe will happend all that troubles of porting the VAX to Alpha.
>A lot of companies will not port from Alpha to Itanium.
>
As others have said elsewhere, I've ported a large number of programs/products
from VAX to Alpha. Once the Alpha compilers had matured some, the porting was
trivial for all but the really low-level kernel-mode code that was tired closely
to the VAX architecture. I would expect that once Compaq has the GEM compiler
back-end generating IA64 code, porting the code should be as simple as a
recompile of the code. (GEM didn't exist on VAX, so moving from VAX to Alpha
wasn't quite as simple; since all (AFAIK) the Alpha VMS compilers use the GEM
back-end, there should be no compiler growing pains in the new architecture,
once GEM has been fixed.)
That's what I hope, anyway.....
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goath...@goatley.com http://www.goatley.com/hunter/
Christopher Smith
> From: John Vottero [mailto:Jo...@mvpsi.com]
> Where's the uncertainty? Yesterday I was wondering if Alpha
> would be able
> to survive and what would happen to VMS if it didn't, today I
> know what the
> future of VMS is. And it looks brighter than it did yesterday.
Well, on the bright side, if there's a low power Itanic being planned, we
may see another VMS-capable notebook.
Hunter Goatley
>>That's what I hope, anyway.....
>
And has the caveat that VMS itself is running well on IA64. ;-)
David J. Dachtera
>
> I believe will happend all that troubles of porting the VAX to Alpha.
> A lot of companies will not port from Alpha to Itanium.
>
> Bad point of view: I believe a lot of customres will jump to long-time
> processors
> as SPARC or Power PC.
>
> Good point of view: I can improve my value $$$ in the market because
> OpenVMS
> professionals will become much more
> rares .....
...and OpenVMS jobs will become even more rare as customers flee from
such a flighty, uncommitted platform. VAX lived 20 years or more. Alpha
less than a decade. x86 and later - going strong since before Alpha.
We could have been banging away at OpenVMS-Intel for a good number of
years already, and Billy would have had to displace those x86 servers
running it.
Sure glad I'm looking into another line of work...
--
David J. Dachtera
dba DJE Systems
http://www.djesys.com/
Unofficial Affordable OpenVMS Home Page and Message Board:
http://www.djesys.com/vms/soho/
This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings
is to be expected.
Feel free to exercise your rights of free speech and expression.
However, attacks against individual posters, or groups of posters, are
strongly discouraged.
JF Mezei
> We're digesting this, just like you are. The good news is that we are all
> pretty much looking at this as something pretty fun/cool to do.
Until you realise that this decision will force most VMS customers to rething
their VMS investment and yo start seing VMS sales come to a scheeching halt.
And when Compaq's financials won't improve, cutbacks will have to be made.
Anyways it is moot. Unless Compaq commits a massive marketing budget for VMS,
VMS isn't strong enough to survive this period of uncertainty.
I feel for the folks who had invested in Wildfires and will now have to look
at lower quality options in the longer term.
David J. Dachtera
>
> > -----Original Message-----
> > From: Fred Kleinsorge [mailto:klein...@star.zko.dec.com]
>
> > Were you buying Alpha? Or were you buying VMS? I think that
> > in the grand
> > scheme of things, you were buying OpenVMS for the things that
> > it brings to
> > the table... if we could build a VAX that was as fast as
> > Alpha - you'd still
> > be on VAX.
>
> That's an interesting remark, but what exactly does Itanic have that Alpha
> does not? I assume it should be significantly cheaper... I only hope that
> there aren't holes in the design so large that even VMS engineering can't
> work around them. I do have lots of respect for the VMS engineering group.
Have a read over on comp.sys.intel. Looks like Intel can't even work
around the "holes" in Itanic.
JF Mezei
> What architecture would you suggest they buy? IA64 isn't available. SPARC
> looks like it's going to be the odd man out. HP has announced the end of
> life for PA-RISC.
They will just put their major purchases that are VMS related on hold. The big
winners in this will be IBM and SUN.
> If you buy an Alpha, in 2004 you can bring an IA64 machine into the cluster,
> recompile and go. That sounds pretty good to me.
VMS was experiencing somewhat of a rebirth thanks to some tokem marketing
efforts done last year. That momentum will now be lost. Do you think that any
ISV who might have seen that rebirth as a sign to port their software to
Alpha-VMS will now bother porting ? No, he will also wait 4 more years at
which point, it might be moot because interest in VMS will have gone down big
time because VMS will have been stagnant for 4 years.
> People have been claiming that VMS revenues were going to dry up "REAL FAST"
> for years, it hasn't happened and isn't about to.
VMS revenus dried up sufficiently that Digital lost money BIG TIME for a long
enough time that Digital has to give itseld away to a small PC company named
Compaq. What is left is but justa small portion of what VMS used to be. And
this period of undertainty will not help VMS.
> Are you suggesting that Compaq should abandon VMS? What do you want them to
> do? They could spend Billions creating EV8, EV9 etc or spend Millions
> porting VMS to IA64.
Instead of cannabalising VMS and Alpha, Compaq should have sold it to an
outfit that wanted to make a go at it and compete against Intel and Microsoft.
Compaq has capitulated to Intel and Microsoft with today's announcement.
Fred Kleinsorge
>Fred Kleinsorge wrote:
>> Were you buying Alpha? Or were you buying VMS?
>
>Combination of both which yielded the quality, stability and configuration
>options that gave VMS an edge.
>
>Consider who is left as VMS customers today. Consider stuff such as
Wildfires,
>and the special ALpha processors that allow wildfires, and the integration
>that allows Galaxy.
Nothing in OpenVMS, or Galaxy, is Alpha-specific, except perhaps in
implementation. With the Alpha design team going to Intel, I would not be
suprised to see things from EV7 and EV8 sneak into Itanium in the future.
> These were made possible because Alpha was mature enough
>to acquire the features to do this. IA674 is an untested kid probably still
>with math errors such as what Pentium had.
>
We're pretty confident that by the time we get things ported, it will be in
good shape, and EV7 systems will give us a good transition padding with a
high-performance Alpha swan-song.
>Andrew Harrisson, is Sun hiring ?
Sun is trying not to lay off.
Brian Wheeler
Speaking of porting...since Itanium is (technically) a commodity processor,
what's the chance of an off-the-street Itanium box running VMS?
Brian Wheeler
bdwh...@indiana.edu
David J. Dachtera
>
> > -----Original Message-----
> > From: John Vottero [mailto:Jo...@mvpsi.com]
>
> > Where's the uncertainty? Yesterday I was wondering if Alpha
> > would be able
> > to survive and what would happen to VMS if it didn't, today I
> > know what the
> > future of VMS is. And it looks brighter than it did yesterday.
>
> Well, on the bright side, if there's a low power Itanic being planned, we
> may see another VMS-capable notebook.
Excellent bit of sunshine, Chris!
WILLIAM WEBB
I try to see things positively.
1) Port to Itanium = the long-awaited $1000 VMS desktop box.
2) Adoption of the IBM model = real resources thrown into the
VMS ASP program. You got ta have software to sell software...
WWWebb
David Mathog
>
>We're digesting this, just like you are. The good news is that we are all
>pretty much looking at this as something pretty fun/cool to do.
Let's hope "incredibly fast" is also in there somewhere. The odds are very
high that thanks to this announcement Compaq just kissed goodbye to at least
80% of its OpenVMS sales for the rest of 2001 and beyond. The money won't
come back until people see that OpenVMS running on Intel works, is
reliable, has software to run on the new platform, has some modicum of
real support from Compaq, and isn't insanely expensive to migrate to.
That's a tall order, but there's no way the bozos who run your company are
going to wait very long in the face of an extreme drop in product revenues for
OpenVMS before they cancel the port. And to shout the obvious:
***COMPAQ MANAGEMENT WANTS THIS PORT TO FAIL***
That way they don't have to sell anything that competes with Microsoft. If
the OpenVMS group is extremely lucky Q management will _maybe_ give them a
year before they pull the plug. Terry has estimated in this forum that it
should take 18 months. And I think Terry's estimate is probably
optimistic - given the size of the project doing it in 18 months would
probably require that OpenVMS engineering was already geared up for the
port now, with compilers for the new platform for all languages _and_
development hardware on hand. Do you have either? When do you expect to
have either?
After the lot of you (meaning OpenVMS Engineering) have taken a few weeks
to look over the options and considered carefully the real level of support
Compaq will provide, and also, critically, how this is going to turn out
for you all personally if/when Compaq management cancels the project, you
may want to consider extreme action to protect yourselves (or if that isn't
found to be possible, to at least go out with your heads held high.) For
instance, if after careful study you folks decide that you're being set up
for the fall, you could probably at least drag most of Compaq management
down with you by going out on strike in the immediate future. Or as a
better result, you might be able to force Q management to guarantee in some
non-negatable (contractual) way sufficient resources to complete the port
or, to at least give the members of OpenVMS Engineering full rights to
OpenVMS and funds for a spin off if Compaq pulls the plug. But for optimum
bargaining power such a strike must occur before the revenues start to
fall. Ie, very soon. And if you strike, and they won't meet any of your
demands, you'll have learned their true intentions before you all spent a
year or two killing yourselves to complete this port.
