Sun Microsystems fulfilled a pledge Tuesday to release UltraSparc chip details in an effort to make it easier to bring Linux and versions of BSD Unix to its systems. Sun announced the availability of the specifications in conjunction with this week’s Open Source Business Conference in San Francisco, at which Sun President Jonathan Schwartz is delivering the opening keynote speech on Tuesday morning.
Great news! Sun hardware is great and I hope this move helps them sell more of it! Thank you Sun! Keep up the great work. You are winning minds and that has real value. I just hope that it pays for you.
“I’ve never been to hell, but I think it just froze over,” Schwartz said of the improbability of Stallman’s support.
That one made me giggle.
And BTW, it’s not just opening oportunities to optimise the kernels fot the sparc chips. If I read the article correctly, it means that anyone with the the right tools and knowledge will be able to make sparc compatible chips without asking anyone permision(well, except for sharing the design modifications to with the comunity, that is)
edit:I wonder what would it take to make a fork of that(surely there are more knowledgable people aroud here who can tell)
Edited 2006-02-14 20:51
Guess Theo and the gangs nagging finally paid off!
>Great news! Sun hardware is great and I hope this
>move helps them sell more of it! Thank you Sun! Keep
>up the great work. You are winning minds and that
>has real value. I just hope that it pays for you.
I can only hope you are being paid for this post.
I doubt if IBM or HP or Intel or AMD are about to start making Sparc processors. Linux is still going to be an x86 OS.
Sun should thank the people behind Apache, Gnome, KDE, MySQL, X, JBoss rather than sucking Linux kernel developer’s you-know-what!.
>I doubt if IBM or HP or Intel or AMD are about to
>start making Sparc processors. Linux is still going
>to be an x86 OS.
Linux is not a x86 OS now.
> Linux is not a x86 OS now.
Linux has MMX/SSE optimizations but where are the PowerPC Altivec or Sparc VIS optimizations?
OK the fact that you can run Linux on any processor doesn’t mean it’s “designed” for that processor.
AIX will ALWAYS beat Linux on POWER
Solaris will ALWAYS beat Linux on SPARC
HPUX will ALWAYS beat Linux on PARISC.
Accept this and move on…..I have!
Edited 2006-02-14 22:36
Linux has MMX/SSE optimizations
Where?
Sun should thank the people behind Apache, Gnome, KDE, MySQL, X, JBoss rather than sucking Linux kernel developer’s you-know-what!.
Linux allready runs on sparc e.g. Debian and there is a Centos in beta.
BTW, Sun do support other open source projects such as Apache, Gnome and Postgresql. I really can’t see how this will have any negative impact on Linux.
Ditch Solaris and move to Linux! *Slaps forehead*
Reasons why Linux/Intel beat Solaris/Sparc:
1.) Feature, stability, security is very similar
2.) Intel hardware is cheaper and faster than Sparc
3.) Linux is cheaper than Solaris *
4.) Linux doesn’t lock you to 1 vendor, Solaris does
5.) No Linux distributor has Jonathan Swartz as any sort of executive, that guy is crazy **
* = OpenSolaris nulls this point.
** = http://blogs.sun.com/jonathan
Solaris is decent and DTrace blows anything Linux has out of the water. It will be interesting to see how things happen in the future with the whole OpenSolaris community gains followers.
>I really can’t see how this will have any negative
>impact on Linux.
It won’t, unless no developers are interested and the port stagnates. At which point, Sun will demonize Linux.
OpenBSD developers have been asking for UltraSPARC specifications for years (see http://www.openbsd.org/lyrics.html#33).
“Solaris will ALWAYS beat Linux on SPARC”
That’s not sure anymore when the SPARC specs are out, that’s what this is all about!
I am really over the this OS beats this OS arguments. At the end of the day the final decision as to an operating system’s viability is determined by the personal preference of the end user.
