In article <Pine.SOL.4.64.0808240924040.19628@marrakesh>,
Rich Teer <rich.teer@rite-group.com> wrote:
>On Sat, 23 Aug 2008, Tim Bradshaw wrote:
>
>> This worries me. I hope that, somewhere, in Sun there is someone who
>
>Same here.
>
>> understands what their market is, and what markets they should attempt
>> to move into. Because desktops / laptops is not either of those.
>
>Not sure I entirely agree with this. I'd rather have (and in fact, do)
>Solaris running on my desktop and laptop than Linux (or any other OS).
>
>--
>Rich Teer, SCSA, SCNA, SCSECA
>
>CEO,
>My Online Home Inventory
>
>URLs: http://www.rite-group.com/rich
> http://www.linkedin.com/in/richteer
> http://www.myonlinehomeinventory.com

ME TOO!


Look -- if you're going to be administering the BIG sun servers,
I'd sure want one of at least the same species as my desktop!

(Why have to try to master some *other* kind of thing?)


David

In article <6h91gpFk4q8mU1@mid.individual.net>,
Ian Collins <ian-news@hotmail.com> wrote:
>David Combs wrote:
>> Like I say:
>>
>> what's with this suicidal rumor -- "NO MORE SPARC WORKSTATIONS!"
>>
>>
>> Huh?
>>
>>
>> Without sparc in workstations, why the hell not just
>> get a plain old PC?
>>
>Why indeed. There's very little development targeted at Sparc that
>can't be done on Solaris x86. A Sparc desktop will never be
>price/performance compatible with an x64 based one.
>
>> I also hear that this new, super-duper "openSolaris" (or whatever
>> it's called) is to be sun's future OS direction -- and yet they
>> haven't got it working on sparc, just x86.
>>
>Yet, it's targeted for the next major release (November?). The SXCE
>builds maintain parity between x64 and Sparc. Don't forget Sun's
>"OpenSolaris" distribution is unencumbered and freely distributable. It
>may be a bigger job getting an unencumbered Sparc release out the door.
>
>--
>Ian Collins.

Does it or doesn't it work on sparc?

(I was told by someone that this opensolaris stuff worked
on only x86. So was that info right or wrong?


Thanks!

David

David Combs wrote:
> In article <Pine.SOL.4.64.0808240924040.19628@marrakesh>,
> Rich Teer <rich.teer@rite-group.com> wrote:
>> On Sat, 23 Aug 2008, Tim Bradshaw wrote:
>>
>>> This worries me. I hope that, somewhere, in Sun there is someone who
>> Same here.
>>
>>> understands what their market is, and what markets they should attempt
>>> to move into. Because desktops / laptops is not either of those.
>> Not sure I entirely agree with this. I'd rather have (and in fact, do)
>> Solaris running on my desktop and laptop than Linux (or any other OS).
>>
[please don't quotes sigs]
>
> ME TOO!
>
>
> Look -- if you're going to be administering the BIG sun servers,
> I'd sure want one of at least the same species as my desktop!
>
> (Why have to try to master some *other* kind of thing?)
>
Most of the sparc specific stuff is big server only. There isn't much
(if anything) you can do on a sparc desktop that you can't on an x64 one.

Or is there?

--
Ian Collins.

In article <6727c041-be86-4342-abd7-3ddb04f56acb@34g2000hsh.googlegroups.com>,
Tim Bradshaw <tfb+google@tfeb.org> wrote:
>I have opinions, though I suspect that they are not worth much (why
>would mine be more interesting than anyone else's?), except for the

Am I correctly assuming that you are a Sun customer or partner?

If so, you're opinions about Sun's current market and market
opportunities are worth quite a bit if I understand the
market research industry correctly.

>"it's not desktop systems" one, which must surely be uncontroversial.

On which systems from Sun's competitors should developers service
your corner of the market and your suggested growth opportunities?

John
groenveld@acm.org

On Mon, 25 Aug 2008, David Combs wrote:

> Does it or doesn't it work on sparc?
>
> (I was told by someone that this opensolaris stuff worked
> on only x86. So was that info right or wrong?

OpenSolaris, as in what was previously known as Project Indiana,
is indeed x86-only at present.

What was previously loosly refered to as OpenSolaris, i.e.,
Solaris Express, does support SPARC in addition to x86. That's
one reason why I refuse to support the Project Indiana universe.

--
Rich Teer, SCSA, SCNA, SCSECA

CEO,
My Online Home Inventory

URLs: http://www.rite-group.com/rich
http://www.linkedin.com/in/richteer
http://www.myonlinehomeinventory.com

In article <g8um8h$fiq$1@f04n12.cac.psu.edu>,
groenvel@cse.psu.edu (John D Groenveld) writes:
> On which systems from Sun's competitors should developers service
> your corner of the market and your suggested growth opportunities?

