- OpenVMS ever to be on Intel chip?
Terry C. Shannon
"Darren Peacock" <daz...@hotmail.com> wrote in message
news:FsEq6.10155$0N3....@news-server.bigpond.net.au...
> Cant remember the detail but back in 1988-89 there was a project in
Digital
> to do just that ..
> My memory is fuzzy but it was called Emerald or Gem ..
> It was descibed as VMS on Intel.
>
> But the initial downsizing then , the project was halted. Funny to see the
> following 24 months some of the members end up in a small word processing
> company called Microsoft.
>
Well, I dragged Charlie Matco away from his usual pursuits of wine, women,
and rumourmongering and persuaded him to come clean on this matter.
There was in fact a DEC project called EMERALD back in the late 1980s. Its
goal: to boldly send VMS where it had not gone before... into IA32-land.
EMERALD was scuttled right around the same time the PRISM RISC project was
killed (late March 1988). Prototype PRISM processors existed at the time,
but apparently there were Big Delays with the complementary MICA operating
system. MICA was, simply stated, a reimplementation of VMS for the PRISM
RISC architecture. Dave Cutler flew the coop right after Ken Olsen pulled
the plug on PRISM/MICA. Word has it that a lot of the MICA code rose from
the dead when Windows NT was born.
Separately, there was a midnight project to port VMS to the Mach kernel. The
project was done (half-baked, actually) at Carnegie Mellon University IIRC.
Some of the incomplete code--which may well have a few facets of Emerald
embedded in it--is said to be floating around somewhere.
And that's all I got from Charlie. Heck, he started mumbling some stuff
about OZIX, MERLIN, QUARTZ, CHEYENNE, and other cryptic codewords from days
gone by. Wish he'd be a tad more forward-looking and spill the beans about
MARVEL... and the system a generation beyond MARVEL.
cheers,
Matco's Handler
daytripper
Um...Don't know how to tell Charlie this, but the Emerald program was to
produce a risc-based 32-bit micro in Hudson. And Cutler's mission in Belvue
was developing "Crystal" - a 64bit 8-way risc machine built out of NEC ecl
gate arrays and AMCC ecl/ttl mixed-tech ASICs.
These were "VMS-on-RISC" programs, nothing to do with IA32.
These were killed - along with the Argonaut program out of Littleton in favor
of the $3B disaster known as Aquarius (aka VAX9000), out of Marlboro.
Had either Crystal or Argonaut (both of which were on the virge of first
silicon) seen the light of day, things might have been different. Instead, in
managing to piss off most of the best talent it had in favor of Glorioso's own
swan song, Digital never recovered...
/daytripper
Bob Kaplow
> "Darren Peacock" <daz...@hotmail.com> wrote in message
> news:FsEq6.10155$0N3....@news-server.bigpond.net.au...
>> Cant remember the detail but back in 1988-89 there was a project in
> Digital
>> to do just that ..
>> My memory is fuzzy but it was called Emerald or Gem ..
>> It was descibed as VMS on Intel.
>>
>> But the initial downsizing then , the project was halted. Funny to see the
>> following 24 months some of the members end up in a small word processing
>> company called Microsoft.
> And that's all I got from Charlie. Heck, he started mumbling some stuff
> about OZIX, MERLIN, QUARTZ, CHEYENNE, and other cryptic codewords from days
> gone by. Wish he'd be a tad more forward-looking and spill the beans about
> MARVEL... and the system a generation beyond MARVEL.
There's always the CHARON VAX emulator for the PC. You can get a free
hobyist version off their web site.
Terry C. Shannon
"daytripper" <day_t...@REMOVEyahoo.com> wrote in message
news:5sematg4mmbhgtjar...@4ax.com...
C'est true about Crystal and its untimely fate. As for Emerald, the codename
is consistent with other DEC Jewels of the era. Charlie claims to have seen
an Emerald manual, but perhaps was mentally impaired or bedazzled by the
comely companion in whose residence he saw the manual.
>
> These were "VMS-on-RISC" programs, nothing to do with IA32.
>
> These were killed - along with the Argonaut program out of Littleton in
favor
> of the $3B disaster known as Aquarius (aka VAX9000), out of Marlboro.
Truly a sad story. I am told that the Aquarians convinced KO and the BoD
that they could sell something like 3K Aquariae per year; a Great
Expectation indeed when IBM was moving less than 2K mainframes per annum.
Bill Demmer (proponent of PRISM and other midrange SMP approaches) gave up
in disgust and let the Aquarians have their way. A total of 455 VAX 9000s
were sold, and the product was rendered obsolete in less than 12 months time
by Calypso, so Demmer got the last laugh.
As you note below, the stockholders and midrange systems engineers were not
laughing.
HappyCanuck
happened to it tho..
--
Bill
Toronto, Ontario Canada
"Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
news:iecpi5...@eisner.encompasserve.org...
Terry C. Shannon
"HappyCanuck" <happy...@altavista.net> wrote in message
news:FbOq6.69886$UZ4.17...@news4.rdc1.on.home.com...
> I can still remember a product, not from DEC, in the Rainbow era called PC
VMS. No idea what
> happened to it tho..
>
Ah yes... PC VMS, from Wendin Software. The firm is, I believe, taking the
dirt nap along with numerous other one-hit wonders of the 1980s. PC VMS
attempted to provide a VMS environment on a DOS PC, but neither Charlie nor
I ever fooled with the thing.
Rob Young
>
>
> about OZIX, MERLIN, QUARTZ, CHEYENNE, and other cryptic codewords from days
> gone by. Wish he'd be a tad more forward-looking and spill the beans about
> MARVEL... and the system a generation beyond MARVEL.
>
>
Which is to say....
We will see on-chip memory controllers on Alpha before we will
see them on Intel boxes. Compaq high-end AlphaServers will look
vastly different in 2-3 years than currently shipping high-end
AlphaServers. Compaq has won 7 out of 8 of the last go round
of Supercomputer bids, some of which are future roll-outs. Which
leads one to conclude Compaq AlphaServer division has a grand
story to tell. Too bad we can't hear it yet.
Rob
Terry C. Shannon
"Rob Young" <you...@encompasserve.org> wrote in message
news:EwxoEg...@eisner.encompasserve.org...
> Which is to say....
>
> We will see on-chip memory controllers on Alpha before we will
> see them on Intel boxes. Compaq high-end AlphaServers will look
> vastly different in 2-3 years than currently shipping high-end
> AlphaServers. Compaq has won 7 out of 8 of the last go round
> of Supercomputer bids, some of which are future roll-outs. Which
> leads one to conclude Compaq AlphaServer division has a grand
> story to tell. Too bad we can't hear it yet.
It will be a MARVELous tale indeed.
To be fair to Compaq, the firm *does* put out press releases on its
accomplishments. Subscribing to the HPTC newsletter from the Q is a good way
to keep up to date. (An even Better Answer is to subscribe to Shannon Knows
Compaq.)
What Compaq does *not* have is a customer evangelist who is empowered to go
out and proactively engage the press to ensure that coverage is granted
where coverage is due.
doh,
charlie matco
Robert Deininger
Shannon" <terrys...@mediaone.net> wrote:
> What Compaq does *not* have is a customer evangelist who is empowered to go
> out and proactively engage the press to ensure that coverage is granted
> where coverage is due.
Even Apple figured that one out eventually. They've done pretty well
since they stopped being suicidal. You hear that, Q?
--
Robert Deininger
rdein...@mindspring.com
Bradford J. Hamilton
>>
>> "Rob Young" <you...@encompasserve.org> wrote in message
>> news:EwxoEg...@eisner.encompasserve.org...
>>
>>> Which is to say....
>>
>>> We will see on-chip memory controllers on Alpha before we will
>>> see them on Intel boxes. Compaq high-end AlphaServers will look
>>> vastly different in 2-3 years than currently shipping high-end
>>> AlphaServers. Compaq has won 7 out of 8 of the last go round
>>> of Supercomputer bids, some of which are future roll-outs. Which
>>> leads one to conclude Compaq AlphaServer division has a grand
>>> story to tell. Too bad we can't hear it yet.
Why not?
>>
>> It will be a MARVELous tale indeed.
