In article , "Tom Linden"
writes:

> On Tue, 30 Oct 2007 07:45:18 -0700, Richard B. Gilbert
> wrote:
>
> > Do consider, however, the fact that RISC architectures are generally
> > faster than CISC.
> >
> Not completely true. For simple things yes. Alpha had a 3:1 bloat factor
> over VAX

Actually, this just comes from being RISC. With memory so cheap these
days, it's not really an issue.

On 10/31/07 02:59, Phillip Helbig---remove CLOTHES to reply wrote:
> In article <4727437E.7060900@comcast.net>, "Richard B. Gilbert"
> writes:
>
>> Do consider, however, the fact that RISC architectures are generally
>> faster than CISC.
>
> Isn't RISC inherently SLOWER than CISC, but in practice this is more
> than made up for by the fact that the time between design and production
> for RISC is so much shorter than for CISC that Moore's law gives RISC an
> advantage? (In other words, we are comparing a RISC processor based on
> one-year-old hardware with a CISC processor based on three-year-old
> hardware, or whatever.)

No.

The genesis of RISC was the observations that most compiler writers
didn't use every exotic opcode, but just strung together sets of the
simple, fast ones.

They used that knowledge to build spare designs:
- only memory instructions are load/store
- no indirect addressing modes
- longword addressing
- *lots* of registers
- no microcode, all opcodes hard-wired

This simplicity just worked better than CISC because microcode is
slow, registers are fast, and the simplicity let them crank up the MHz.

Of course, RISC has been getting more complex ever since those first
ROMP, MIPS and SPARC chips. But that's "allowable" because chips
now have so many transistors that you can still maintain the essence
of RISC.

--
Ron Johnson, Jr.
Jefferson LA USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

On Wed, 31 Oct 2007 01:00:34 -0700, Phillip Helbig---remove CLOTHES to
reply wrote:

> In article , "Tom Linden"
> writes:
>
>> On Tue, 30 Oct 2007 07:45:18 -0700, Richard B. Gilbert
>> wrote:
>>
>> > Do consider, however, the fact that RISC architectures are generally
>> > faster than CISC.
>> >
>> Not completely true. For simple things yes. Alpha had a 3:1 bloat
>> factor
>> over VAX
>
> Actually, this just comes from being RISC. With memory so cheap these
> days, it's not really an issue.
>
It is not just memory, it has other design requirements such as bigger
memory bandwidth and larger caches



--
PL/I for OpenVMS
www.kednos.com

On Wed, 31 Oct 2007 03:13:59 -0700, Ron Johnson
wrote:

> On 10/31/07 02:59, Phillip Helbig---remove CLOTHES to reply wrote:
>> In article <4727437E.7060900@comcast.net>, "Richard B. Gilbert"
>> writes:
>>
>>> Do consider, however, the fact that RISC architectures are generally
>>> faster than CISC.
>>
>> Isn't RISC inherently SLOWER than CISC, but in practice this is more
>> than made up for by the fact that the time between design and production
>> for RISC is so much shorter than for CISC that Moore's law gives RISC an
>> advantage? (In other words, we are comparing a RISC processor based on
>> one-year-old hardware with a CISC processor based on three-year-old
>> hardware, or whatever.)
>
> No.
>
> The genesis of RISC was the observations that most compiler writers
> didn't use every exotic opcode, but just strung together sets of the
> simple, fast ones.
>
Well those observations were not particularly keen and largely based, AIR,
on analysis of C, Pascal and F77 and ignore Cobol and PL/I. Note that
the only Risc design to have done so was Power from IBM, which even
allowed unaligned access for a minimal (one tick) cost

> They used that knowledge to build spare designs:
> - only memory instructions are load/store
> - no indirect addressing modes
> - longword addressing
> - *lots* of registers
> - no microcode, all opcodes hard-wired
>
> This simplicity just worked better than CISC because microcode is
> slow, registers are fast, and the simplicity let them crank up the MHz.
>
> Of course, RISC has been getting more complex ever since those first
> ROMP, MIPS and SPARC chips. But that's "allowable" because chips
> now have so many transistors that you can still maintain the essence
> of RISC.

Actually if Dec had continued to a 64 bit VAX it would have looked that way
as well, compare z-series with 370
>



--
PL/I for OpenVMS
www.kednos.com

david20@alpha2.mdx.ac.uk wrote:
> VAX -> Alpha was seen as producing improved performance.
> Alpha -> Itanium is seen as a forced move where continued Alpha development
> would have produced better performance.

There are many more factors in today's computing world than raw
performance -- like price/performance, for one.

Many customers upgrading from Alpha to Itanium today are telling us they
are getting twice the performance for half the price. That's a 4X
improvement in price/performance as a reward for moving. And moving
across from Alpha to Itanium is much easier than VAX to Alpha was.