Sparc servers nowadays are all multi-processor, and much
more multi-processor than any desktop systems. Your typical
sparc desktop system just isn't sufficiently representitive
of a server system any more for it to be a good enough
development platform. An application developed on a desktop
system can too easily find it can't use more than a couple
of cores on a server, as the author missed the importance of
designing for multi-processor systems.

Much better to get something like a small Niagra system and
stick it in computer room, and front this up with a Solaris
x86 desktop for access and windowing support. Then the
developer will see the importance of designing for multi-
processor systems in order to get the Niagra performing
well, and the application should then perform well on all
sparc server systems.

I'm struggling to see what the point of a sparc desktop is
nowadays, unless you're looking at something like an M series
supporting very many desktops via SunRay.

--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

David Combs wrote:
> In article <6h91gpFk4q8mU1@mid.individual.net>,
> Ian Collins <ian-news@hotmail.com> wrote:
>> David Combs wrote:
>>
>>> I also hear that this new, super-duper "openSolaris" (or whatever
>>> it's called) is to be sun's future OS direction -- and yet they
>>> haven't got it working on sparc, just x86.
>>>
>> Yet, it's targeted for the next major release (November?). The SXCE
>> builds maintain parity between x64 and Sparc. Don't forget Sun's
>> "OpenSolaris" distribution is unencumbered and freely distributable. It
>> may be a bigger job getting an unencumbered Sparc release out the door.
>>
[please don't quote sigs]
>
> Does it or doesn't it work on sparc?
>
Which part of "it's targeted for the next major release" didn't you get?

It only takes a couple of seconds to check out
http://opensolaris.org/os/downloads/

--
Ian Collins.

Ian Collins <ian-news@hotmail.com> wrote:
> Which part of "it's targeted for the next major release" didn't you get?
>
> It only takes a couple of seconds to check out
> http://opensolaris.org/os/downloads/

But, I'd like to point out that the majority of that page assumes you'll
be running on x86. There's only one instance of "SPARC" on the entire
page.

It's a matter of mindset. Someone used to Solaris, or even SX{C,D}E, will
be used to explicit identifiers, "this for x86, this for SPARC". Even
adding a few "(x86 only)" would improve things immensely.

And the roadmap to SPARC for the Sun OpenSolaris distribution is not
obvious. I know, I did quite a bit of digging before I found the "2008.11"
note myself.

--
Brandon Hume - hume -> BOFH.Ca, http://WWW.BOFH.Ca/

On Aug 25, 5:22*pm, groen...@cse.psu.edu (John D Groenveld) wrote:

> Am I correctly assuming that you are a Sun customer or partner?

No, but my relationship is at least that close.

> If so, you're opinions about Sun's current market and market
> opportunities are worth quite a bit if I understand the
> market research industry correctly.

Well, OK, what I meant was: if I had opinions I wanted to voice (other
than the desktop one) I'd do it to Sun, not in a newsgroup.

> On which systems from Sun's competitors should developers service
> your corner of the market and your suggested growth opportunities?

Well, there are two answers to this.

Firstly. When I last wrote code for my living I had a Linux desktop.
But I did not then and do not now think that Linux companies should be
concentrating on desktop / laptop systems. Note 1: that does not mean
they should not provide a GUI and IDEs, just that, for instance, they
probably should be expending more effort on other areas (edge servers,
say). Note 2: actually, there may be niches where Linux will thrive
on desktops/laptops/phones (the EEE for instance, as well as probably
an increasing chink of the phone market). It is probably a bit late
now for Sun to try and push Solaris in those markets.

Secondly. Due to the wonders of networks and networked window systems,
I really don't care very much what sits on a developer's desk any
more, and actually it hasn't mattered very much for a long time.
Indeed, when I had the Linux desktop I mentioned above, I don't think
I ever edited or compiled code on it. Actually, nothing is new here: I
didn't run much but the GUI on my first Sun worksation either (who
would want to on a 3/50?)

--tim

On Aug 25, 7:03*pm, and...@cucumber.demon.co.uk (Andrew Gabriel)
wrote:

> Much better to get something like a small Niagra system and
> stick it in computer room, and front this up with a Solaris
> x86 desktop for access and windowing support. Then the
> developer will see the importance of designing for multi-
> processor systems in order to get the Niagra performing
> well, and the application should then perform well on all
> sparc server systems.

Are modern compilers & build environments (I guess I mean Sun's, since
I'm not sure gcc counts as "modern", at least from what I hear
compiler people say) up to exploiting parallelism in designs like
Niagara / Rock? By this I mean can *builds* benefit from this, not
the built system? The reason for asking is people building big C/C++
systems used to want blazingly-fast single-threaded performance so
they did not die of caffeine poisoning while the system was building,
and this would not match Niagara very well.

It seams to me that this should be a pretty straightforward problem -
at least for multifile builds it should be trivial, and for single-
file builds it should be possible to parallelise compilation of
different functions relatively easily (at least for low optimisation
settings).

Other than that possible caveat, LDoms with their backing store on ZFS
sound like a fantastic dev/test platform to me - snapshots are a huge
win for these things.

(Obviously I agree with you about the rest...)