Before anybody jumps on me for suggesting such a disruptive action,
please keep in mind that at this point I'm concerned only about the welfare
of the folks in OpenVMS engineering. They really need to think long and
hard about their personal stakes in this port, this company, and VMS in
general, especially considering how Compaq management's "support" for Alpha
has turned out, and given its similar level of "support" for OpenVMS. They
really need to consider all their options at this point, and blindly
following orders will turn out to be a poor choice if general Mike only
wanted them for cannon fodder.
Regards,
David Mathog
mat...@caltech.edu
Manager, sequence analysis facility, biology division, Caltech
**************************************************************************
* RIP VMS & ALPHA *
**************************************************************************
Christopher Smith
Regards,
Chris
Christopher Smith, Perl Developer
Amdocs - Champaign, IL
/usr/bin/perl -e '
print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");
'
> -----Original Message-----
> From: David J. Dachtera [mailto:djesys...@fsi.net]
Larry Kilgallen
> In article <99MZ6.117$rc5....@news.cpqcorp.net>, "Fred Kleinsorge" <klein...@star.zko.dec.com> writes:
>>
>>We're digesting this, just like you are. The good news is that we are all
>>pretty much looking at this as something pretty fun/cool to do.
>
> Let's hope "incredibly fast" is also in there somewhere.
No, let's hope "incredibly correct" is in there.
IA-64 already has lesser operating systems.
Larry Kilgallen
--
==============================================================================
Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> Clusters
==============================================================================
Jordan Henderson
David Mathog <mat...@seqaxp.bio.caltech.edu> wrote:
>[snip]
>
> ***COMPAQ MANAGEMENT WANTS THIS PORT TO FAIL***
>
That's as absurd as anything that's been posted here recently. And that's
saying a LOT.
Compaq being unable to manage a port to the processor that it looks like
everyone will be supporting in the future (see recent HP announcement, heck
even Sun has hedged their bets by having a working Solaris port _today_)
will not bode well for their ability to sell into those high-margin
Enterprise accounts that they are obviously trying to nurture with
their entire new strategy.
>[snip]
>
>Before anybody jumps on me for suggesting such a disruptive action,
>please keep in mind that at this point I'm concerned only about the welfare
>of the folks in OpenVMS engineering. They really need to think long and
>hard about their personal stakes in this port, this company, and VMS in
>general, especially considering how Compaq management's "support" for Alpha
>has turned out, and given its similar level of "support" for OpenVMS. They
>really need to consider all their options at this point, and blindly
>following orders will turn out to be a poor choice if general Mike only
>wanted them for cannon fodder.
>
People have been jumping up and down, red-in-the-face, screaming in this
newsgroup for VMS on commodity hardware. It looks like Compaq has taken
a HUGE step in that direction and many of the same people are screaming
"woe is me, the end of VMS is nigh!"
I remember a quote from the Commodore management when the Amiga died
(effectively). The manager was quoted as saying something to the
effect "Our staunch advocates in the community didn't do us any favors
by being so shrill all the time."
People need to turn down the paranoia a little.
>Regards,
>
>David Mathog
>mat...@caltech.edu
>Manager, sequence analysis facility, biology division, Caltech
>**************************************************************************
>* RIP VMS & ALPHA *
>**************************************************************************
-Jordan Henderson
jor...@greenapple.com
Terry C Shannon
On Mon, 25 Jun 2001, Robert Deininger wrote:
> In article
> <3B55D7F383B0D31197D9...@cmiexch1.cmi.itds.com>,
> Christopher Smith <csm...@amdocs.com> wrote:
>
>
> > That's an interesting remark, but what exactly does Itanic have that Alpha
> > does not?
>
> Competent management and marketing. But alpha has had the best engineers.
>
Alpha's Omega was rendered inevitable by Digital's outright marketing
malfeasance and a string of Stupid Strategy Tricks. Compaq's initial
indecisiveness and tepid support for Alpha (and the support was far less
than tepid amongst the Compaq Classic box-pushers) delayed a turn-around
in the business, which actually has grown during the past two years.
Too little, too late.
Christof Brass
>
> I have finally, after 10 years or so of trying, finally prodded our
> development group (kicking and screaming) into converting from VAX to
> Alpha. Last week, we ordered several Alpha's to replace our VAX 7860s.
> The conversion should be completed this year.
>
> What a kick in the pants this is!
>
> I hope I still have a job.
>
> VMS to Itanium, and how easy the port from AlphaVMS will be. The
> competitors are already spreading the word that Alpha (and thus VMS) is
> dead.
>
> >
> > Name: John B. Macallister E-mail: j.macal...@physics.ox.ac.uk
> > Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UK
> > Phone: +44-1865-273388 (direct) 273333 (reception) 273418 (Fax)
Well done! I don't think that your job is at risk. Image how
long it would take to migrate to IA63? With a good Alpha
solution your company could well operate additional 10 to 15
years depending of the demands and apps available to the
platform. But given that there are still a lot of VAX versions
around (compared to the low number of VMS apps) I wouldn't worry
about not having Alpha apps around.
Christof Brass
>
> john nixon wrote in message ...
> >development group (kicking and screaming) into converting from VAX to
> >Alpha. Last week, we ordered several Alpha's to replace our VAX 7860s.
> >The conversion should be completed this year.
> >
> >What a kick in the pants this is!
> >
> >I hope I still have a job.
> >
>
> Remember, we will continue and implement the EV7 family and system
> platforms. If you get from VAX to Alpha, the "port" to IPF should, we
> believe, be a recompile & relink. It is also a 99.8% probability that there
> will be mixed cluster capability.
>
> >VMS to Itanium, and how easy the port from AlphaVMS will be. The
> >competitors are already spreading the word that Alpha (and thus VMS) is
> >dead.
>
> positive about this. At the same time as it was announced, we also were
> told our headcount would increase by a number that we will be hard pressed
> to find engineers to fill quickly enough.
>
> Having ported VMS once, we know we can do it again. There will be some
> challenges ahead, but we don't know of any show-stoppers. We'll know just
> how "easy" it is in a little while. The OS has some significant challenges
> that most user code doesn't know or care about (like the console program,
> etc).
Okay then.
May I ask for a multiplatform port? :-)
Would you guys retain Alpha while porting to IA63?
Would you guys do the port in a way that you have a PowerPC and
a SPARC version also at the same time?
I'm SW engineer and I know that this is not a simle thing. But
we all know that supporting VAX, Alpha and IA63 is the hard
part. Well done porting should allow for additional HW platforms
at marginal costs.
My experience with porting app SW to different platforms is that
each platform reduces the number of generic bugs and design
weaknesses but adds platform specific bugs while revealing
certain bugs that didn't show up on other OSs. At the end the
quality of the SW is much higher.
John Santos
> John Macallister wrote:
> > Today's announcement says that there will be a "Full port of ... VMS .. to
> > Itanium ... starting now." .
> >
> > That, to me, appears to be a fairly positive commitment to the longer term
> > future of VMS.
>
> 1- Alpha is dead as of TODAY. Who in their right mind will invest in a million
> dollar wildfire system today knowing the architecture is dead ?
By this reasoning, P4 is dead. Why would anyone invest in a million dollars
worth of P4 NT servers?
(I left out "in their right mind" on purpose. ;-)
> 2-With nobody buying Alpha systems, nobody will be buying VMS systems. VMS
> revenus will dry up REAL FAST. Remember, it will take 4 (and knowing intel,
> probably 8) years before IA64 has the bit in it to support VMS and more
> importantly NSK. Meanwhile, NT will happily chug along on IA64.
This is the great danger with this strategy, I think. They need to
convince potential customers that switching from Alpha to IA64 in 3 years
will be an upgrade, not a migration, and will be utterly painless.
One thing I found with VAX->Alpha: It's a lot easier if you haven't lost
your sources and if you have working build procedures to recompile
everything!
> 3- This means that Compaq is wasting money porting VMS to IA64 (the port is
> probably just to save face). By the time VMS is ready on IA64, its market will
> be greatly diminshed.
The port is essential to my answer to 2, and thus to preventing the market
from diminishing.
> 4-The only way Compaq can offload their current stock of Alpha is to lower the
> prices WAY DOWN. I guess thsi is great news for hobyists.
No, unfortunately. I'd like to get a decent Alpha for home use, so this
means they will stay expensive. Now, if I were to buy an expensive
Alpha for home use, the price would immediately plummet. (This has
happens whenever I buy PC hardware, but I've never noticed this
effect on software. Instead, the vendor releases a new version with
critical bugs fixed about one week after any free or cheap upgrade
period expires.)
> In one feel swooop, Compaq has managed to kill the revival effort that had
> been made for VMS in the past year. The uncertainty over VMS that will last
> many years will make sure VMS returns into major downsizing mode. I personally
> do not beleive that VMS will survive. (Sure, Compaq will continue to support
> it, but it will have fewer and fewer customers).
I hope you are wrong. It depends on both technology (having the
port go smoothly) and marketing (convincing the customers of that.)
I have a lot of faith in VMS engineering getting things right,
though not always quickly enough to keep me perfectly happy, but
marketing is another story... :-(
--
John Santos
Evans Griffiths & Hart, Inc.
781-861-0670 ext 539
John Santos
I don't know what the internal interface between the compilers
and the GEM back-end looks like, but since all the compiles on
Alpha use it, it must be fairly clean.
How difficult would it be to split the compilers from GEM so
that developers could ship intermediate code and the customer
could do the GEM phase at installation time? Then you could
ship a common kit for both Alpha and IA64, with the target
platform selected by the customer's GEM compiler (shipped as
part of VMS.)