>>
>> To be fair to Compaq, the firm *does* put out press releases on its
>> accomplishments. Subscribing to the HPTC newsletter from the Q is a good way
>> to keep up to date. (An even Better Answer is to subscribe to Shannon Knows
>> Compaq.)
>>
>> What Compaq does *not* have is a customer evangelist who is empowered to go
>> out and proactively engage the press to ensure that coverage is granted
>> where coverage is due.
What would that entail?
>>
>> doh,
>>
>> charlie matco
>>
>>
Terry C. Shannon
"Bradford J. Hamilton" <hami...@encompasserve.org> wrote in message
news:c6lGfl...@eisner.encompasserve.org...
What would proactive engagement entail? Well, you could start with
establishing relationships with Key Influencers in the media, calling them
up when you've got news or developments with a potential news hook, etc.
Adopting a Clinton/Carville "War Room" strategy for addressing competitive
developments and competitive FUD might help, too. All too often Compaq
conducts its activities in a reactive, after-the-fact, fashion. If at all.
David J. Dachtera
I've asked RM to hire me into exactly that position on more than one
occasion. The results, so far, have been less than positive.
--
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.
Terry C. Shannon
"David J. Dachtera" <djesys...@earthlink.net> wrote in message
news:3AAC1518...@earthlink.net...
> "Terry C. Shannon" wrote:
> >
> > "Rob Young" <you...@encompasserve.org> wrote in message
> > news:EwxoEg...@eisner.encompasserve.org...
> >
> > > Which is to say....
> > >
> > > We will see on-chip memory controllers on Alpha before we will
> > > see them on Intel boxes. Compaq high-end AlphaServers will look
> > > vastly different in 2-3 years than currently shipping high-end
> > > AlphaServers. Compaq has won 7 out of 8 of the last go round
> > > of Supercomputer bids, some of which are future roll-outs. Which
> > > leads one to conclude Compaq AlphaServer division has a grand
> > > story to tell. Too bad we can't hear it yet.
> >
> > It will be a MARVELous tale indeed.
> >
> > To be fair to Compaq, the firm *does* put out press releases on its
> > accomplishments. Subscribing to the HPTC newsletter from the Q is a good
way
> > to keep up to date. (An even Better Answer is to subscribe to Shannon
Knows
> > Compaq.)
> >
> > What Compaq does *not* have is a customer evangelist who is empowered to
go
> > out and proactively engage the press to ensure that coverage is granted
> > where coverage is due.
>
> I've asked RM to hire me into exactly that position on more than one
> occasion. The results, so far, have been less than positive.
>
The OpenVMS Group already has a Marketing Director.
David J. Dachtera
Seems to me those would be two separate positions. Not sure who would
have authority over what, however.
Paul Repacholi
> > I've asked RM to hire me into exactly that position on more than
> > one occasion. The results, so far, have been less than positive.
Hey Dave, amybe they didn't want one with an axe an a weakess to be
suddenly overcome by attacks of candor ;)
> The OpenVMS Group already has a Marketing Director.
We know. That most of the problem.
--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
Raw, Cooked or Well-done, it's all half baked.
daytripper
> Which is to say....
>
> We will see on-chip memory controllers on Alpha before we will
> see them on Intel boxes.
*10* Rambus channels per processor, in fact...
> Compaq high-end AlphaServers will look
> vastly different in 2-3 years than currently shipping high-end
> AlphaServers. Compaq has won 7 out of 8 of the last go round
> of Supercomputer bids, some of which are future roll-outs. Which
> leads one to conclude Compaq AlphaServer division has a grand
> story to tell. Too bad we can't hear it yet.
>
> Rob
Well....Remember they pre-sold a 256-processor box to the Feds - got mucho
funding bucks for the deal - but by the time they actually built something
they could ship, the architecture could only get to 32 processors.
"Ooops! Sorry 'bout that - can you use a cluster instead?"
The last three years have been pretty weird for Marlboro. The first 18 months
they were battered by The Big Lie out of Houston - that "the Digital merger"
was dragging down the corporate bottom line. Every time the rolled-up numbers
lurched south, corporate hacks told every media outlet that their core
business (selling shitty peecees) was just fine - all of the money problems
were due to the pit in Marlboro...
Well, you can work an engineer to death - as long as he thinks he's respected
for the effort. The folks at MRO clearly weren't. And the result - man, the
attrition was awesome, as a *huge* percentage of the talent discovered The
Outside World - some for the first time in 20-something years - and the notion
of stock that actually appreciated ;-)
Meanwhile, the Oberfuerher is taken out, Mike C steps in, and as his first
significant action, unravels the books in public. Whoa Nellie - what have we
here? "Houston, we have a BIG problem!" The core peecee business was tanking,
big time, and no bottom in sight...
Meanwhile the midrange Alphabits and Tandems were actually turning serious
profits (into the 9 digit range) and in fact were holding Q's head above the
sea of "industry standard" red ink.
Gotta love the irony ;-)
As for Marlboro: well, some things never change:
- the volume server segment that *carried* Alpha for the last two years is
kaput - ALL of the engineers bailed - so the dual and quad bidness will be
based on reselling API hardware (assuming *that* effort ever bears fruit -
which is questionable - lots o' luck on chipset development).
- Wildfire (aka Aquarius-II with $2000 blowers) *finally* starts shipping -
three years late and still on a wing and a prayer. Will it ever recoup the
development costs - especially with the world-wide economy taking an extended
breather?
- And now Marvel - yet another Platform For All Segments (No! No! We've *seen*
that movie!) - is stumbling along in the pipeline. Another technological tour
de force - in an era of cheap hardware. The honchos will tell you they'll be
able to sell these from Slates to Crates. Good luck making any money at either
end...
/daytripper (If you listen closely, you can hear Don Meredith's voice...)
Josef Kolbitsch
> Even Apple figured that one out eventually. They've done pretty well
> since they stopped being suicidal. You hear that, Q?
well, i'd say, that steven jobs always wsa _the_man_ on the stage.
when he left the company, there was noone to really promote the
products.
unfortunately jobs didn't succeed with his nextstep, but the
presentations were just as those at apple now ... :)
--
---------------------------------------------------------
josef kolbitsch
> sk...@sbox.tu-graz.ac.at
> http://www.sbox.tu-graz.ac.at/home/s/skol
---------------------------------------------------------
andrew harrison
Why wait.
The Ultra III has an onchip memory controller now.
Incedentally I do admire your logic Rob. 7 out of
8 wins for supercomputer business = a great future
for Alphaservers sounds great until you remember
that sucess in the HPC space does not guarantee
sucess elsewhere.
If it did then Crays and SGI MIPS based
machines would be widely used for as commercial
servers, which they arn't.
Its also good to see that you havn't stopped
selling futures. I particularly liked you
advising the guy with performance issues on
WildFire to wait for Marvel.
Regards
Andrew Harrison
Enterprise IT Architect
Rob Young
> On 11 Mar 2001 16:40:36 -0500, you...@encompasserve.org (Rob Young) wrote:
>> Which is to say....
>>
>> We will see on-chip memory controllers on Alpha before we will
>> see them on Intel boxes.
>
> *10* Rambus channels per processor, in fact...
>
>> Compaq high-end AlphaServers will look
>> vastly different in 2-3 years than currently shipping high-end
>> AlphaServers. Compaq has won 7 out of 8 of the last go round
>> of Supercomputer bids, some of which are future roll-outs. Which
>> leads one to conclude Compaq AlphaServer division has a grand
>> story to tell. Too bad we can't hear it yet.
>>
>
> As for Marlboro: well, some things never change:
>
> - the volume server segment that *carried* Alpha for the last two years is
> kaput - ALL of the engineers bailed - so the dual and quad bidness will be
> based on reselling API hardware (assuming *that* effort ever bears fruit -
> which is questionable - lots o' luck on chipset development).
>
I disagree. The EV7 will make for a very nice quad-based server
and so maybe "one size fits all" isn't too far off the mark.
> - Wildfire (aka Aquarius-II with $2000 blowers) *finally* starts shipping -
> three years late and still on a wing and a prayer. Will it ever recoup the
> development costs - especially with the world-wide economy taking an extended
> breather?