John Smith wrote:
> HP may well become Itanic's only customer, which puts it in the same boat as
> Alpha was in.

Alpha lost Windows support, and with it the associated volumes and
economies of scale.

Itanium runs Windows, Linux, HP-UX and other UNIX flavors, OpenVMS, and
NonStop, and benefits from the volumes of all those platforms. And
OpenVMS customers benefit from the lower hardware costs that volume allows.

JF Mezei wrote:
> The main reason that IA64 failed commercially is that it isn't
> compatible with windows,

Don't tell that to our many customers running Windows on Superdomes --
they might be very disappointed to find what they're doing is impossible.

When you can't scale-out with clusters like you can with OpenVMS,
sometimes the only solution is to scale-up, within a box, hence Windows
on Superdome.

On 11/01/07 15:57, Keith Parris wrote:
> John Smith wrote:
>> HP may well become Itanic's only customer, which puts it in the same
>> boat as
>> Alpha was in.
>
> Alpha lost Windows support, and with it the associated volumes and
> economies of scale.
>
> Itanium runs Windows, Linux, HP-UX and other UNIX flavors, OpenVMS, and
> NonStop, and benefits from the volumes of all those platforms. And
> OpenVMS customers benefit from the lower hardware costs that volume allows.

But -- compared to how many Opterons are sold each week, or Xeons
sold every day -- how many Itania (plural of Itanium?) are sold each
QUARTER?

--
Ron Johnson, Jr.
Jefferson LA USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

On 11/01/07 16:00, Keith Parris wrote:
> JF Mezei wrote:
>> The main reason that IA64 failed commercially is that it isn't
>> compatible with windows,
>
> Don't tell that to our many customers running Windows on Superdomes --

The very thought of Windows on Real Hardware just makes me ill.

> they might be very disappointed to find what they're doing is impossible.
>
> When you can't scale-out with clusters like you can with OpenVMS,
> sometimes the only solution is to scale-up, within a box, hence Windows
> on Superdome.


--
Ron Johnson, Jr.
Jefferson LA USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

Keith Parris wrote:
> Don't tell that to our many customers running Windows on Superdomes --
> they might be very disappointed to find what they're doing is impossible.

You're allowed to think what you want. Yes, there is still some old
version of Windows with very limited amount of software that runs on
that IA64 contraption.

But that is a far cry from IA64 claiming "imdustry standard" or "high
volume" status.

In terms of costs, if HP had wanted to build Alpha systems at low cost,
they could have. They simply elected to keep older designs. And they
artificially placed a price difference between Alpha and those IA64
things to help force customers choose the unwanted IA46 over the more
popular Alpha. Similarly, they changed VMS licensing to make it more
affordable for some, but only for IA64, not for the other platfoms.

On 11/01/07 19:43, JF Mezei wrote:
> Keith Parris wrote:
>> Don't tell that to our many customers running Windows on Superdomes --
>> they might be very disappointed to find what they're doing is impossible.
>
> You're allowed to think what you want. Yes, there is still some old
> version of Windows with very limited amount of software that runs on
> that IA64 contraption.

You might know a lot about VMS, but you apparently know squat about
Windows.

All of the current versions of the "big" Server software (Windows,
Exchange & SQL Server) are all released for ia64. And ia64 is a VC
builds target, so in-house development obviously goes on.

> But that is a far cry from IA64 claiming "imdustry standard" or "high
> volume" status.
>
> In terms of costs, if HP had wanted to build Alpha systems at low cost,
> they could have. They simply elected to keep older designs. And they
> artificially placed a price difference between Alpha and those IA64
> things to help force customers choose the unwanted IA46 over the more
> popular Alpha. Similarly, they changed VMS licensing to make it more
> affordable for some, but only for IA64, not for the other platfoms.


--
Ron Johnson, Jr.
Jefferson LA USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

"Jan-Erik Söderholm" wrote in message
news:3vNVi.12580$ZA.8128@newsb.telia.net...
> JF Mezei wrote:
>
> > The main reason that IA64 failed commercially...
>
> Did it ?
> Or do you want it to ?

Are your eyes closed, or is the content of JF's message just too
uncomfortable for you?

Of course IA64 has failed commercially to achieve its stated primary goal,
the one which was used as an excuse to kill Alpha in its favour. The goal
stated by Intel and HP was that IA64 was going to be the "industry standard"
64 bit architecture (e.g. volume dominant, most affordable, best supported,
etc). IA64 is not volume dominant, it is not most affordable, it is not best
supported in terms of choice of applications - it has failed in all of
those, AMD64 has won hands down in all of those, and AMD64 is likely to stay
successful for longer than IA64 is likely to be around.