A second benefit is (I think) most or all the optimization stuff
is done by GEM, not by the front-end compilers. (Some of the Compaq
compiler-writers, Steve Lionel I think, have mentioned that the
optimization qualifiers, e.g. for the Fortran compiler are just
passed through to GEM, which is what leads me to believe this.) So
you could pre-compile on whatever platform the developer has, but
final-compile for not only the architecture but the specific chip
on the end user's system. If you really wanted to get every ounce
of speed, you could do it without having to ship multiple .OLB's
optimized (with /optimize=tune=xxx) for each architecture.
It might even be possible to create a VAX back-end (no optimizing,
except maybe removing redundant instructions or other trivial stuff),
so you wouldn't even need to ship a separate VAX version.
This might just be a pipe-dream though. There doesn't seem to a
separate "GEM" shareable image, which you would expect if there
was a really clean interface between the compiler's front-end
and GEM. It could also be GEM is still a moving targer, and the
compiler developers always want to ship the latest version with
their compiler since it has new features they need?
Alphaman
news:acKZ6.98$rc5....@news.cpqcorp.net...
> I believe that the VAX port probably was the most tramatic for everyone.
> Going from Alpha to IPF (Itanium Processor Family) will be less tramatic.
I
But traumatic, nonetheless. We are still cranking out both VAX and Alpha
code to this day and into the foreseeable future, still supporting VAX
customers. Now, we have to make it 3 platforms? Why port to VMS on IA64
when one could port to NT on IA64? Contrary to facts, the perception is
that NT is less expensive, and 2000 is "13 times more reliable". And hey,
IA64/Windoze 2000 developers' systems are available in beta today -- why not
get a head start instead of waiting for OpenVMS-IA64?
> While I am, and remain, an Alpha supporter. You now will have a chance to
> see VMS on a "open" hardware platform.
Just how open? Will VMS run on a Dell box, Taiwanese clone, or Q only? And
what is your definition of "open"? You mean it won't be like Alpha, just
one processor family controlled by one company? Oops, like, say, Intel?
That's "open"? Or are you saying that work is afoot to make IA64 just the
first platform, with Sledgehammer, PowerPC, SPARC, and others to follow, in
a true open model?
Not having a choice in processors is definitely NOT "open", IMHO. OpenVMS
remains an oxymoron, just that after EV7, part of the money goes to Intel
instead of Compaq.
There's an old adage that applies very well today. "Fool me once, shame on
you. Fool me twice, shame on me." Forgive me, as I'm trying to decide if
the Q has fooled me a second time, and what I should do before the third
time strikes.
Nothing against you or your team mates, Fred -- you do fine work.
Aaron, who does not like being made a fool of.
--
Aaron Sakovich http://members.home.net/sakovich/alphaman.html
Make April 15 just another day: http://www.fairtax.org/
"The supersonic boom should hit you in just a few seconds." (Apollo 440)
Larry Kilgallen
> This might just be a pipe-dream though. There doesn't seem to a
> separate "GEM" shareable image, which you would expect if there
> was a really clean interface between the compiler's front-end
> and GEM. It could also be GEM is still a moving targer, and the
> compiler developers always want to ship the latest version with
> their compiler since it has new features they need?
Indeed, I have received bug report answers for certain languages
that say "we can't do that because our compiler is running on a
different version of GEM than the languages that support that".
And a compiler language group would certainly not want bug reports
from someone running their front end against a version of GEM they
had not tested.
Larry Kilgallen
> Just how open? Will VMS run on a Dell box, Taiwanese clone, or Q only?
Fred already answered that in this newsgroup, saying they would try to
avoid technical designs that prevented compatibility. But, as always,
there are the issues of "qualification" and "support".
Scott Vieth
fabio_...@ep-bc.petrobras.com.br wrote:
> I believe will happend all that troubles of porting the VAX to Alpha.
> A lot of companies will not port from Alpha to Itanium.
>
Yes they will. When the fastest AlphaServer is no longer fast enough, they
will move to a hot Intel-based box. Same thing that happened when people hit
a performance ceiling with the VAX line. A fully-loaded VAX 7800 only goes so
fast.
-Scott
Scott Vieth
Brian Wheeler wrote:
> Speaking of porting...since Itanium is (technically) a commodity processor,
> what's the chance of an off-the-street Itanium box running VMS?
What are the chances of it running or what are the chances of it being supported?
I have a feeling that VMS-on-Intel will end up being like Win 2000 DataCenter:
There will be a very
small list of supported systems and options. If you don't have hardware from that
list, then
you don't get support.
It's the same deal as today when Hoff gets a question about running VMS on a
system that
was never meant to run anything other than NT-Alpha: "Yes, VMS may load on that
system but
it never was and never will be supported." You...are...on....your...own.
I predict that the Itanium-based servers will be big expensive hogs just like the
multi-processor
Unisys Intel systems that are currently on the market (the model that Compaq
*used* to sell). After
shelling out $100,000 for one of these beasties, you won't want to put a $19 NIC
in there just to
cut corners. You'll choose a NIC from the small list of NICs that is "approved"
by Compaq.
I doubt that VMS will every run (and be supported) on a regular desktop system
from Gateway. :^)
-Scott :^)
Scott Vieth
It's gonna be a while before they get things just right. You'll be able to
get your money's worth out of the GS system and then buy another AlphaServer
before you are ready to migrate to VMS on Intel.
If you can afford to wait four years for VMS on Intel, then you didn't need the
GS in the first place.
You must have plenty of horsepower already on-hand...
-scott
fabio_...@ep-bc.petrobras.com.br wrote:
> The company I am working for is / was planning to buy a GS machine (not
> sure
> if will be 80 or 160).
>
> I believe the company will wait...
>
> Everybody should wait ! !
>
> Reg
>
> FC
>
> "Fred Kleinsorge" <klein...@star.zko.dec.com> em 25/06/2001 14:05:36
>
> Favor responder a "Fred Kleinsorge" <klein...@star.zko.dec.com>
>
> Info...@Mvb.Saic.Com
>
> Assunto: Re: Full port of VMS to Itanium.
>
> I believe that the VAX port probably was the most tramatic for everyone.
> Going from Alpha to IPF (Itanium Processor Family) will be less tramatic.
> I
> believe that if you have gotten to Alpha, odds are it will be a recompile
> and go.
>
> Were you buying Alpha? Or were you buying VMS? I think that in the grand
> scheme of things, you were buying OpenVMS for the things that it brings to
> the table... if we could build a VAX that was as fast as Alpha - you'd
> still
> be on VAX.
>
> While I am, and remain, an Alpha supporter. You now will have a chance to
> see VMS on a "open" hardware platform.
>
> fabio_...@ep-bc.petrobras.com.br wrote in message ...
> >I believe will happend all that troubles of porting the VAX to Alpha.
> >A lot of companies will not port from Alpha to Itanium.
> >
> >Bad point of view: I believe a lot of customres will jump to long-time
> >processors
> > as SPARC or Power PC.
> >
> >Good point of view: I can improve my value $$$ in the market because
> >OpenVMS
> > professionals will become much more
> >rares .....
> >
> >
> >Regards
> >
> >FC
> >
> >
> >
> >
> >John Macallister <J.Macal...@physics.ox.ac.uk> em 25/06/2001 10:18:02
> >
> >Favor responder a John Macallister <J.Macal...@physics.ox.ac.uk>
> >
> >
> >
> > Info...@Mvb.Saic.Com
> >
> >
> >
> >Assunto: Full port of VMS to Itanium.
> >
> >
> >Today's announcement says that there will be a "Full port of ... VMS .. to
> >Itanium ... starting now." .
> >
> >
> >That, to me, appears to be a fairly positive commitment to the longer term
> >future of VMS.
> >
> >The Powerpoint slides of the announcement are at URL:
> >
> >http://www.compaq.com/hps/ipf-enterprise/download/webcast.ppt
Main, Kerry
>>> Why port to VMS on IA64 when one could port to NT on IA64? <<
Perhaps if one wanted high end clustering and felt simple fail over
clustering was not enough to meet their requirements? For some Customers it
will be enough and it will have a market. For others, these capabilities
will not be enough and will require an OS with more capabilities.
Perhaps if one needed a 64bit OS solution that a long history of proven
64bit based reliability. Keep in mind that the final release date for the
1.0 release of Win64 has still not been determined.
With a semi-common HW platform, the various OS's will have to compete on the
basis of their capabilities.
Obviously, it is not without challenges, but I suspect that given the same
price ranges, and the same HW that is marketed by Intel, OpenVMS will more
than hold its own weight. Obviously, apps play a part in this as well.
Fwiw, and while it is way to early to know if this will be possible, but
under "would be great if ..", I suspect ISV's could be see a benefit in this
as they will have Itanium servers connected to a SAN and they can choose to
boot what ever OS they want that is supported.
Interesting times...
Regards,
Kerry Main
Senior Consultant
Compaq Canada Inc.
Professional Services
Voice: 613-592-4660
Fax : 819-772-7036
Email: Kerry...@Compaq.com
-----Original Message-----
From: Alphaman [mailto:alpha...@nixspam-home.com]
Sent: June 25, 2001 9:07 PM
To: Info...@Mvb.Saic.Com
Subject: Re: Full port of VMS to Itanium.
Fred Kleinsorge <klein...@star.zko.dec.com> wrote in message
news:acKZ6.98$rc5....@news.cpqcorp.net...
> I believe that the VAX port probably was the most tramatic for everyone.
> Going from Alpha to IPF (Itanium Processor Family) will be less tramatic.