>
> - And now Marvel - yet another Platform For All Segments (No! No! We've *seen*
> that movie!) - is stumbling along in the pipeline. Another technological tour
> de force - in an era of cheap hardware. The honchos will tell you they'll be
> able to sell these from Slates to Crates. Good luck making any money at either
> end...
>
Ah yes. Attack of the rack-em stack-em servers. 4 or 5 years ago
it was attack of the killer-PCs. And we have "dot NET" to see us
through with "distributed computing made right" by the Borg with
Infiniband giving us truly distributed data so compute cycles
are "out there" and storage is "out there".
There are several problems with all this of course. Last I checked,
the per port cost for switched storage will probably cost more than
the server itself, maybe Infiniband's per port cost is significantly
cheaper? I doubt that. Storage over-IP? Read Pfister's follow-up
to that in comp.arch:
Infiniband is for the Datacenter. Google's solution of course is
to have storage local (2 IDE drives internal to the thin-server) but
this is a special. So now build a much larger infastructure, etc.
etc.
Oh.. Marvel? As we know, the money in a lot of these deals is
made on the service end. I think Marvel will be the best high-end
box in 2-3 years and that has made Sun a ton of money as we can
see. We'll see how much of the industry starts loading up on
Marvel boxes. Tru64 + Oracle should jumpstart them pretty well,
it seems. Not only that, there are many industry segments that would
write large checks today if they can get a 10% decrease in runtimes
(or conversely 10% increase in real performance) so there will no
doubt be a need for high-end boxes in the age of systems like
Google. Google: Destroyer of the Environment. No doubt contributing
mightily to California's energy crisis (if you toss in Yahoo too).
Rob
Rob Young
> Rob Young wrote:
>>
>> In article <VhFq6.9855$5f.29...@typhoon.ne.mediaone.net>, "Terry C. Shannon" <terrys...@mediaone.net> writes:
>> >
>> >
>> > And that's all I got from Charlie. Heck, he started mumbling some stuff
>> > about OZIX, MERLIN, QUARTZ, CHEYENNE, and other cryptic codewords from days
>> > gone by. Wish he'd be a tad more forward-looking and spill the beans about
>> > MARVEL... and the system a generation beyond MARVEL.
>> >
>> >
>>
>> Which is to say....
>>
>> We will see on-chip memory controllers on Alpha before we will
>> see them on Intel boxes. Compaq high-end AlphaServers will look
>> vastly different in 2-3 years than currently shipping high-end
>> AlphaServers. Compaq has won 7 out of 8 of the last go round
>> of Supercomputer bids, some of which are future roll-outs. Which
>> leads one to conclude Compaq AlphaServer division has a grand
>> story to tell. Too bad we can't hear it yet.
>>
>
> Why wait.
>
> The Ultra III has an onchip memory controller now.
>
But they aren't shipping are they? Or.. more accurately, folks
have ordered SunBlades in September and are still waiting on
them. Is that the typical lead-time for a SunBlade 1000?
> Incedentally I do admire your logic Rob. 7 out of
> 8 wins for supercomputer business = a great future
> for Alphaservers sounds great until you remember
> that sucess in the HPC space does not guarantee
> sucess elsewhere.
>
Very true. But it does give sales and marketing folks something
very nice to talk about and if 2 or 3 of those sites allow themselves
to be "at the ready" to discuss why they went with Compaq, etc. it
does make for a good testimonial (better than an eBay testimonial
wouldn't you say? Caution: land mines abound. Have at it Andy!)
> If it did then Crays and SGI MIPS based
> machines would be widely used for as commercial
> servers, which they arn't.
Umm.. we know why that isn't the case. Care to share with the
audience or do you wish me to follow-up on this point? Why
raise this point?
> Its also good to see that you havn't stopped
> selling futures. I particularly liked you
> advising the guy with performance issues on
> WildFire to wait for Marvel.
Absolutely.
Rob
Jan Vorbrueggen
> > We will see on-chip memory controllers on Alpha before we will
> > see them on Intel boxes.
> *10* Rambus channels per processor, in fact...
The presentation at the last ISSC said 8 of 'em.
Jan
andrew harrison
>
> In article <3AACDAA5...@uk.sun.com>, andrew harrison <andrew...@uk.sun.com> writes:
> > Rob Young wrote:
> >>
> >> In article <VhFq6.9855$5f.29...@typhoon.ne.mediaone.net>, "Terry C. Shannon" <terrys...@mediaone.net> writes:
> >> >
> >> >
> >> > And that's all I got from Charlie. Heck, he started mumbling some stuff
> >> > about OZIX, MERLIN, QUARTZ, CHEYENNE, and other cryptic codewords from days
> >> > gone by. Wish he'd be a tad more forward-looking and spill the beans about
> >> > MARVEL... and the system a generation beyond MARVEL.
> >> >
> >> >
> >>
> >> Which is to say....
> >>
> >> We will see on-chip memory controllers on Alpha before we will
> >> see them on Intel boxes. Compaq high-end AlphaServers will look
> >> vastly different in 2-3 years than currently shipping high-end
> >> AlphaServers. Compaq has won 7 out of 8 of the last go round
> >> of Supercomputer bids, some of which are future roll-outs. Which
> >> leads one to conclude Compaq AlphaServer division has a grand
> >> story to tell. Too bad we can't hear it yet.
> >>
> >
> > Why wait.
> >
> > The Ultra III has an onchip memory controller now.
> >
>
> But they aren't shipping are they? Or.. more accurately, folks
> have ordered SunBlades in September and are still waiting on
> them. Is that the typical lead-time for a SunBlade 1000?
>
Yes they are, I saw 3 SunBlade 1000's recently in
one of the offices of the customer I advise. Of
course they could have been cardboard mockups but
they wern't.
I don't know what the lead time is perhaps you
should place an order and find out :):)
> > Incedentally I do admire your logic Rob. 7 out of
> > 8 wins for supercomputer business = a great future
> > for Alphaservers sounds great until you remember
> > that sucess in the HPC space does not guarantee
> > sucess elsewhere.
> >
>
> Very true. But it does give sales and marketing folks something
> very nice to talk about and if 2 or 3 of those sites allow themselves
> to be "at the ready" to discuss why they went with Compaq, etc. it
> does make for a good testimonial (better than an eBay testimonial
> wouldn't you say? Caution: land mines abound. Have at it Andy!)
>
> > If it did then Crays and SGI MIPS based
> > machines would be widely used for as commercial
> > servers, which they arn't.
>
> Umm.. we know why that isn't the case. Care to share with the
> audience or do you wish me to follow-up on this point? Why
> raise this point?
>
Why arn't SGI MIPS machines used much in the commercial market ??
Could it be poor market acceptance, poor marketing, lack of
available commercial software, lack of a salesforce who can
open commercial customers doors.
Anything ring a bell with you and do any of the reasons why
SGI's arn't sucessfull as commercial servers have anything
to do with technology ????
> > Its also good to see that you havn't stopped
> > selling futures. I particularly liked you
> > advising the guy with performance issues on
> > WildFire to wait for Marvel.
>
> Absolutely.
>
I am waiting with some anticipation for the moment
when you actually recommend an existing Compaq product
to solve someones problem. What follows Marvel ???
Terry C. Shannon
"andrew harrison" <andrew...@uk.sun.com> wrote in message
news:3AACFBC3...@uk.sun.com...
> Rob Young wrote:
<snip>
> >
>
> Why arn't SGI MIPS machines used much in the commercial market ??
>
> Could it be poor market acceptance, poor marketing, lack of
> available commercial software, lack of a salesforce who can
More likely the fact that Sun bought the family jewels from SGI's Cray brain
trust and very capably exploited same in the form of the Starfire. Which, if
memory serves me correctly, has enjoyed some of the highest shipment rates
in the enterprise server space.
>
> I am waiting with some anticipation for the moment
> when you actually recommend an existing Compaq product
> to solve someones problem. What follows Marvel ???
As your competitive intel folks no doubt are aware, the Marvel followon will
be similar to the Serengeti followon. Mass quantities of CPUs on a big-arse
switched fabric. CPQ and SUNW have very similar Server Utility strategies.