The vast majority of folks buying IA64 are folks who are already heavily
committed to stuff which only comes from HP, folks who have no realistic
alternative, folks who are committed to NonStop, HPUX, and VMS. Those folks
are largely not interested in IA64 as such, they are interested in the
software and system benefits which VMS etc bring to their businesses. IA64
just happens to be HP's required silicon for the time being.

What features and benefits does IA64 silicon bring to the table for these
folk? Are any of them things which a true "industry standard" chip choice
couldn't also deliver? Those who have the choice of something other than
IA64 pretty much all choose to look elsewhere these days, and that is not
likely to change. HPTC is one of the few exceptions where Itanium has a few
wins, and you have to ask how much Intel marketing funding is used to tilt
the playing field here (as is common practice with Intel). Can IA64 really
be a viable medium term business for Intel and/or HP, when it is apparently
in the same no-USP, non-competitive-product "death spiral" which Alpha was
allegedly going down?

Anyway, once the Common System Interconnect Itaniums come out, you'll be
able to plug a true "industry standard" 64bit CPU (y'know, the AMD64 ones)
into an Itanium socket in an Itanium system and it'll work just fine, once
the software's been ported, and it'll magically have all the same "system
level" reliability features as an Itanium would, and superb system-level
performance, not to mention a better choice of software than Itanium, won't
it? Where does that leave IA64, where does it leave HP's Itanium-centric
system designers?


Wrt Unisys: er, afaict Itanium's on the way out. Unisys's "mainframe class"
kit now offers Xeon-based systems, not just IA64 (eg as reported
http://www.unisys.com/about__unisys/...s/05158777.htm). Some bloke
called Rich Marcello took charge of Unisys's enterprise server and software
business earlier this year, does he know anything about the enterprise
market, is he likely to know anything about whether customers who have a
free choice would prefer their hardware to be IA64 or AMD64?


"OpenVMS customers benefit from the lower hardware costs that volume
allows."

If one accepts that is the case, then surely it follows that OpenVMS on
mid/high-end Proliant-class boxes can do the "economies of scale" thing even
better, *unless* there is some technical uniqueness which only Itanium
brings to the OpenVMS table, or *unless* there is some other consideration
not yet mentioned. No one's found me a significant technical uniqueness yet
except maybe some of the aforementioned niche HPTC stuff, which isn't VMS
territory anyway. Therefore one might imagine that OpenVMS on mid/high-end
AMD64 Proliant would deliver the kind of hardware availability that keeps
98% of the server market happy today at the same hardware prices they're
paying today, and maybe for a little extra in terms of OS cost to buy (and
initial cost to learn) OpenVMS could provide software stability those
customers can only dream about today, with TCO's they can also only dream
about ("fit and forget" operations and no "Windows DataCentre" maintenance
costs), and sales volumes (and profits) for the VMS vendor that Marcello etc
could only ever dream about too.

So why doesn't it happen? I can't help wondering if maybe actually marketing
OpenVMS on non-niche hardware might risk seriously upsetting Microsoft, and
HP HQ wouldn't want that to happen would they? Redmond might tolerate HP
doing a bit of background stuff with Linux, NonStop is too niche for Redmond
to worry about, but HP actively marketing an HP-owned credible server OS on
non-niche hardware, to potentially compete with some aspects of Windows
Server high end, might not go down well in Redmond, and that could be bad
news for HP in general - not just systems, but services too, are dependent
on Microsoft-related revenues.

Will Microsoft still be around 30 years after the launch of Windows? What
state will Windows be in?

2p
John Wallace

Ron Johnson wrote:
> All of the current versions of the "big" Server software (Windows,
> Exchange & SQL Server) are all released for ia64. And ia64 is a VC
> builds target, so in-house development obviously goes on.

Unless Microsoft has reversed its previous decision, its support of IA64
is still very limited with only a few "server" packages available, and
nothing like Office or others popular software. And if I remember
correctly, Ia64 versions were to come later than for the industry high
volume stuff.

And frankly, do you serously believe that someone with pure business
goals would choose to run windows on some IA64 box ?

I can understand shops who have marketing associations with either intel
or microsoft who would deploy some windows on ia64. I can understand
some very high profile high performance shop getting some deal where the
subsidies for having windows on IA64 would be greater than running linux
on 8086 for instance. But do people really want windows on IA64 on
their own free will ?

On 11/01/07 23:08, JF Mezei wrote:
> Ron Johnson wrote:
>> All of the current versions of the "big" Server software (Windows,
>> Exchange & SQL Server) are all released for ia64. And ia64 is a VC
>> builds target, so in-house development obviously goes on.
>
> Unless Microsoft has reversed its previous decision, its support of IA64
> is still very limited with only a few "server" packages available, and
> nothing like Office or others popular software. And if I remember

Of course MSO won't be available for ia64.