I
But traumatic, nonetheless. We are still cranking out both VAX and Alpha
code to this day and into the foreseeable future, still supporting VAX
customers. Now, we have to make it 3 platforms? Why port to VMS on IA64
when one could port to NT on IA64? Contrary to facts, the perception is
that NT is less expensive, and 2000 is "13 times more reliable". And hey,
IA64/Windoze 2000 developers' systems are available in beta today -- why not
get a head start instead of waiting for OpenVMS-IA64?
> While I am, and remain, an Alpha supporter. You now will have a chance to
> see VMS on a "open" hardware platform.
Just how open? Will VMS run on a Dell box, Taiwanese clone, or Q only? And
JF Mezei
> Perhaps if one needed a 64bit OS solution that a long history of proven
> 64bit based reliability. Keep in mind that the final release date for the
> 1.0 release of Win64 has still not been determined.
By the time VMS becomes available to customers on IA64, NT will have have more
maturity and more history on that platform compared to VMS. Same with Linux
and other unix systems.
> With a semi-common HW platform, the various OS's will have to compete on the
> basis of their capabilities.
When you look at Tandem systems, you realise that this won't be the case. For
Tandem to maintain their reliability, their systems may have the CPU based on
IA74, but I would expect their motherboards, bus interfaces etc to be quite
different than commodity hardware in order to acheive the fault tolerance that
is required and expected from those customers.
The big question is whether VMS will be able to run on commodity stuff or if
it will also need to have systems built specifically to run (and boot) VMS.
In my mind, Compaq won't really be saving that much money for these types of
systems because they will still need to be built differently than those built
for simpler OS such as NT and Unix.
> Interesting times...
Correct. We now know not to trust any commitments made by Compaq because they
can go and change them whenever convenient for its parent companies (MS and
Intel). So all the commitments Compaq made for VMS along with those on Alpha
are now meaningless.
Will they complete the port to IA64 ? Probably. Will they market VMS
agressively: probably not.
I think that the Tandem migration will be far more important to Compaq than
the VMS one since Tandem
is the real "mission critical" operating system with lots of visibility,
especially on Wall Street.
Hipenbecker, Doug
Itanium and have the working incarnations of the OS's ready to demonstrate
*BEFORE* announcing the abandonment of the only chip platforms they operate
on...I find it hard to believe that a savvy business decision maker would
throw his "trust" into Compaq's following through on its promise to actually
port the OS's. This is also the reason that this announcement will severly
hurt Compaq in the pocketbook immediately in terms of Alpha and NSK
sales...they will practically vanish...how stupid can a company be?
Doug Hipenbecker
Miller Brewing Co.
-----Original Message-----
From: JF Mezei [mailto:jfmezei...@videotron.ca]
Sent: Monday, June 25, 2001 11:14 PM
To: Info...@Mvb.Saic.Com
Subject: Re: Full port of VMS to Itanium.
Larry Kilgallen
> Wouldn't it have made more sense to have ported OpenVMS, Tru64, NSK to
> Itanium and have the working incarnations of the OS's ready to demonstrate
> *BEFORE* announcing the abandonment of the only chip platforms they operate
> on.
It would seem the money that would have been spent on EV8 will instead
be spent on porting to IA64. Compaq does not have an infinite amount
of money, even though they have more than you or I.
cjt & trefoil
>
<snip>
>
> While I am, and remain, an Alpha supporter. You now will have a chance to
> see VMS on a "open" hardware platform.
>
I seriously question whether Compaq is going to want to see VMS running on,
e.g, Dell hardware.
cjt & trefoil
I think getting a solid compiler could take a while.
Hunter Goatley wrote:
>
> On Mon, 25 Jun 2001 10:41:14 -0300, fabio_...@ep-bc.petrobras.com.br wrote:
>
> >I believe will happend all that troubles of porting the VAX to Alpha.
> >A lot of companies will not port from Alpha to Itanium.
> >
Larry Kilgallen
While hardware revenues are nice, so is software support revenue.
Some companies have a fixation on buying hardware from a particular
manufacturer, and no such rules about software (and other companies
are different, I know).
Would Compaq prefer to get $n from a Dell box or $0. I know that I
prefer $n (for positive values of n).
cjt & trefoil
<snip>
>
> People have been jumping up and down, red-in-the-face, screaming in this
> newsgroup for VMS on commodity hardware. It looks like Compaq has taken
> a HUGE step in that direction and many of the same people are screaming
> "woe is me, the end of VMS is nigh!"
>
It is a little ironic that the particular commodity hardware is the result of
an HP initiative, IMHO.
Larry Kilgallen
> Don't forget that the IA-64 is VLIW, and requires a lot of compiler smarts.
> I think getting a solid compiler could take a while.
>
> Hunter Goatley wrote:
>>
>> On Mon, 25 Jun 2001 10:41:14 -0300, fabio_...@ep-bc.petrobras.com.br wrote:
>>
>> >I believe will happend all that troubles of porting the VAX to Alpha.
>> >A lot of companies will not port from Alpha to Itanium.
>> >
>> As others have said elsewhere, I've ported a large number of programs/products
>> from VAX to Alpha. Once the Alpha compilers had matured some, the porting was
>> trivial for all but the really low-level kernel-mode code that was tired closely
>> to the VAX architecture. I would expect that once Compaq has the GEM compiler
>> back-end generating IA64 code, porting the code should be as simple as a
>> recompile of the code. (GEM didn't exist on VAX, so moving from VAX to Alpha
>> wasn't quite as simple; since all (AFAIK) the Alpha VMS compilers use the GEM
>> back-end, there should be no compiler growing pains in the new architecture,
>> once GEM has been fixed.)
Certainly they must have used GEM in their IA64 experiments for
Tru64/Digital Unix. They must use GEM for the IA32 version of
Compaq Fortran, and they provided some compiler to Oracle for
their experiment at Rdb on IA32. The GEM part seems further
advanced than any other part of the process.
Larry Kilgallen
Imagine how HP feels !!!
cjt & trefoil
Bob Koehler wrote:
>
> In article <OFFB99FC77.F49E9947...@ep-bc.petrobras.com.br>, fabio_...@ep-bc.petrobras.com.br writes:
>
> > Bad point of view: I believe a lot of customres will jump to long-time
> > processors
> > as SPARC or Power PC.
>
> I think both PowerPC and IA-64 will easily outlive SPARC.
>
> ----------------------------------------------------------------------
> Bob Koehler | Computer Sciences Corporation
> NASA GSFC Flight Software | Federal Sector, Civil Group
> | please remove ".aspm" when replying
Larry Kilgallen
> Compaq is much more devious than I ever could have imagined. Does
> this mean we will be running on code developed in China?
>
> http://www.compaq.com/alphaserver/news/idc/20606.htm
>
> Tru64's Future Is on Alpha, Not on Intel's IA64
>
> Compaq recently disclosed its decision to keep the 64-bit Tru64 Unix
> operating system (formerly known as Digital Unix) on Alpha alone. The
> company decided not to "productize" the Tru64 port to Intel's 64-bit
> IA64 microprocessor, although dev elopment on that port will continue
> in China as part of a contractual agreement with a Chinese government
> agency. This decision reduces the number of Unix server operating
> environments (SOEs) that have been announced for IA64 to three:
> Hewlett-Packard's H P-UX, Sun's Solaris, and IBM/SCO/Sequent's Project
> Monterey.
>
> Compaq announced in April that it would support Project Monterey Unix
> on IA64 as an upgrade for the tens of thousands of Compaq ProLiant
> servers sold each year. That move stemmed from the fact that Monterey
> will be the preferred upgrade path for long-t ime users of ProLiant
> servers running SCO UnixWare. IDC believes that Compaq's April
> decision forced the recent Tru64 move on IA64.
>
> Business conditions were very different in January 1998, when the port
> of then-Digital Unix to IA64 was announced as part of Project Bravo by
> Digital Equipment Corp. Once Compaq acquired Digital Equipment in
> mid-1998, it was forced to take a long, hard look at trimming the
> number of overlapping product offerings in its portfolio and reducing
> R&D costs. It would be difficult for one company -- Compaq -- to
> explain why it needed to sell two Unix operating systems for IA64. Had
> that been the case when IA64 shipped, it might have confused
> independent software vendors (ISVs) considering ports to both the
> Monterey/IA64 and Tru64/IA64 platforms.
>
> The decision to halt the Tru64/IA64 marketing effort has the dual
> effect of trimming development costs for Compaq's "value" servers
> while simplifying Compaq's plans for IA64 as a platform for "volume"
> servers running Windows 2000, N ovell NetWare, Linux, and Project
> Monterey Unix.
>
> There is one downside to the end of the Tru64/IA64 port that should be
> noted, however: It effectively leaves more than 5,000 fully 64-bit
> Unix applications temporarily stranded on the Alpha platform.
> Together, those 5,000+ applications running on IA64 would have given
> Compaq an early lead in its planned Tru64/IA64 rivalry with HP, Sun,
> and IBM. Instead, IDC expects that ISVs will recompile many of the
> Alpha-only applications so that they can be moved, over time, to other
> vendors' Unix/IA64 platforms.
>
John Macallister
> That's an interesting remark, but what exactly does Itanic have that Alpha
> does not?
You'll be able to buy an Itanium system from "any computer shop". Who knows
you may even be able to buy a VMS CD set from the shop too? The move to
Itanium opens up a whole new world of opportunity for VMS that was excluded
from the VAX-Alpha setup.