Of course, neither Serengeti nor Marvel are soup yet (if all goes well for
Sun, Serengeti should precede Marvel by as much as a year).
Rob Young
>> >
>> > The Ultra III has an onchip memory controller now.
>> >
>>
>> But they aren't shipping are they? Or.. more accurately, folks
>> have ordered SunBlades in September and are still waiting on
>> them. Is that the typical lead-time for a SunBlade 1000?
>>
>
> Yes they are, I saw 3 SunBlade 1000's recently in
> one of the offices of the customer I advise. Of
> course they could have been cardboard mockups but
> they wern't.
>
> I don't know what the lead time is perhaps you
> should place an order and find out :):)
>
Seems to be at least 6 months:
http://www.aceshardware.com/board/general/read.php?message_id=30008802
"Hello all
I ordered two Sunblade 1000 in September. I Have still not received them.
Suns estimate is in late Mars. Since the machines I ordered cost around
11000$ each, Sun should have good margins. "
There is another one out there that says the same. Ordered
in September and still waiting. Those SunBlades you "saw", were
they ordered in August?
This is a big matrix. But one intersection whereby I recommend
Compaq product is if they are having difficulty with acceptance
testing of Sun boxes. From what I understand, many folks have
creative workarounds for Sun crashing problems (eBay for example)
but for others (APAC for example) workarounds don't cut it. So
in summary, Compaq is a far superior solution if the problem you
are trying to solve is a server uptime problem. (One example
among many but a very good one).
Rob
Paul Repacholi
Interesting Andrew. Perhaps you could explain the Cray/SGI design and
how it is going in the comercial arena. If it stays up long enough for
you to post your reply that is.
Bob Kaplow
> machines would be widely used for as commercial
> servers, which they arn't.
Didn't Sun buy Cray, or at least the rights to the UE10K cache error box. It
*IS* a Cray design. Which is probably why it's such a bad commercial
server...
Terry C. Shannon
"Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
>
It
> *IS* a Cray design. Which is probably why it's such a bad commercial
> server...
Indeed Sun did purchase the Cray technology that begat the UE10K from SGI
(remember them?). I don't think think the rights to, or the roots of,
Cachegate hail from Mendota, though.
Bob Kaplow
The lack of parity in Cray designs goes back at least to his days at CDC in
the early 60s. [The first machine I used in college was a CDC 6000, one of
his cool early designs. RISC, functional units, pipelining, etc were all
present in this design, WAY ahead of its time.] Any error detection or
correction slowed down the machine, and his goal was always "full speed
ahead". Cost was never a factor, and reliability rarely was.
Paul DeMone
andrew harrison wrote:
>
Hey look who crawled out of the woodwork. Hanging out at
c.s.i looking for hints of when merced s(l/h)ips?
> Why wait.
>
> The Ultra III has an onchip memory controller now.
And the performance shows. Sure couldn't get those
SPECmarks with a system bus. ;^)
Hmmm, connecting CPU pins to SDRAM? I guess Sun has
a lot more cheap package pins than good ideas. Perhaps
SDRAM was state of the art when the 3 was architected.
>
> Incedentally I do admire your logic Rob. 7 out of
> 8 wins for supercomputer business = a great future
> for Alphaservers sounds great until you remember
> that sucess in the HPC space does not guarantee
> sucess elsewhere.
I don't know. The Sun replacement business is certainly
a potential growth area. SPARC reminds me of the heyday
of the VAX. Big, noisy, lots of press, lots of revenue.
Slow CPUs held buoyant by software, seamless networking,
but mostly fading reputation. And ripe for the big fall.
BTW, has the US5 team come out of reset yet?
>
> If it did then Crays and SGI MIPS based
> machines would be widely used for as commercial
> servers, which they arn't.
>
> Its also good to see that you havn't stopped
> selling futures.
That stuff is due in the near future. Like most of Sun's
US3 line. Except the EV68/EV7 stuff won't be born obsolete.
--
Paul W. DeMone The 801 experiment SPARCed an ARMs race of EPIC
Kanata, Ontario proportions to put more PRECISION and POWER into
dem...@mosaid.com architectures with MIPSed results but ALPHA's well
pde...@igs.net that ends well.
David Mathog
>
>The OpenVMS Group already has a Marketing Director.
Good to hear that the invisible man is gainfully employed.
David Mathog
mat...@seqaxp.bio.caltech.edu
Manager, sequence analysis facility, biology division, Caltech
**************************************************************************
* RIP VMS *
**************************************************************************
Paul Repacholi
> From what I understand, many folks have
> creative workarounds for Sun crashing problems (eBay for
> example)...
Ah, So thats why Cray/SGI don't do wwell as commercial servers. Knew
there would be a reason. Guess all those Origins out there must be
working for charity. And you'll never find an Indigo in a graphics
of TV/Film studio...
David Mathog
Can somebody please verify the details related in this story? In
particular, the level of attrition at Marlboro and the hollowing out of the
small/midrange machine engineering pool. This makes future alpha systems
development appear to be even more of an iffy proposition than it already
was.
>
>Well....Remember they pre-sold a 256-processor box to the Feds - got mucho
>funding bucks for the deal - but by the time they actually built something
>they could ship, the architecture could only get to 32 processors.
>"Ooops! Sorry 'bout that - can you use a cluster instead?"
>
>The last three years have been pretty weird for Marlboro. The first 18 months
>they were battered by The Big Lie out of Houston - that "the Digital merger"
>was dragging down the corporate bottom line. Every time the rolled-up numbers
>lurched south, corporate hacks told every media outlet that their core
>business (selling shitty peecees) was just fine - all of the money problems
>were due to the pit in Marlboro...
>
>Well, you can work an engineer to death - as long as he thinks he's respected
>for the effort. The folks at MRO clearly weren't. And the result - man, the
>attrition was awesome, as a *huge* percentage of the talent discovered The
>Outside World - some for the first time in 20-something years - and the notion
>of stock that actually appreciated ;-)
It's just plain nuts to spend $9B on what is essentially a technology
company - and then drive all the technologists away.
>
>Meanwhile, the Oberfuerher is taken out, Mike C steps in, and as his first
>significant action, unravels the books in public. Whoa Nellie - what have we
>here? "Houston, we have a BIG problem!" The core peecee business was tanking,
>big time, and no bottom in sight...
>
>Meanwhile the midrange Alphabits and Tandems were actually turning serious
>profits (into the 9 digit range) and in fact were holding Q's head above the
>sea of "industry standard" red ink.
>
>Gotta love the irony ;-)
>
>As for Marlboro: well, some things never change:
>
>- the volume server segment that *carried* Alpha for the last two years is
>kaput - ALL of the engineers bailed - so the dual and quad bidness will be
>based on reselling API hardware (assuming *that* effort ever bears fruit -
>which is questionable - lots o' luck on chipset development).
This is the part that's really scary. The local experience with API
machinery has been pretty dismal. (These weren't my machines but I've been
observing their "progress" at close hand.) Quality has been bad (systems
that won't run at rated speed, systems that won't run with graphics cards
that they supposedly supported) and the price was only ever marginally
lower than that of real Compaq kit. And the API site doesn't indicate that
they can run VMS or Tru64, although I'm pretty sure that the UP2000 can,
but not sure at what level of support that is. (Ie, some of the NT 164
systems will run VMS too, but that doesn't mean that anybody supports
that.) Nor has API ever built anything with >2 processors, and it's a big
step in complexity up from 2 CPUs.
>
>- Wildfire (aka Aquarius-II with $2000 blowers) *finally* starts shipping -
>three years late and still on a wing and a prayer. Will it ever recoup the
>development costs - especially with the world-wide economy taking an extended
>breather?
>
>- And now Marvel - yet another Platform For All Segments (No! No! We've *seen*
>that movie!) - is stumbling along in the pipeline. Another technological tour
>de force - in an era of cheap hardware. The honchos will tell you they'll be
>able to sell these from Slates to Crates. Good luck making any money at either
>end...
>
The outcome for both of the above products will depend more on marketing
than on technology. Which, unfortunately, is not a good thing, considering
who it is that needs to do the marketing.