> correctly, Ia64 versions were to come later than for the industry high
> volume stuff.
>
> And frankly, do you serously believe that someone with pure business
> goals would choose to run windows on some IA64 box ?
>
> I can understand shops who have marketing associations with either intel
> or microsoft who would deploy some windows on ia64. I can understand
> some very high profile high performance shop getting some deal where the
> subsidies for having windows on IA64 would be greater than running linux
> on 8086 for instance. But do people really want windows on IA64 on
> their own free will ?


--
Ron Johnson, Jr.
Jefferson LA USA

Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!

In article , Keith Parris writes:
>david20@alpha2.mdx.ac.uk wrote:
>> VAX -> Alpha was seen as producing improved performance.
>> Alpha -> Itanium is seen as a forced move where continued Alpha development
>> would have produced better performance.
>
>There are many more factors in today's computing world than raw
>performance -- like price/performance, for one.
>
>Many customers upgrading from Alpha to Itanium today are telling us they
>are getting twice the performance for half the price. That's a 4X
>improvement in price/performance as a reward for moving. And moving
>across from Alpha to Itanium is much easier than VAX to Alpha was.

Did you not see that I wrote "where continued Alpha development would have
produced better performance." Alpha development was stopped quite sometime ago
now.

David Webb
Security team leader
CCSS
Middlesex University

This post is meant as a QUESTION, not a comment.

Has anyone seen references to the 30th anniversary of VMS in various
media outlets ?

I didn't see it in Inquirer, nor in the HP press newsroom (aka: HP's
real press releases).

> -----Original Message-----
> From: JF Mezei [mailto:jfmezei.spamnot@vaxination.ca]
> Sent: November 1, 2007 11:09 PM
> To: Info-VAX@Mvb.Saic.Com
> Subject: Re: Happy Anniversary VMS - 30 years young
>
> Ron Johnson wrote:
> > All of the current versions of the "big" Server software (Windows,
> > Exchange & SQL Server) are all released for ia64. And ia64 is a VC
> > builds target, so in-house development obviously goes on.
>
> Unless Microsoft has reversed its previous decision, its support of
> IA64
> is still very limited with only a few "server" packages available, and
> nothing like Office or others popular software. And if I remember
> correctly, Ia64 versions were to come later than for the industry high
> volume stuff.
>

Ummm, since when did Office become a server application?

How many people install Office on Xeon based x86 servers?

Did I miss something?

> And frankly, do you serously believe that someone with pure business
> goals would choose to run windows on some IA64 box ?
>

As Keith stated, the answer is yes.

> I can understand shops who have marketing associations with either
> intel
> or microsoft who would deploy some windows on ia64. I can understand
> some very high profile high performance shop getting some deal where
> the
> subsidies for having windows on IA64 would be greater than running
> linux
> on 8086 for instance. But do people really want windows on IA64 on
> their own free will ?

As Keith stated, the answer is yes.

Regards


Kerry Main
Senior Consultant
HP Services Canada
Voice: 613-592-4660
Fax: 613-591-4477
kerryDOTmainAThpDOTcom
(remove the DOT's and AT)

OpenVMS - the secure, multi-site OS that just works.

"Peter 'EPLAN' LANGSTOeGER" wrote in message
news:47225da8$1@news.langstoeger.at...
> In article
> ,
> "Main, Kerry" writes:
>>What is this fascination you have with the main page on HP.com?
>
> What is your fascination with JF postings?
> Only because he is still putting the finger in the wound
> (while all other got bored repeating itself pointing to the obvious
> and stopped years ago)?
>
> Don't shoot the messenger...
>
> (though I often get bored of JF postings, too)
>


To be honest I seldom reply or pay attention to COV anymore because of a
couple posters (JF being high among them) who have made it their mission to
turn every positive into a negative, and to dwell on real and imagined
events in the past, and to imply evil intent to everything we do and
everything that happens.

"Richard Maher" wrote in message
news:fg1f6f$pdf$1@news-01.bur.connect.com.au...
> Hi Paul,
>
> In my opinion there is rarely a call, or justification, for the use of
> such
> intemperate language in civilized society; not that that has anything to
> do
> with COV (or the world at large :-)
>
> But I'm just curious as to why Robert Whininger hasn't come out in
> sympathy
> with you. I'd certainly hate to think that I was being stalked or that it
> was only my Web Pages that he felt compelled to critique publicly!
>

I can't imagine why Robert would be anxious to insult or be insulted by such
a swell guy like yourself. You are a model of civilized discourse.

But more than likely - the fires in San Diego caused a shutdown of the
facility that housed the news server where the feeds come in/out of HP. A
number of us stopped checking after a few days and only now are tuning back
in at all.