I would like to think this is the end of the beginning for VMS as a World
market player. Now that a sensible approach to hardware has been taken it
will take another leap to provide a "Windows compatible" user interface.
Just as users didn't want weird hardware they don't want software which is
different from everyone else's and that is particularly true of third party
sofwtare vendors.
If the second, software, leap is taken in a timely manner it may herald the
end of Microsoft as the dominant market player. The World may once again
know what it's like to have sofwtare which behaves in a predictable way ...
John
'
Patrick MOREAU, CENA Athis, Tel: 01.69.57.64.40
> As others have said elsewhere, I've ported a large number of programs/products
> from VAX to Alpha. Once the Alpha compilers had matured some, the porting was
> trivial for all but the really low-level kernel-mode code that was tired closely
> to the VAX architecture. I would expect that once Compaq has the GEM compiler
> back-end generating IA64 code, porting the code should be as simple as a
> recompile of the code. (GEM didn't exist on VAX, so moving from VAX to Alpha
> wasn't quite as simple; since all (AFAIK) the Alpha VMS compilers use the GEM
> back-end, there should be no compiler growing pains in the new architecture,
> once GEM has been fixed.)
I'm already thinking of recompiling the DECW archive on VMS IA64 ...
It may be fun.
Patrick
--
===============================================================================
pmo...@cena.dgac.fr (CENA) ______ ___ _ (Patrick MOREAU)
more...@decus.fr (DECUS) / / / / /| /|
CENA/Athis-Mons FRANCE / /___/ / / | / | __ __ __ __
BP 205 / / / / |/ | | | |__| |__ |__| | |
94542 ORLY AEROGARE CEDEX / / :: / / | |__| | \ |__ | | |__|
http://www.ath.cena.fr/~pmoreau/ http://www.multimania.com/pmoreau/
===============================================================================
John Macallister
million
>dollar wildfire system today knowing the architecture is dead ?
The reasons why people bought Alpha yesterday have not changed. It's the
software which is important: people bought Alphas not because they were
Alphas but because they run VMS. Where once there was no long term
commercial future for either Alpha or VMS there is now, with the Itanium
announcement, a whole new horizon in view. We'll have to wait and see of
course but I wonder how long it will be before we see:
iVMS
?
I never did take to OpenVMS.
John Macallister
> recertifying it, and duplicating/updating the support stuff such as
> documentation,
Of course software will have to be recertified as it has to be with every
new release. However, duplicating the support stuff should simply mean
running the docs through a photocopier or, at most, changing OpenVMS to
iVMS. The VMS docs I have are not VAX or Alpha specific: the one set does
for both. The VAX->Alpha transition was less than perfect because all VAX
devices were not taken over to Alpha and there were some unnecessary VMS
changes in the transition (e.g. SYSGEN AUTO ALL to SYSMAN IO AUTO ) which
made some people believe that AlphaVMS and VAX/VMS were different OS's.
Some people believed that a VAX-Alpha transition would be a major effort.
For most people the effort involved was close to zero and most ordinary
users couldn't tell whether they were on a VAX or an Alpha.
I think the people who worry so much about the Alpha-Itanium transition
would probably also have spent some time trying to tabulate the differences
between VMS and OpenVMS. By the way, as a marketing shot in the launch of
the Alpha-Itanium transition, Compaq are running a competition to see
whether anyone can find more differences between VMS and iVMS than between
VMS and OpenVMS. The prize, I believe, is a kit which can be applied to any
version of the OS which makes user software compatible across
VMS-OpenVMS-iVMS. Details are on the Compaq Web site but may be hard to find
at the moment until the port to iVMS is complete. I understand that Compaq
will apply the kit to their Web site so that "page not found" errors will be
a thing of the past as there will be only one version of everything relating
to VMS!
paddy....@zzz.tg.nsw.gov.au
>>
><snip>
>>
>> While I am, and remain, an Alpha supporter. You now will have a chance to
>> see VMS on a "open" hardware platform.
>>
><snip>
>
>I seriously question whether Compaq is going to want to see VMS running on,
>e.g, Dell hardware.
It has been seriously suggested that Compaq is no longer interested in seeing
VMS running -- or Tru64. The time frames suggest that they are after a
death, and there is already a thread about a lost contract/sale.
My greatest sympathy goes to people like Fred and Hoff. The pay-packet is
still there, but to work on a project that will die?? Many, many years ago I
did that, and it is soul destroying. Isn't that a reason for Dave Cutler
going to Micro$oft?
Over on comp.lang.fortran, Steve Lionel is expressing similar platitudes.
The personalities of the three mentioned (and others we know) are such that
they want to look after their customers, but how much will they be
ham-strung.
SL has said that of the compiler teams to go to Inhel, Fortran will be one of
the first. Though I'm not a PC person, DVF -> CVF -> IVF??? A problem that
(if I read him correctly) Steve has is how much they will be able to support
their other platforms VMS and Tru64, perhaps even windows in the transition
(definitely my conjecture here re windows support).
Regards, Paddy
paddy....@zzz.tg.nsw.gov.au
>prefer $n (for positive values of n).
Especially for large values of n :-) Something like the old saw (in varying
combinations) 2+2 = 4, even for large values of 2.
Regards, Paddy
Carl Perkins
}I believe that the VAX port probably was the most tramatic for everyone.
}Going from Alpha to IPF (Itanium Processor Family) will be less tramatic. I
}and go.
}
}Were you buying Alpha? Or were you buying VMS? I think that in the grand
}scheme of things, you were buying OpenVMS for the things that it brings to
}the table... if we could build a VAX that was as fast as Alpha - you'd still
}be on VAX.
}
}see VMS on a "open" hardware platform.
So are you saying that we will be able to buy VMS to run on Dell
IA64 systems?
If not, then it isn't an "open" harware platform - just another
proprietary hardware platform that happens to use a CPU made by
Intel instead of Compaq.
If so, then the "Open" in OpenVMS will be considerably more meaningful.
--- Carl
Carl Perkins
}>That's an interesting remark, but what exactly does Itanic have that Alpha
}>does not?
}
There is no evidence of this. Intel has killed off a number of other
processors in the past. What heppens to VMS if that happens? How
fast can you port VMS to whatever it is they might replace it with?
}Look. I am an Alpha supporter. I think that EV68 is the best processor on
}the market today. I think EV7 will be a kick-ass chip, and system. The
}Itanium has nothing on it. Until recently, there was a strong question of
}if they would *ever* get IA64 to work well enough to be interesting. But
}the recent benchmark numbers show that they willl be in the ballpark soon.
In the ballpark of what? Today's Alpha or the (what would have been) the
Alpha of the same timeframe?
I seriously doubt that the IA64 processor that will be current when EV8
would have show up will be in the same ballpark for performance.
Making the IA64 faster is not so easy - EPIC is a seriously iffy
architecture to bet your future on.
}So, project it out. Sometime in the mid-late 2000's, IA64 is very
}competetive with Alpha on performance, but with HP and Microsoft on it, it
}sells many times more than Alpha. The price is lower than Alpha. Not only
}that, we don't have to build two seperate lines of HW for IA64 and Alpha.
}Lots cheaper. Heck, *maybe*, just *maybe* it means that if we do it right,
}we could put VMS on any Itanium box that we want to qual it on, including
}the "cheap" ones too.
The problem is that if consolidating down to one type of hardware is
good, what makes you think that the people running your company won't
decide that consolidating down to one OS isn't just as good? Instead
of having NT, VMS, True64, Linux, and NSK, why not just NT?
And BTW - the price is not lower than Alpha. At least, not yet. You
can actually buy a DS10 running VMS for what it costs to get an Itanium
based system running Windows. Yeah, the DS10 is slower - lots slower
at floating point and somewhat less slow at integer (the Itanium's
integer performance truely stinks) - but that is because Compaq doesn't
do upgrades. Comapq should sell a DS10 with at least an 833MHz Alpha
and it should have the 1GHz version available from day one for that chips
availability. But it won't for purely marketing reasons. Likewise I doubt
that the low end VMS system will see the fast upgrades that a PC sees
(where the processor comes out and you can just use it) once it is IA64
based either. The reason is the very same marketing reason: avoiding
having the high end system sales canibalized by the low end systems.
}I'm pretty sure we can pull thins thing off. It's just a matter of how
}easy/hard it will be.
I'm certain that it is possible. I am just about as certain that
is is not a good idea.
I was considering buying some Compaq stock as it was looking very
cheap. I certainly will not be doing so now.
--- Carl
Brian Schenkenberger, VAXman-
>In article <9h8bl3$o...@gap.cco.caltech.edu>,
>David Mathog <mat...@seqaxp.bio.caltech.edu> wrote:
>>[snip]
>>
>> ***COMPAQ MANAGEMENT WANTS THIS PORT TO FAIL***
>>
>
>That's as absurd as anything that's been posted here recently. And that's
>saying a LOT.
As absurd as killing off Alpha and jumping on the me too intel bandwagon?
I don't see the future well for this Itanium processor with respect for
OpenVMS. Compaq has given away key Alpha technology and intel will in-
corporates pieces which fill in the gaps and holes in their sieve design.
When Compaq requires a certain enhancement to make an OpenVMS port a re-
ality, will intel put on the brakes and introduce these enhancements? I
see, if the aformentioned happens, other vendors having their way too and
eventually the Itanium will become the equivalent Billybloatware in Sil-
icon. Let's just call it Titanium now hope that there's a lifeboat that
can rescue us..