Regards,
Sharkonwheels
"Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
> In article <VhFq6.9855$5f.29...@typhoon.ne.mediaone.net>, "Terry C.
Shannon" <terrys...@mediaone.net> writes:
> > "Darren Peacock" <daz...@hotmail.com> wrote in message
> > news:FsEq6.10155$0N3....@news-server.bigpond.net.au...
>
> There's always the CHARON VAX emulator for the PC. You can get a free
> hobyist version off their web site.
Anything in this era is useless without network support. Even
my Amiga emulator supports TCP/IP handed down to Winsock.
Charon was one app that practically set a record for the speed in
which it hit my recycle bin.
Tony
tonymATcompusourceDOTnet
Sharkonwheels
"Terry C. Shannon" <terrys...@mediaone.net> wrote in message
news:c3Vq6.12962$5f.33...@typhoon.ne.mediaone.net...
>
> "David J. Dachtera" <djesys...@earthlink.net> wrote in message
> news:3AAC1518...@earthlink.net...
> > "Terry C. Shannon" wrote:
> > >
> > > "Rob Young" <you...@encompasserve.org> wrote in message
> > > news:EwxoEg...@eisner.encompasserve.org...
> > >
> > occasion. The results, so far, have been less than positive.
>
> The OpenVMS Group already has a Marketing Director.
>
He....I wondered if that "Marketing Director" has Commodore on
his resume.....'bout the same productivity.
Tony
tonymATcompusourceDOTnet
Bob Kaplow
> "Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
>> hobyist version off their web site.
>
> Anything in this era is useless without network support. Even
> my Amiga emulator supports TCP/IP handed down to Winsock.
> Charon was one app that practically set a record for the speed in
> which it hit my recycle bin.
From what I heard last week, the hobby kit is network-less, but the full
commercial version does have ethernet support. IIRC you can run DECnet out
the same port as your PC uses for TCPIP, but if you wnat to run TCPIP you
need to use a seperate card.
daytripper
>
> Seems to be at least 6 months:
>
>http://www.aceshardware.com/board/general/read.php?message_id=30008802
>
>"Hello all
>
> I ordered two Sunblade 1000 in September. I Have still not received them.
daytripper
wrote:
>In article <c3Vq6.12962$5f.33...@typhoon.ne.mediaone.net>, "Terry C. Shannon" <terrys...@mediaone.net> writes:
>>
>>The OpenVMS Group already has a Marketing Director.
>
>Good to hear that the invisible man is gainfully employed.
Post Of The Week candidate.
Bill Todd
...
> There are several problems with all this of course. Last I checked,
> the per port cost for switched storage will probably cost more than
> the server itself, maybe Infiniband's per port cost is significantly
> cheaper? I doubt that.
You might want to rethink that doubt. Intel seems hell-bent on making
Infiniband a volume technology (perhaps not desktop-PC-volume, but
commodity-server-volume), and supposedly plans to make it a motherboard
component to help get costs in line.
Storage over-IP? Read Pfister's follow-up
> to that in comp.arch:
>
>
http://groups.google.com/groups?q=pfister+infiniband+group:comp.arch&hl=en&l
r=&safe=off&rnum=1&seld=924890734&ic=1
The observation that IP protocols tend to stress the host processor more
than IB will, while correct, decreases in importance at least on the storage
end, where the host processor really doesn't have all that much else to do.
And while Greg wasn't optimistic about NIC-level hardware taking up this
load, there are already 100 Mb and Gigabit Ethernet cards available that do
so (and pricing for the 100 Mbit cards was IIRC very reasonable).
>
> Infiniband is for the Datacenter.
Not in Intel's plans: they want it on every local server.
Google's solution of course is
> to have storage local (2 IDE drives internal to the thin-server) but
> this is a special.
Their application may be a special, but there are a lot of other
applications that it could be usefully applied to. As well as a lot that it
couldn't, of course. And doubling their drive density would be easy (e.g.,
2U server boxes with 9 hot-swap front-loading drives are on the market),
which also helps alleviate your power concerns (the more drives in the box,
the less the CPU contributes to the overall load - and in any event a lot of
uses don't require that fast a CPU in the box, which drops its power
requirement a lot).
- bill
Hunter Goatley
Kaplow) wrote:
>In article <1Gfr6.2393$3e.30...@news3.mco>, "Sharkonwheels" <to...@compusourceDOT.net> writes:
>> "Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
>>> There's always the CHARON VAX emulator for the PC. You can get a free
>>> hobyist version off their web site.
>>
>> Anything in this era is useless without network support. Even
>> my Amiga emulator supports TCP/IP handed down to Winsock.
>> Charon was one app that practically set a record for the speed in
>> which it hit my recycle bin.
>
>From what I heard last week, the hobby kit is network-less, but the full
>commercial version does have ethernet support.
That is correct.
Personally, I think it's great that they've released a hobbyist version.
While it would be nicer if it included network support, it's still a
great alternative to only having Windows on a laptop. It just means
you have to be more creative about getting data in and out. And since
the hobbyist version can read VMS CDs in your PC's CD-ROM or DVD-ROM
drive, that makes it even easier to get data in (and least initially).
>IIRC you can run DECnet out
>the same port as your PC uses for TCPIP, but if you wnat to run TCPIP you
>need to use a seperate card.
Or disable TCP/IP as an active protocol for Windows on your single card.
You can even cluster the commercial version with your other VMS systems,
if you have any.
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goath...@goatley.com http://www.goatley.com/hunter/
Rob Young
Great follow-up. I appreciate this contrarian view as we see
Pfister and know him as a well-respected author and engineer
in his own right (I suppose a PhD from MIT should also merit
a dollop of respect too). Thanks for the heads-up on Infiniband
cost or future costs, makes sense. Food for thought all round!
Rob
Paul Repacholi
See the early Alpha papers. *Correction* costs time, *detection* need
not. The rule of the CDC 6000s and Cybers with ECS was to never put
any data into it that you'd want to read back! I don't think the 6000
had any pipelined units, about the only 'modern' idea it didn't have!
Bob Kaplow
>> The lack of parity in Cray designs goes back at least to his days at
>> CDC in the early 60s. [The first machine I used in college was a CDC
>> 6000, one of his cool early designs. RISC, functional units,
>> pipelining, etc were all present in this design, WAY ahead of its
>> time.] Any error detection or correction slowed down the machine,
>> and his goal was always "full speed ahead". Cost was never a factor,
>> and reliability rarely was.
>
> See the early Alpha papers. *Correction* costs time, *detection* need
> not. The rule of the CDC 6000s and Cybers with ECS was to never put
> any data into it that you'd want to read back! I don't think the 6000
> had any pipelined units, about the only 'modern' idea it didn't have!
Hmmm, you may be right about that. I thought the Multiply / Divide units
might have been pipelined. One other modern feature it did have that I
forgot to list was out of order execution. And a RISC style FORTRAN compiler
that ripped apart your code and rebuilt it for optimal execution among the
functional units.
Bob Kaplow
>>IIRC you can run DECnet out
>>the same port as your PC uses for TCPIP, but if you wnat to run TCPIP you
>>need to use a seperate card.
>
> Or disable TCP/IP as an active protocol for Windows on your single card.
>
> You can even cluster the commercial version with your other VMS systems,
> if you have any.
There's a version of CHARON-VAX in the works for Alpha VMS. Run Galaxy on
uour VMS box; in one partition run CHARON, and have that run VAX/VMS;
cluster the VAX emulator instance with the rest of the Galaxy!
Word is that VMS engineering will SUPPORT the CHARON-VAX as a valid
"hardware" environment for VMS. Just the thing for those apps still stuck in
VAXland in the 21st century.
NewsReader
max. Surely this is a little limiting?
"Bob Kaplow" <kapl...@eisner.encompasserve.org.mars> wrote in message
news:eMHfuP...@eisner.encompasserve.org...
Bob Kaplow
> I read the tech .pdfs for Charon & it says for PC memory emulation is 16Mb
> max. Surely this is a little limiting?
The current version emulates a MicroVAX-II. The MV2 was limited to 16 MB. So
it's as limiting as the hardware it's emulating.