Even in enterprise space, the manager/systems designers opinion will affect the final decision. Also haven’t you guys heard about the right tool for the right job?
A post about UltraSparc Specifications being released starts (excuse the foul language) p*ssing contests everywhere. I am glad no one has gone on to turn this from Linux > Solaris or Solaris > Linux into licence wars. The world is big enough for us all, atleast the OpenSource software world is…
Anyone here realise that this is kinda pointless given that they’re eventually going to move to SPARC V processors from Fujitsu?
Not that I think its a bad thing, but at the same time, this appearst to be more of a ‘here is some stuff’ to the OSS community to keep the kernel kiddies happy.
> ‘here is some stuff’ to the OSS community to keep the kernel kiddies happy.
More like keep the squeaky wheels like the Slashdot crowd happy and say nice things about Sun.
Actually Sun aren’t ‘going to move to the SPARC V
processors from Fujitsu’. This is a collaboration which,
currently, has a very definite shelf life (in terms of
product introductions). Sun still maintains one of the
largest (it was the second largest, not sure now) in-
house processor design teams and is actively working on
future generations of it’s own SPARC processors (i.e.
Rock and Niagara 2 names which are already in the public
domain). The APL was essentially to fill in a hole in
the processor roadmap left by the cancellation of the
UltrasparcV project. You will still be able to buy USIV+
, UIII+, T1 based systems from Sun.
Actually Sun aren’t ‘going to move to the SPARC V
processors from Fujitsu’. This is a collaboration which,
currently, has a very definite shelf life (in terms of
product introductions). Sun still maintains one of the
largest (it was the second largest, not sure now) in-
house processor design teams and is actively working on
future generations of it’s own SPARC processors (i.e.
Rock and Niagara 2 names which are already in the public domain). The APL was essentially to fill in a hole in the processor roadmap left by the cancellation of the UltrasparcV project. You will still be able to buy USIV+, UIII+, T1 based systems from Sun.
Now, not to sound bitter or anything, but how come the price/performance of SUN’s single processor line so, to use a rather blunt word, shit, when compared to what Fujitsu has to offer?
I would have thought the most logical thing to do would be to embrace SPARC V as a good successor to the current line up, and start (FINALLY!) updating those SPARC based workstations.
No word by SUN has been made one way or another on the future of their SPARC workstations and where Opteron fits into the large sceme of things – Opterons seem to be hanging off the company like a spare prick at a wedding – no real definitive role in the grand sceme of things.
Is the eventual move to phase out the UltraSPARC workstations in favour of using the Niagara 1 and 2 for servers, whilst relegating Opteron to workstations and servers where Niagara isn’t suitable? why all the secrecy surrounding the future moves of SUN? It seems to have all the stupidity and paranoia of Apple without the associated zing and hype that comes with an Apple product line up.
I think the bigger question ontop of that is this; when are we going to see SUN actually get off its behind and start pushing to get more ISV’s to develop software for Solaris as a workstation operating system alternative to Linux and other UNIX’s out there? Do SUN realise that the world doesn’t simply revolve around servers, because shock horror, IT infrasctucture consists of more than just that!?
Now, not to sound bitter or anything, but how come the price/performance of SUN’s single processor line so, to use a rather blunt word, shit, when compared to what Fujitsu has to offer?
Is it? It depends on what you are talking about – can
you give some examples of where USIV+ price/performance
is ‘shit’ compared to FJ?
I would have thought the most logical thing to do would be to embrace SPARC V as a good successor to the current line up, and start (FINALLY!) updating those SPARC based workstations.
How is that logical? Do you mean do exactly what HP
have done and become beholden to Intel to keep them in
the server game? As I said, processor design is very
much alive at Sun. An alliance with FJ will only
persist as long as it is beneficial to both parties.
I think Sun with the T1 based servers has shown it
can innovate and offer something different – it’s a
bit early to say jump to FJ.