>Compaq being unable to manage a port to the processor that it looks like
>everyone will be supporting in the future (see recent HP announcement, heck
>even Sun has hedged their bets by having a working Solaris port _today_)
>will not bode well for their ability to sell into those high-margin
>Enterprise accounts that they are obviously trying to nurture with
>their entire new strategy.
Compaq has a strategy?
>People have been jumping up and down, red-in-the-face, screaming in this
>newsgroup for VMS on commodity hardware. It looks like Compaq has taken
>a HUGE step in that direction and many of the same people are screaming
>"woe is me, the end of VMS is nigh!"
My feet have been planted firmly on the ground both figuratively and
in reality. I had and still have great faith in Alpha. Digital did
it right in addressing the needs of OpenVMS, OSF-1 and that other M$
thing. It was simple and the design was clean. It improved on other
RISC technologies by observing other RISC designs and opting out of
the implementations in their designs which hindered really stream-
lining the Alpha for performance. I want OpenVMS on a solid, sound
bit of engineering so that it is stable and reliable and not neces-
sarily cheap. I have a bit of the "commodity hardware" you speak of
in my home (sprinkling holy water over the demon's seed as I mention
it) and it is so "solid" that it sounds like a popcorn maker when it
is in use.
--
VAXman- OpenVMS APE certification number: AAA-0001 VAXman(at)TMESIS(dot)COM
city, n., 1. a place where trees are cut down and streets are named after them.
Vance Haemmerle
>
> On Mon, 25 Jun 2001, Robert Deininger wrote:
>
> > Competent management and marketing. But alpha has had the best engineers.
> >
>
> Alpha's Omega was rendered inevitable by Digital's outright marketing
> malfeasance and a string of Stupid Strategy Tricks. Compaq's initial
> indecisiveness and tepid support for Alpha (and the support was far less
> than tepid amongst the Compaq Classic box-pushers) delayed a turn-around
> in the business, which actually has grown during the past two years.
>
> Too little, too late.
Compaq is a company in search of something they do well. They
couldn't
enhance the value of the Altavista search engine, so they sold it. No
longer tops in PC manufacturing, they bought several Enterprise
companies
to change the company's vision. Failing to do that well they now plan
to
buy several software and service companies... maybe they can do
something
with them. You'd think with all the restructurings, they'd at least
be
good at them by now!
Message to Michael Capellas: Congratulations, you've won the Robert
Palmer "Help, we've lost our core compentecy and we can't get up"
award!
Compaq is the next Lucent.
--
Vance Haemmerle Internet va...@toyvax.Tucson.AZ.US
Tucson, AZ Web
http://toyvax.Tucson.AZ.US/~vance/
Vance Haemmerle
> Well, on the bright side, if there's a low power Itanic being planned, we
> may see another VMS-capable notebook.
Do you REALLY think Compaq will allow OpenVMS on commodity
hardware? Think about it, remember the Alpha white boxes meant to
run Windows NT and how they wouldn't run OpenVMS? Compaq will most
certainly have two Itanium server lines, a commodity-priced one for
Windows NT-based systems and one that will be able to run OpenVMS
for substantially more money. After all, that would mean competition
with Microsoft. My other prediction: OpenVMS and Tru64 will not be
able to run on Itanium servers of other manufacturers. How much
support was there for Alpha clones? Perhaps OpenVMS systems will require
some special ROM, ala the old Mac OS ROM, I'm sure they'll think of
something.
OpenVMS engineering will never be given the resources to qualify commodity
peripherals and graphics. In the windows way, the manufacturers of the
peripherals and graphics write their own drivers... do you think they'll
write one for VMS? OpenVMS on a laptop again? Dream on.
Without the hardware subsidy on OS development, how much more do you
think people will have to pay for OpenVMS?
Brian Schenkenberger, VAXman-
>>Fred Kleinsorge wrote:
>>>
>><snip>
>>>
>>> While I am, and remain, an Alpha supporter. You now will have a chance to
>>> see VMS on a "open" hardware platform.
>>>
>><snip>
>>
>>I seriously question whether Compaq is going to want to see VMS running on,
>>e.g, Dell hardware.
>
>It has been seriously suggested that Compaq is no longer interested in seeing
>VMS running -- or Tru64. The time frames suggest that they are after a
>death, and there is already a thread about a lost contract/sale.
...and for those of us here making our living off of VMS. It's getting
harder and harder to mount this now lame once mighty and magestic steed
named VMS.
Career change time?
Vance Haemmerle
>
> In article <9h8eca$5vt$1...@lisa.gemair.com>, jor...@lisa.gemair.com (Jordan Henderson) writes:
> >Compaq being unable to manage a port to the processor that it looks like
> >everyone will be supporting in the future (see recent HP announcement, heck
> >even Sun has hedged their bets by having a working Solaris port _today_)
> >will not bode well for their ability to sell into those high-margin
> >Enterprise accounts that they are obviously trying to nurture with
> >their entire new strategy.
>
> Compaq has a strategy?
Lets port Tru64 to IA-64. No wait, lets not port Tru64 to IA-64. Hey,
lets port Tru64 to IA-64 again! Compaq strategy at its best!
mist dragon
> In comp.unix.tru64 article <2d9340de.01062...@posting.google.com>, morto...@hotmail.com (John Morton) writes:
>
> > Compaq recently disclosed its decision to keep the 64-bit Tru64 Unix
> > operating system (formerly known as Digital Unix) on Alpha alone. The
> > company decided not to "productize" the Tru64 port to Intel's 64-bit
> > IA64 microprocessor, although dev elopment on that port will continue
> > in China as part of a contractual agreement with a Chinese government
> > agency. This decision reduces the number of Unix server operating
> > environments (SOEs) that have been announced for IA64 to three:
> > Hewlett-Packard's H P-UX, Sun's Solaris, and IBM/SCO/Sequent's Project
> > Monterey.
> >
Interesting notion - when I recieved the news and this one, I thought
to myself of how close Itanium port already is ? Afaik, Tru 64 already
has the possibility to run Linux applications with compatibility lib.
I suppose Linux could have similar Tru 64 compatibility box or
libraries and a lot of porting would be saved. For Tandem it could be
the same - writing device drivers and providing special kernel
services. And who knows - would it be possible to have a VMS emulator
on Linux like Alpha Nt had Intel emulator. HP 3000 was an emulator
over HP-UX when 3000 was moved to risc. Apple has 68k emulator on
risc.
In my experience of porting between different unix systems, in reality
the differences are really small unless you have a different bit order
and have used structs with exactlength digits. Code made with VMS c
compiler and then ported to risc machine like Aix crashes easily
because VMS c compiler is 'smart' and corrects the mistakes
automatically whereas Aix thinks that 'if you made it, you mean it'.
So I used the develop in aix and then port to vms because in that way
the result was easily guaranteed. The rest of the code not doing
anything with stucts should be really easy.
Now, if we compare tru 64 and linux, the differences must be small. So
in principle what I have described should be possible (of course I am
talking about IA64 port of Linux). For openvms - if they support only
the compability without really booting to it - again it could be done.
steven...@quintiles.com
I doubt that the Tandem customers would be happy with that method. The
present architecture does everything in a lock step technique such that at
least two processors do the same job at any one time. If one processor
gives a wrong answer then it is failed out.
Tandem takes the must not fail idea to the extreme which is why customers
use it in their specific applications. It also extends to disks which use
514 byte sectors rather than the usual 512 as the two extra bytes contain
parity information for the 512 bytes of real data.
Steve.
Bob Koehler
> I believe that the VAX port probably was the most tramatic for everyone.
> Going from Alpha to IPF (Itanium Processor Family) will be less tramatic. I
> believe that if you have gotten to Alpha, odds are it will be a recompile
> and go.
I don't think so. Many of the basic architectural concepts of Alpha are
compatable with VAX, and the Alpha was designed with the knowledge that
VMS would have to run on it. This is not true for IA-64.
Bob Koehler
>
> You'll be able to buy an Itanium system from "any computer shop".
Yes, but will VMS run on an "any shop" Itanium? And will EDT be useable
on and "any shop" keyboard, or will we still need three fingers to hit
"gold"?
Bob Koehler
As long as they're doing anoth port, how about confusing the world with
another name change?
VAX/VMS
VAX-VMS
VAX VMS
OpenVMS VAX
OpenVMS Alpha AXP
OpenVMS Alpha
How about just plain "VMS" ?
Alan B.
** What do I tell my boss now about my plan for migrating our old VAX
apps to Alpha, as opposed to a PC OS as has happened to other apps? He
was quite receptive and told me to go ahead. Now that Alpha's future
(and what would be required to move to Itanium) is uncertain, this
puts quite a damper on things. ** I feel somewhat screwed by CPQ!
William...@acml.com
How about
OpenVMS Omega
(tongue firmly planted in cheek)
koehler@encompa
sserve.org To: Info...@Mvb.Saic.Com
cc:
06/26/2001 Subject: RE: Full port of VMS to Itanium.
09:32 AM
______________________________________________________________________
The information contained in this transmission may contain
privileged and confidential information and is intended only
for the use of the person(s) named above. If you are not the
intended recipient, or an employee or agent responsible for
delivering this message to the intended recipient, any review,
dissemination, distribution or duplication of this communication
is strictly prohibited. If you are not the intended recipient,
please contact the sender immediately by reply e-mail and destroy
all copies of the original message.