An MV3000 is supposed to be in the works. And some folks probably need a
6000 or 7000 class emulator. I wonder if they can do SMP emulation :-)
daytripper
>kapl...@eisner.encompasserve.org.mars (Bob Kaplow) writes:
>> The lack of parity in Cray designs goes back at least to his days at
>> CDC in the early 60s. [The first machine I used in college was a CDC
>> 6000, one of his cool early designs. RISC, functional units,
>> pipelining, etc were all present in this design, WAY ahead of its
>> time.] Any error detection or correction slowed down the machine,
>> and his goal was always "full speed ahead". Cost was never a factor,
>> and reliability rarely was.
>
>See the early Alpha papers. *Correction* costs time, *detection* need
>not.
Don't know what the papers said, but it is true that one can optimize a
Hamming code read path such that data correctors propagate one gate delay
faster than the error syndromes.
/daytripper
andrew harrison
>
> andrew harrison wrote:
> >
>
> Hey look who crawled out of the woodwork. Hanging out at
> c.s.i looking for hints of when merced s(l/h)ips?
Since I have never subscribed to comp.sys.intel, I can only
assume that you are refering to someone else.
> > Why wait.
> >
> > The Ultra III has an onchip memory controller now.
>
> And the performance shows. Sure couldn't get those
> SPECmarks with a system bus. ;^)
>
> Hmmm, connecting CPU pins to SDRAM? I guess Sun has
> a lot more cheap package pins than good ideas. Perhaps
> SDRAM was state of the art when the 3 was architected.
>
Humm so what would you recommend ?? RAMBUS, a fairly
safe recommendation on comp.os.vms/comp.sys.intel but
one that would garner much less support on comp.arch
as you know.
> >
> > Incedentally I do admire your logic Rob. 7 out of
> > 8 wins for supercomputer business = a great future
> > for Alphaservers sounds great until you remember
> > that sucess in the HPC space does not guarantee
> > sucess elsewhere.
>
> I don't know. The Sun replacement business is certainly
> a potential growth area. SPARC reminds me of the heyday
> of the VAX. Big, noisy, lots of press, lots of revenue.
> Slow CPUs held buoyant by software, seamless networking,
> but mostly fading reputation. And ripe for the big fall.
> BTW, has the US5 team come out of reset yet?
>
So who would you choose instead.
HP, nah no one could be that dumb. Big first level
cache, great SPECint/TPC-C performance and nothing
much else. Crippled by relatively small cache sizes
in a CPU that was designed to be "low cost" because
we should all be running IA-64 by now. Waiting
to put their customers through the HP-PA to IA-64
transition while frantically re-spinning the 8XXXX.
I have just had the great joy of having to organise
5 different apps benchmarks for 2 customers because
HP tried to convince them that HP-PA was twice as
fast as anything else on the market, so the customers
would only need 50% of the CPU's. The apps ranged
from HPC/FP to OLTP and Batch. 3 inhouse developed
on HP, 2 off the shelf. In none of them and in
none of the sub-runs did Sun require more CPU's
than HP to be as fast of faster than HP and in
a couple of cases the HP's were embarassingly
slow. In all cases HP benchmarked the best they
had.
Ever wondered why HP never published SPECrate_fp
numbers for the V-class, if you did look at the
numbers for the N-class it will give you a hint.
Compaq, nah no one could be that dumb. Crippled
by WildFire beginning to trail Marvel as the
solution to all the WildFire problems.
So what would you recommend ???????
> >
> > If it did then Crays and SGI MIPS based
> > machines would be widely used for as commercial
> > servers, which they arn't.
> >
> > Its also good to see that you havn't stopped
> > selling futures.
>
> That stuff is due in the near future. Like most of Sun's
> US3 line. Except the EV68/EV7 stuff won't be born obsolete.
>
Really so how come WildFire was ?????????
Its not much point having a CPU which isn't obselete
if the only system you can run it in is.
Jon Morgan
<terrys...@mediaone.net> wrote:
>The OpenVMS Group already has a Marketing Director.
Does it?
-jon.
Dave Weatherall
Could one run ELN on it I wonder.
--
Cheers - Dave.
Paul Repacholi
> Great infiniband looks like some old concepts ;) Looks like and
> update of an RH20 channel for the new century ;)
True. Read Tony in RH20 land? :)
Mark Garrett
you...@encompasserve.org wrote on 14/03/2001 00:42:
>>
>>>
>> http://groups.google.com/groups?q=pfister+infiniband+group:comp.arch&hl=en&l
>> r=&safe=off&rnum=1&seld=924890734&ic=1
>>
Great infiniband looks like some old concepts ;) Looks like and update of
an RH20 channel for the new century ;)
Cheers
mark ;)
Mark Garrett
pr...@prep.synonet.com wrote on 16/03/2001 22:05:
> Mark Garrett <Mark.G...@wedontwantyourspam.com.au> writes:
>
>> Great infiniband looks like some old concepts ;) Looks like and
>> update of an RH20 channel for the new century ;)
>
> True. Read Tony in RH20 land? :)
Read it a couple of times, I suspect Tony was just pissed off that the lack
of info and the design being anything but for TOPS-10, when he had a
perfectly good RH10 driver and they did this RH20 to him :). I think the
RH20 was a good enough channel controller though, just not what he expected
and certainly a few quirks (aka bugs) in the hardware to work around.
Cheers
Mark ;)
Roland Roberts
Terry> Separately, there was a midnight project to port VMS to the
Terry> Mach kernel. The project was done (half-baked, actually) at
Terry> Carnegie Mellon University IIRC. Some of the incomplete
Terry> code--which may well have a few facets of Emerald embedded
Terry> in it--is said to be floating around somewhere.
Once upon a time, there were some papers on the web on this project.
Richard Levitte used to have copies or links on his web site
(http://www.lp.se/), but I can't find them there anymore. They were
discussed as part of a "FreeVMS" project that died out when DEC and
then Compaq decide to open up the hobbyist program.
I think I have hardcopies at home; if anyone is interested, I can try
to post a reference. The project ran VMS services on a Mach
microkernel which was running BSD as the primary OS. The hardware was
still a VAX as the authors wanted to be able to avoid having to write
everything from scratch: using VAX hardware meant VMS binaries could
be run....
roland
--
PGP Key ID: 66 BC 3B CD
Roland B. Roberts, PhD RL Enterprises
rol...@rlenter.com 76-15 113th Street, Apt 3B
rbro...@acm.org Forest Hills, NY 11375
Bob Koehler
> I think I have hardcopies at home; if anyone is interested, I can try
> to post a reference. The project ran VMS services on a Mach
> microkernel which was running BSD as the primary OS. The hardware was
> still a VAX as the authors wanted to be able to avoid having to write
> everything from scratch: using VAX hardware meant VMS binaries could
> be run....
I read those papers. The effort was limitted to getting a VAX/VMS copy
of COPY.EXE to run on Mach, and only for selected file organizations.
Most of the VMS kernel, much of RMS, and much of the CLI was not
attempted. Other parts of VMS were not needed for the demonstration.
Wanting to make sure COPY.EXE ran unmodified in the demonstration was
part of the reason for doing the demonstration on a VAX.
----------------------------------------------------------------------
Bob Koehler | Computer Sciences Corporation
NASA GSFC Flight Software | Federal Sector, Civil Group
| please remove ".aspm" when replying
Ove Axelsson
And read:
CHARON-VAX Hobbyist release notes
TERMS OF USE
The copyright of this hobbyist edition of the CHARON-VAX emulator is
the
property of Software Resources International. You are authorized to
download,
use and freely copy this software without charge for PRIVATE,
NON-COMMERCIAL
USE provided that the trademark acknowledgement is affixed to each
copy and
the content of the software is not altered.
Software Resources International neither makes representations that
the use
of CHARON-VAX will not infringe on existing or future patent rights
nor will
it provide guarantees and therefore disclaims liability concerning
performance
or reliability of the CHARON-VAX hobbyist software. You use this
software at
your own risk and/or responsibility.
Hardware and Software Requirements:
A personal computer running Windows 98, NT or 2000 with 160 MB of free
disk space.
CHARON-VAX is distributed as a compressed Zip file. Install CHARON-VAX
by
running the setup.exe included in this distribution.
CHARON-VAX can be run or un-installed from the entries that are added
to your
start menu.
Release 10 November 2000
----------------------------------------
The following additions and changes have been made:
Date limit change; this release is now operational until May 1, 2001.
This is a consolidated distribution replacing the separate W98 and
WNT/2000
releases. Physical media access (CD or Floppy disk) does not work on
W98 due
to design constraints. Media access is checked and a corresponding
error
message is generated when running under W98.
Some DVD drives will work as well, but only limited testing has taken
place.
Release 12 October 2000
----------------------------------------
Add halt_enable and power_up_mode switches to CPU device. Both
represent
corresponding switches of CPU I/O insert module of real MicroVAX.
Please check
user manual for detailed description of these two options.
To conform with future hobbyist versions on other platforms and the
CHARON-VAX
product strategy, the number of operational DHV11 lines in the
hobbyist versions
is limited to 2 (line 0 and 1). the other lines are not operational.
Release 4 October 2000
----------------------------------------
Minor changes to direct CD/Floppy support and to CPU module.
Change names of devices to make them consistent with other releases
which use
predefined hardware configuration and with names of controllers which
are
assigned to them by OpenVMS/VAX.
Release 18 August 2000
----------------------------------------
Windows NT 4.0 & Windows 2000 only:
Add ability to read CD ROM directly from CHARON-VAX. CHARON-VAX will
recognize
CD-ROM as DEC RRD45. Boot from CD is also supported. It means that
CHARON-VAX
can boot directly from OpenVMS/VAX Hobbyist CD and install OpenVMS/VAX
operating system onto empty disk image file which is supplied as
additional
component during installation.
You should use CD-ROM drive letter and edit CHARON.INI file to
configure
CHARON-VAX to read a CD. For example (provided that E: stands for CD
drive):
set DUA file[3]="\\.\E:"
Also supported direct access to floppy drives and floppy disks
formatted
under OpenVMS operating system, both VAX and Alpha. Floppy disks
formatted
under CHARON-VAX are also readable and writable under OpenVMS/VAX and
OpenVMS/Alpha. Various floppy drives are recognized by CHARON-VAX as
DEC RX23 or DEC RX26.
You should use floppy drive letter and edit CHARON.INI file to
configure
CHARON-VAX to read a floppy. For example (provided that A: stands for
a
floppy drive):
set DUA file[2]="\\.\A:"
Release 12 August 2000
----------------------------------------
Console port logic fixed. Using a telnet connection will no longer
interfere
with the console port diagnostics and occasionaly generate a
diagnostic
error.
Xon/Xoff synchronization now by default enabled on console port and
DHV11
lines, which can be disabled if needed in the CHARON.INI file
(e.g. set OPA ttsync=no hsync=no, or set TXA ttsync[1]=no
hsync[1]=no).
Cut and Paste into a telnet-connected terminal emulator window will
now work
properly provided that HOSTSYNC is enabled for VMS terminal device and
for
serial line in the CHARON.INI file (e.g. "set OPA hsync=yes" for the
CHARON.INI file and "set term/perm/hostsync/inq" for VMS terminal
device).
Release 24 July 2000
----------------------------------------
Floating point error fixed.
Beta 2 release 30 June 2000
----------------------------------------
This release of the CHARON-VAX hobby kit is distributed to a limited
audience
for preliminary testing. These tests should assess whether this kit is
now
ready for general release.
Fixes applied:
Errors related to modem status on lines assigned to DHV11.
We tested the Hyperterminal Private Edition from Hillgraeve as a
console
terminal using a local telnet connction on Windows 2000. This works
well and is
documented in the manual.
NOTE FOR SERIAL LINES CONFIGURATION
If you are assigning the specific COMx: port with the SET TXA
LINE[0]="COMx:"
instruction in the CHARON.INI file, the corresponding port must be
available in
your Windows configuration. Otherwise the system will crash.
Copyright (C) 2000 Software Resources International
# 30-17-002
On 19 Mar 2001 14:59:04 -0500, Roland Roberts <rol...@rlenter.com>
wrote:
Bill Pedersen
Am currently working with a couple customers with the emulator on OpenVMS
Alpha for support of OpenVMS VAX applications which they do not wish to
migrate (or can not) to Alpha.
Newest version (MV3500/3600) which supports 64MB is still in beta.
Looks JUST LIKE A VAX.
Issues are, of course, the extra network adapter and a console device. On
workstation console can be another x-window. On server have developed
technique of using a LAT device is applicaiton mode as the "console".
--
Bill Pedersen
CCSS Corporation
www.CCSScorp.com
831-336-2708
"Dave Weatherall" <djw...@attglobal.net> wrote in message
news:DTiotGxQ0bj6-pn2-dO13EFYdIgvz@localhost...
bill davidsen
Bill Pedersen <pede...@ccsscorp.com> wrote:
| ELN has been run on the Charon-VAX emulator.
|
| Am currently working with a couple customers with the emulator on OpenVMS
| Alpha for support of OpenVMS VAX applications which they do not wish to
| migrate (or can not) to Alpha.
|
| Newest version (MV3500/3600) which supports 64MB is still in beta.
|
| Looks JUST LIKE A VAX.
|
| Issues are, of course, the extra network adapter and a console device. On
| workstation console can be another x-window. On server have developed
| technique of using a LAT device is applicaiton mode as the "console".
I'd re-ask the previous question, can SMP be used to improve
performance? And given the relative price and performance of ia86 vs.
Alpha, isn't the market limited to people who need to run VAX software
(VMS? OpenUNIX?) but can't or won't use a real VAX? If the markey is
people who can't afford a VAX, I would guess the price of the software
is going to have to be pretty low or they can't afford that, either.
Those are questions, not comments or anything, I'm just curious about
the implementation and motivation.
--
bill davidsen <davi...@tmr.com> CTO, TMR Associates, Inc
"I am lost. I am out looking for myself. If I should come back before I
return, please ask me to wait." -seen in a doctor's office
Larry Kilgallen
> I'd re-ask the previous question, can SMP be used to improve
> performance? And given the relative price and performance of ia86 vs.
> Alpha, isn't the market limited to people who need to run VAX software
> (VMS? OpenUNIX?) but can't or won't use a real VAX? If the markey is
> people who can't afford a VAX, I would guess the price of the software
> is going to have to be pretty low or they can't afford that, either.
I believe the market is people who _cannot_ change, for instance
because they depend on software whose source is lost (such as being
from a vendor that no longer exists).
Bill Pedersen
1) the continued support of existing OpenVMS VAX based application
without port to OpenVMS Alpha - especially long term given there are no
longer any VAX systems being manufactured.
2) small, portable development environments.
These are the primary draw that I see for the product at present.
There does seem to be some interest in it, in talking with the folks at
Software Resources International they see between 100 and 200 downloads of
the hobbyiest version DAILY.
SMP is not the direction being pursued at the present time. It uses SMP
already in that each thread in the emulator (each device) is a thread.
--
Bill Pedersen
CCSS Corporation
www.CCSScorp.com
831-336-2708
"bill davidsen" <davi...@tmr.com> wrote in message
news:9aktmn$2nkq$1...@newssvr05-en0.news.prodigy.com...
bill davidsen
Bill Pedersen <pede...@ccsscorp.com> wrote:
| Motivation is:
|
| 1) the continued support of existing OpenVMS VAX based application
| without port to OpenVMS Alpha - especially long term given there are no
| longer any VAX systems being manufactured.
|
| 2) small, portable development environments.
|
| These are the primary draw that I see for the product at present.
|
| There does seem to be some interest in it, in talking with the folks at
| Software Resources International they see between 100 and 200 downloads of
| the hobbyiest version DAILY.
|
| SMP is not the direction being pursued at the present time. It uses SMP
| already in that each thread in the emulator (each device) is a thread.
That's what I was asking, if the emulator runs faster on N processors
for larger values of N. And it sounds like yes to me.
I do know some people who have written large systems for VMS, I'm not
sure if they have source, given that they lost the support documentation
I can't believe they do. They are probably a potential customer.