John Macallister
>Yes, but will VMS run on an "any shop" Itanium? And will EDT be useable
>on and "any shop" keyboard, or will we still need three fingers to hit
>"gold"?
Why should one shop's Itanium be different from another's re the running of
VMS any more than with Windows? Windows runs on any shop Pentiums,etc
because so many people use Windows and the market makes it worthwhile to
cater for different hardware configurations. iVMS could be in a similar or
better position.
I usually only have to hit one key for "Gold" when using EDT through my
Windows PC running Exceed. When I occasionally find another product where
one finger doesn't work I don't bother going to three fingers as I find that
two fingers, not pointing at the keyboard, are sufficient!
Dan O'Reilly
> > > You'll be able to buy an Itanium system from "any computer shop".
>
> >Yes, but will VMS run on an "any shop" Itanium? And will EDT be useable
> >on and "any shop" keyboard, or will we still need three fingers to hit
> >"gold"?
>
>Why should one shop's Itanium be different from another's re the running of
>VMS any more than with Windows? Windows runs on any shop Pentiums,etc
>because so many people use Windows and the market makes it worthwhile to
>cater for different hardware configurations. iVMS could be in a similar or
>better position.
Why would one shop's alpha run NT and not VMS? Console/microcode support.
------
+-------------------------------+---------------------------------------+
| Dan O'Reilly | |
| Principal Engineer | "Why should I care about posterity? |
| Process Software | What's posterity ever done for me?" |
| http://www.process.com | -- Groucho Marx |
+-------------------------------+---------------------------------------+
John Macallister
Microcode can be written. It's a one-off job for each variation of hardware
type. The reason some shops did not write in VMS console support was that
they did not see a market for VMS on Alpha. iVMS on Itanium, or any Intel
processor, would offer the potential of significant returns for the
investment in microcode. The significant returns won't appear, of course,
until there's good, compatible application software available. The next few
months may interesting as ASP's comment on iVMS ...
John W. Hom
> > > You'll be able to buy an Itanium system from "any computer shop".
> VMS any more than with Windows? Windows runs on any shop Pentiums,etc
> because so many people use Windows and the market makes it worthwhile to
> cater for different hardware configurations. iVMS could be in a similar or
> better position.
I think I remember reading somewhere that the Pentium processors
have two
bits for privilege, but was hobbled to one bit in order to work
with the
more common OSes which use User/System modes. If Compaq is
indeed intent on
porting VMS to the IA64, pressure must to placed to keep the four
execution
modes required (User/Supervisor/Executive/Kernel). I don't know
if the
motherboards and the various other cards need to be changed, but
I'm sure
a physical jumper could be placed somewhere that would
enable/disable the
extra bit.
It would be interesting to see what becomes of this project. If
they are
successful and do port the OS to IA64, I'm building my own box,
and for
the first time, I'll be able to run VMS at home. If not, it's
probably
the end.
John
--
John W. Hom
j.h...@alumni.nyu.edu
Beware of all enterprises that require new clothes.
-- Henry David Thoreau
Mark D. Jilson
knows what Intel will eventually deliver post IA-64 as we all know that
IA-64 in it's current form won't be good enough to use.
Jordan Henderson wrote:
> <snipped a bunch>
> People have been jumping up and down, red-in-the-face, screaming in this
> newsgroup for VMS on commodity hardware. It looks like Compaq has taken
> a HUGE step in that direction and many of the same people are screaming
> "woe is me, the end of VMS is nigh!"
>
--
Jilly - Working from Home in the Chemung River Valley - Lockwood, NY
- ji...@clarityconnect.com - Brett Bodine fan
- Mark....@Compaq.com - since 1975 or so
- http://www.jilly.baka.com -
Dan O'Reilly
> > Why would one shop's alpha run NT and not VMS? Console/microcode support.
>
>Microcode can be written. It's a one-off job for each variation of hardware
>type. The reason some shops did not write in VMS console support was that
>they did not see a market for VMS on Alpha. iVMS on Itanium, or any Intel
>processor, would offer the potential of significant returns for the
>investment in microcode. The significant returns won't appear, of course,
>until there's good, compatible application software available. The next few
>months may interesting as ASP's comment on iVMS ...
But the point is, a HUGE reason for VMS' success is because the hardware
platforms were predictable and well-qualified. You won't (I strongly hope)
be able to go into your local CompUSA (or its equivalent overseas) and buy
VMS off-the-shelf, it just doesn't make any sense to even consider that.
David Mathog
>Wouldn't it have made more sense to have ported OpenVMS, Tru64, NSK to
>Itanium and have the working incarnations of the OS's ready to demonstrate
>*BEFORE* announcing the abandonment of the only chip platforms they operate
>on...
Yes, it would have. Either Q management really has no intention of
completing these ports (that is, they're intentionally torching these
businesses for who knows what twisted and perverse MBA reason, and these
"ports" are designed to falsely reassure the remaining customers so that
the Q can milk the last few pennies out of them before selling those
customers' souls to CA) or they don't have any institutional memory of the
VAX->Alpha transition.
>I find it hard to believe that a savvy business decision maker would
>throw his "trust" into Compaq's following through on its promise to actually
>port the OS's. This is also the reason that this announcement will severly
>hurt Compaq in the pocketbook immediately in terms of Alpha and NSK
>sales...they will practically vanish...how stupid can a company be?
Michael Crichton coined this term for the entertainment industry but
it seems appropriate to apply it here as well: fabulously stupid.
Regards,
David Mathog
mat...@caltech.edu
Manager, sequence analysis facility, biology division, Caltech
**************************************************************************
* RIP VMS & ALPHA *
**************************************************************************
John Macallister
>be able to go into your local CompUSA (or its equivalent overseas) and buy
>VMS off-the-shelf, it just doesn't make any sense to even consider that.
I strongly hope people WILL be able to into any computer shop and purchase
hardware and an iVMS CD kit as easily installable as the other OS's
available in the shop. People should not have to be computer experts to
install and own an iVMS PC.
There will, of course, be "enterprise systems" and "SERVERS" which may run
specialised hardware or software requiring significant specialised support.
The home market with iVMS and the server market with sVMS/eVMS (?) should
support each other.
We all have to get away from (a) the doom and gloom culture at Alpha
reaching the end of its development cycle and (b) the mean culture of not
wanting to share iVMS with the rest of the World lest we end up not being
special in some way.
The rest of the World is now going to look at news groups like this for
impressions of iVMS and so we must attempt to have positive discussions
looking forward to the new horizons and opportunities the full port to
Itanium can offer. If all they see are people jumping or threatening to jump
ship nobody will come aboard and iVMS will be scuttled by those already on
board and not sunk by an indifferent World.
iVMS : let's get on with it ...
Dan O'Reilly
> >You won't (I strongly hope)
> >be able to go into your local CompUSA (or its equivalent overseas) and buy
> >VMS off-the-shelf, it just doesn't make any sense to even consider that.
>
>I strongly hope people WILL be able to into any computer shop and purchase
>hardware and an iVMS CD kit as easily installable as the other OS's
>available in the shop. People should not have to be computer experts to
>install and own an iVMS PC.
Agreed. However, just one thing's missing: software. Just how many people
are going to give up their Microsoft stuff so they can run right out and buy
VMS without it? For that matter, what system have more people heard of?
VMS or Windows? Who has the track record? VMS or Windows? It doesn't matter
HOW technically superior VMS is to Windows, very few people who want to run
a spreadsheet give a damn about that. There is literally ZERO window of
opportunity to make VMS a system for the masses. It simply can't compete.
The other concern, of course, is that if Joe Everyman can walk into a store
and buy his VMS system off the shelf, and it doesn't work with his (disk/tape/
scanner/digital camera/controller/whatever-widget-you-want), that starts bad
press about "look how lousy the support for this is". And eventually that word
gets back to the IT manager who's looking at buying VMS: "hmmmmm...I've heard
that this thing just isn't reliable anymore, better buy something else". And
it won't matter that for HIS application it would be just fine. Perception is
reality. I know, I ran up against countless managers and beancounters with
that mentality at MCI, BT, whatever you wish.
>iVMS : let's get on with it ...
Yea, verily, brother!
Robert Deininger
<jfmezei...@videotron.ca> wrote:
> It was bad enough that Digital
> failed in its transition of VAX->ALPHA (the fact that so many vaxes remain in
> production 10 years after is a testament to that failure, with demand for new
> VAXes continuing),
Nope. Now you go too far. Digital's migration from vax to alpha was a
success. 95% of user-mode code was compile-and-go. The VMS tranistion was
easy for most of us. (I leave out things like marketing and product
development, which stank.) All those vaxes still in service are a
testament to the quality of their design and construction. They weren't
made to be disposable.
> You could hire all the Alpha engineers. I bet and I hope that many of them
> will refuse to work for Intel.
Hoping for that is definitely cutting off your nose to spite your face.
Those engineers are the best chance to make a future intel CPU that is
worth cheering for.
Many of the engineers may not like having their leases yanked in a new
direction, and may leave. But Intel is quickly becoming the only place to
work if you want to design state of the art CPUs for a living. I expect
IBM will make a bid for these folks, but beyond that they don't have many
options.
Perhaps Intel will bend over backwards to make these people happy. They
are among the best in the world. If there were lots of people this good
available, intel would have already hired them to help with the
Unobtanium.
The engineers and intel may have to stay together because of mutual necessity.