Lock Horsburgh
(to you) topic, but I never heard of this VMS on Intel before.
Does this really mean you can run VAX stuff on a PC?
Under windows/dos/linux, or instead of?
Dual boot possible?
Does this mean you can carry a VMS app around on a laptop, or is it not
quite that simple?
What does it cost?
You mentioned downloads and Software Resources International.
Is there a URL where I could find out more?
Lock Horsburgh
"Bill Pedersen" <pede...@ccsscorp.com> wrote in message
news:3acf59c3$0$830$8ee...@newsreader.tycho.net...
Paul Repacholi
> I'm new to this NG, so please forgive me if I am going back over a
> familiar (to you) topic, but I never heard of this VMS on Intel
> before.
> Does this really mean you can run VAX stuff on a PC?
No
> Under windows/dos/linux, or instead of?
No
> Dual boot possible?
No
> Does this mean you can carry a VMS app around on a laptop, or is it not
> quite that simple?
Yes, see below.
> What does it cost?
You need to find an Alphabook owner, club the 20 people lined up
behind him, and then beat him to death and steal the Alphabook. Beware
of c.o.vms readers behind you :)
On a slightly more serious note. There is a Vax emulation, Charon Vax,
for PC machines.
Hunter Goatley
<lo...@lhorsburgh.NO.RUBBISH.freeserve.co.uk> wrote:
>I'm new to this NG, so please forgive me if I am going back over a familiar
>(to you) topic, but I never heard of this VMS on Intel before.
>
>Does this really mean you can run VAX stuff on a PC?
Yes.
>Under windows/dos/linux, or instead of?
Under Windows 2000.
>Dual boot possible?
CHARON-VAX runs as an application under Windows, which means you
can still do Windows too. CHARON-VAX does chew up CPU time, though,
so your Windows apps will be slower.
>Does this mean you can carry a VMS app around on a laptop, or is it not
>quite that simple?
>
It means you can carry a VMS system around on a laptop, running whatever
apps you want to run. A container file is used for the system disk and
other disks.
>What does it cost?
>
Check the web site.
>You mentioned downloads and Software Resources International.
>
>Is there a URL where I could find out more?
>
Hunter Goatley
>On Fri, 20 Apr 2001 14:02:59 +0100, "Lock Horsburgh"
><lo...@lhorsburgh.NO.RUBBISH.freeserve.co.uk> wrote:
>
>>I'm new to this NG, so please forgive me if I am going back over a familiar
>>(to you) topic, but I never heard of this VMS on Intel before.
>>
>>Does this really mean you can run VAX stuff on a PC?
>
>Yes.
>
After seeing Paul's reply, I realized that I answered based on
the assumption that Lock was asking specifically about CHARON-VAX.
Hence the yes/no matching on our replies. ;-)
Paul Repacholi
So as the old saw goes, 'what answer would you like?' :)
Carl Nelson
home computer.
But my home computer is a Mac, not a PC.
But I *do* have Virtual PC running on my Mac.
My first question is: Why does my head hurt when I think about this?
My next question is: Does anyone have a Mac emulator that runs under
VMS?
Lock Horsburgh
I visited the software resources site and read up on CHARON-VAX.
Unfortunately they have discontinued the "hobbyist" download,
which would work on Win95, and it says the downloaded copies
will only work till April 2001.
The cheapest version seems to be $1900.
But what I didn't find was, how do you get a licence to run VMS etc
other than the hobbyist licence?
What do Compaq think of this, is it an opportunity for them to charge
whatever they like, cos you must be desperate? Or are they hostile,
cos they want you to buy Alpha boxes? Or don't they care?
Is anybody using this live, for commercial use, yet?
Lock.
Djordje Vukovic
>CHARON-VAX runs as an application under Windows, which means you
>can still do Windows too. CHARON-VAX does chew up CPU time, though,
>so your Windows apps will be slower.
How slow (or fast) will Charon itself feel on an average Wintel box-
compared to a real MicroVAX it emulates?
On a more general note- What is the general opinion on Charon- is it
good? Good enough to substitute a MicroVAX where it is really needed,
or is it just showing that it might be done?
Jan Vorbrueggen
> But my home computer is a Mac, not a PC.
> But I *do* have Virtual PC running on my Mac.
> My first question is: Why does my head hurt when I think about this?
It don't know - missing exposure to IBM's VM - now called OS/390, I believe -
which allows you to do this as a matter of course...like run VM again (for
instance a new version) as a virtual machine on top of VM and hosting another
virtual machine such as CMS or MVS...once virtual, always virtual.
Jan
Hunter Goatley
<lo...@lhorsburgh.NO.RUBBISH.freeserve.co.uk> wrote:
>Thanks to those who replied.
>
>I visited the software resources site and read up on CHARON-VAX.
>Unfortunately they have discontinued the "hobbyist" download,
>which would work on Win95, and it says the downloaded copies
>will only work till April 2001.
>
The Hobbyist version will return sometime soon....
>The cheapest version seems to be $1900.
>
Sounds about right.
>But what I didn't find was, how do you get a licence to run VMS etc
>other than the hobbyist licence?
>What do Compaq think of this, is it an opportunity for them to charge
>whatever they like, cos you must be desperate? Or are they hostile,
>cos they want you to buy Alpha boxes? Or don't they care?
>
I'm not sure what the answer to this is. Compaq is aware of CHARON-VAX;
presumably VMS would be licensed for it just as it would be licensed for
the hardware being emulated. But that's just my guess.
bill davidsen
This works really well... I once had a client who *had* to run an
Apple ][ program against some data. So we got an Apple ][ emmulator
program for CP/M on Z80, and a CP/M on Z80 emulator for UNIX, and ran
the whole thing under two levels of virtual machine.
It ran considerably faster that the original hardware.
Hans Vlems
VM is the Virtual Machine OS. It can run as host for as many guests as you
like. We used to run a couple of VSE guests under VM. And when a new VM
version
came out, it ran as a guest itself (second level VM).
The performance was surprisingly good: IBM knows this sport pretty well.
VMS might be adapted to do the same I think. But VM is a much simplerOS
than VMS was (since V2.0 or so)
Hans Vlems
Jan Vorbrueggen <j...@mailhost.neuroinformatik.ruhr-uni-bochum.de> wrote in
message news:y41yqkg...@mailhost.neuroinformatik.ruhr-uni-bochum.de...
Larry Kilgallen
> OS/390 is the latest name for MVS.
> VM is the Virtual Machine OS. It can run as host for as many guests as you
> like. We used to run a couple of VSE guests under VM. And when a new VM
> version
> came out, it ran as a guest itself (second level VM).
> The performance was surprisingly good: IBM knows this sport pretty well.
>
> VMS might be adapted to do the same I think. But VM is a much simplerOS
> than VMS was (since V2.0 or so)
The VAX instruction set as it stands cannot be virtualized on itself.
Kenn Humborg
This makes me curious... What is required for an architecture
to be virtualized on itself and what features of the VAX
instruction set prevent this?
References to any relevant online docs would be much
appreciated.
Later,
Kenn
Larry Kilgallen
> Larry Kilgallen <Kilg...@eisner.decus.org.nospam> wrote:
>>In article <9c20v4$buc$2...@news.IAEhv.nl>, "Hans Vlems" <hvl...@iae.nl> writes:
>>> VM is the Virtual Machine OS. It can run as host for as many guests as you
>>> like. We used to run a couple of VSE guests under VM. And when a new VM
>>> version
>>> came out, it ran as a guest itself (second level VM).
>>> The performance was surprisingly good: IBM knows this sport pretty well.
>>>
>>> VMS might be adapted to do the same I think. But VM is a much simplerOS
>>> than VMS was (since V2.0 or so)
>>
>
> This makes me curious... What is required for an architecture
> to be virtualized on itself and what features of the VAX
> instruction set prevent this?
Failed PROBE instructions cannot be intercepted as a fault.
Jan Vorbrueggen
> References to any relevant online docs would be much appreciated.
A privileged mode instruction must not be able to discern that it was in fact
executing in a less privileged mode. The PROBE problem Larry mentions is one
such case. Programmed I/O also often brings problems with it.
Jan