--
Robert Deininger
rdein...@mindspring.com
Robert Deininger
(Hunter Goatley) wrote:
> As others have said elsewhere, I've ported a large number of programs/products
> from VAX to Alpha. Once the Alpha compilers had matured some, the porting was
> trivial for all but the really low-level kernel-mode code that was tired
closely
> to the VAX architecture. I would expect that once Compaq has the GEM compiler
> back-end generating IA64 code, porting the code should be as simple as a
> recompile of the code. (GEM didn't exist on VAX, so moving from VAX to Alpha
> wasn't quite as simple; since all (AFAIK) the Alpha VMS compilers use the GEM
> back-end, there should be no compiler growing pains in the new architecture,
> once GEM has been fixed.)
Evidently there are various generations of GEM, and some compilers seem
stuck on older ones. Compaq Ada comes to mind. I guess plugging in a
newer GEM isn't completely trivial, or there wouldn't be stranded
compilers.
I hope the IA64 version of GEM is available as a "plug-in", perhaps under
license, with a documented interface. That would be better than "internal
use only". But I guess GEM goes to intel along with alpha. Compaq's
software will no longer have the advantage of the best compiler
technology. Intel will give everyone equal access.
> That's what I hope, anyway.....
So do I.
--
Robert Deininger
rdein...@mindspring.com
John Macallister
It's clear that iVMS will have little impact without user software and that
software must be compatible (i.e. same GUI) with that on Windows. People
just don't want another different set of packages: they want the same
package on Windows or iVMS.
A massive surge of enthusiasm from this news group, DECUS, VMS lobbyists
everywhere is now needed to ensure that there is sufficient momentum to
convince third party vendors that it will be worthwhile porting packages to
iVMS.
Let's go trekking ...
Bob Kaplow
> Think of it as a truck, it is the load that you are carrying that is
> important,
> provide of course that the truck performs to specs. You just traded in an
> old on for a newer one, which will at some point be replaced again, and
> MAYBE it won't have such severe alignement requirements. That's how you
> sell it and keep your job. :->
Except in this case, the truck is a Ford Explorer, and the and the out of
allignment tires are Firestone ATX. Looks like a roll-over waiting to
happen.
Reminds me of a line in an internal DEC notes file just before all the
layoffs started in the early 90s:
Downsize, Rightsize, Capsize.
Bob Kaplow
> For the next several years, Alpha is arguably the best HW available.
> Remember, we will continue and implement the EV7 family and system
> platforms. If you get from VAX to Alpha, the "port" to IPF should, we
> believe, be a recompile & relink. It is also a 99.8% probability that there
> will be mixed cluster capability.
Why would anyone looking to build something new today pick a platform where
the hardware is at a declared end-of-life already and will require an
expensive port in a couple years. Hell, if I wanted to go through all of
that, I'd be using HP-UX!
Bob Kaplow
> Nothing in OpenVMS, or Galaxy, is Alpha-specific, except perhaps in
> implementation. With the Alpha design team going to Intel, I would not be
> suprised to see things from EV7 and EV8 sneak into Itanium in the future.
Will any of them go? And wouldn't Intel have just stolen the technology like
they did earlier with everything else.
> We're pretty confident that by the time we get things ported, it will be in
> good shape, and EV7 systems will give us a good transition padding with a
> high-performance Alpha swan-song.
>
>>Andrew Harrisson, is Sun hiring ?
>
> Sun is trying not to lay off.
But Andrew and his buddies will FUD Compaq, VMS, and Tru64 to death before
the first Itanium / VMS system is sold.
Bob Koehler
> I usually only have to hit one key for "Gold" when using EDT through my
> Windows PC running Exceed. When I occasionally find another product where
> one finger doesn't work I don't bother going to three fingers as I find that
> two fingers, not pointing at the keyboard, are sufficient!
My Eve keypad uses all the LK401 keypad keys (including the missing "," on
the PC), and F17 - F20. I've got everything except F17 - F20 on a Mac,
but on PC's I've never found ",". I've not yet found one PC based
terminal emulator which I'm happy with, despite having tried many
recommended here in c.o.v.
JF Mezei
> VAX/VMS
> VAX-VMS
> VAX VMS
> OpenVMS VAX
> OpenVMS Alpha AXP
> OpenVMS Alpha
>
> How about just plain "VMS" ?
VMS's name should be decided by the new onwer of VMS when Compaq finally
decides to sell VMS.
Perhaps SUN-VMS ? IBM-VMS ?
While we're at it, why not port VMS to the IBM 390 architecture , and have VMS
run as a partition in a large sysplex ?
David J. Dachtera
> [snip]
> My Eve keypad uses all the LK401 keypad keys (including the missing "," on
> the PC), and F17 - F20. I've got everything except F17 - F20 on a Mac,
> but on PC's I've never found ",". I've not yet found one PC based
> terminal emulator which I'm happy with, despite having tried many
> recommended here in c.o.v.
What do you find lacking in Reflection/2 or /4 (aside from
affordability, I mean)?
--
David J. Dachtera
dba DJE Systems
http://www.djesys.com/
Unofficial Affordable OpenVMS Home Page and Message Board:
http://www.djesys.com/vms/soho/
This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings
is to be expected.
Feel free to exercise your rights of free speech and expression.
However, attacks against individual posters, or groups of posters, are
strongly discouraged.
JF Mezei
> Why should one shop's Itanium be different from another's re the running of
> VMS any more than with Windows? Windows runs on any shop Pentiums,etc
You'll find that when you buy Compaq gear, you'll most likely be getting a
version of Windows that has been customized by Compaq with its own little
special drivers for the various devices on it. At at least that was the case
in the past.
In terms of VMS on IA64, will Compaq provide the >>> prompt on all its Intel
machines from now on ? I doubt it.
Bill Gunshannon
Dan O'Reilly <da...@process.com> writes:
|> At 08:46 AM 6/26/2001, John Macallister wrote:
|> > >You won't (I strongly hope)
|> > >be able to go into your local CompUSA (or its equivalent overseas) and buy
|> > >VMS off-the-shelf, it just doesn't make any sense to even consider that.
|> >
|> >I strongly hope people WILL be able to into any computer shop and purchase
|> >hardware and an iVMS CD kit as easily installable as the other OS's
|> >available in the shop. People should not have to be computer experts to
|> >install and own an iVMS PC.
|>
|> Agreed. However, just one thing's missing: software. Just how many people
|> are going to give up their Microsoft stuff so they can run right out and buy
|> VMS without it?
But this is solvable and Compaq has even expressed an interest in
changing this recently. I would assume anything at the application
level that gets ported to VMS on the Alpha today would require little
more than a re-compile to move onto VMS on the IA64. The only possible
problems might be differences in the size of data types and one could
hope that the compiler might flag these as compatability problems.
|> For that matter, what system have more people heard of?
|> VMS or Windows? Who has the track record? VMS or Windows? It doesn't matter
|> HOW technically superior VMS is to Windows, very few people who want to run
|> a spreadsheet give a damn about that. There is literally ZERO window of
|> opportunity to make VMS a system for the masses. It simply can't compete.
Boy, and I thought I was a cynic. Why is something as badly written, as
unstable and as unreliable as Linux taking over the world by storm??
Marketing.
Why are the BSD's which are so much better from the base design on up
not in the same position??
Lack of marketing.
What would it take to make VMS a mainstream OS??
First, it has to run on the same hardware as everything else, with
no price premium required.
Second, it has to be priced in the same catagory. (Yes, Linux and BSD
are free, but service and support cost real money. I am not saying
VMS must be free, but it has to be no more expensive than Win2000.)
Third and most important, it has to be taken out from under the
bushel basket and placed high oh a pedestal where everyone can see it.
Marketing, marketing, marketing.
It can be done.
|>
|> The other concern, of course, is that if Joe Everyman can walk into a store
|> and buy his VMS system off the shelf, and it doesn't work with his (disk/tape/
|> scanner/digital camera/controller/whatever-widget-you-want), that starts bad
|> press about "look how lousy the support for this is". And eventually that word
|> gets back to the IT manager who's looking at buying VMS: "hmmmmm...I've heard
|> that this thing just isn't reliable anymore, better buy something else". And
|> it won't matter that for HIS application it would be just fine. Perception is
|> reality. I know, I ran up against countless managers and beancounters with
|> that mentality at MCI, BT, whatever you wish.
And how is this scenario any differnt for Windows boxes?? My boss has gone
through three different CD/RW drives, including a major brand name (HP) and
three differnt software packages and has yet to find one that isn't better
at making coaters than CD's. We upgraded our labs to Win2K from Win98. Do
you realize how much hardware that obsoleted?? Printers, scanners, CD/RW's,
webcams, tape backup units, etc. And the same is true for Linux and the BSD's.
While it is noce to think that these OSes miraculously support everything,
it just isn't true. First of all, the hardware is changing at a rate so
high no one could keep up with it. But if places that sell VMS sell it as
a package where they bundle it with hardware from the supported list, then
the problem is minimized. And as it becomes more mainstream, vendors may
even start taking it into consideration when they design new hardware, much
like they do for Linux today. Maybe even providing VMS drivers!!
|>
|> >iVMS : let's get on with it ...
|>
|> Yea, verily, brother!
It's one thing to be negative about what Compaq has done, but it
doesn't mean we have to give up on VMS entirely. In the long run,
it is going to how many people actually decide to use VSM that is
going to make this port sink or swim. I think we have brought most
of the issues out in the open, now it's time to start looking at
what it is going to take to resolve them.
bill
--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bi...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>