No word by SUN has been made one way or another on the future of their SPARC workstations and where Opteron fits into the large sceme of things – Opterons seem to be hanging off the company like a spare prick at a wedding – no real definitive role in the grand sceme of things.
Sorry, I REALLY don’t get this. Can you explain what
you mean a bit better please. It seems to me that
Sun sell Opteron and SPARC based workstations (new
models appearing recently in both lines) – you buy
what you need. Same for servers. It’s called choice.
It makes a lot more sense if you look at it from the
perspective of a buyer who want’s run a particular
class of application (TLP, no TLP, FP, no FP etc).
This is obviously simplified but the point is that
hardware choice should be motivated by applications.
Is the eventual move to phase out the UltraSPARC workstations in favour of using the Niagara 1 and 2 for servers, whilst relegating Opteron to workstations and servers where Niagara isn’t suitable? why all the secrecy surrounding the future moves of SUN? It seems to have all the stupidity and paranoia of Apple without the associated zing and hype that comes with an Apple product line up.
I don’t think it’s secrecy. I think it’s lack of a
marketing and PR budget. The computer press in general
only seems to want to write ‘Sun is almost dead’ or
‘Solaris vs. Linux) stories.
I think the bigger question ontop of that is this; when are we going to see SUN actually get off its behind and start pushing to get more ISV’s to develop software for Solaris as a workstation operating system alternative to Linux and other UNIX’s out there? Do SUN realise that the world doesn’t simply revolve around servers, because shock horror, IT infrasctucture consists of more than just that!?
Well http://www.sun.com/desktop/index.jsp has a choice
of 7 workstations. I would say that the current ones
are the 20, 40 and 45 – thats two AMD64 based
workstations and one USIIIi based. How many do you
need?
> Anyone here realise that this is kinda pointless given
> that they’re eventually going to move to SPARC V
> processors from Fujitsu?
Is the SPARC V the same as the ‘UltraSparc T1 “Niagara”‘ mentioned in the article?
Last time I looked at Suns I think they were named UltraSPARC III, IIIx (or something), and IV.
Maybe the opening up of the specs is broader than you think? Or maybe it’s just all for show. I usually just wait to hear what the kernel uber-hackers have to say about it. If they give it the thumbs up, then I’ll have a look.
UltraSPARC is SUNs name for their SPARC based products, where as Fujitsu refer to theirs as SPARC64 followed by a revision number. Niagara is branded as T1 ‘with cool threads’ (yes, you’re not the only one who thinks that sounds really sad).
I would have thought that the bigger problem was the issue of the chipsets because right now, one could simply just write for the V9 SPARC specifications if they wanted to.
Bundle that with the fact that UltraSPARC IV is simply two UltraSPARC III cores dumped onto a single processor catridge, and as for VIS, they’ve of only any real use in the case where number crunching is required, for every day operations, I doubt they would yield much of an advantage.
The article mentions a love/hate relationship “with the GPL”. My guess is that it’s more like, Sun is really two small companies glued together: the OS side doing the software, and the hardware side designing and building servers.
The hardware folks probably love the GPL, especially wrt their hardware, since it makes it easier to use the hardware, and so will sell more of it.
The software folks probably don’t like the GPL, and thus GNU/Linux, since it bites into their domain. They’ve put *so* much work into Solaris, it’s gotta be very difficult to see it losing ground to this young upstart (GNU/Linux). *That’s* their problem.
Perhaps Sun should consider splitting the company, hardware and software. Let the hardware side open up specs and sell tons of hardware, and let the Solaris side compete directly with GNU/Linux to run on that hardware.
Personally, when I look at buying server hardware, I only care about 2 things: (1), Will I easily be able to install Debian Stable on it without having to dig around for drivers (which includes, will I be able to easily find a RAID monitoring util for it)? And (2), Will it be reliable enough that I won’t be taking many trips to the data center?
My guess is that most admins are like me in that regard. The problem is that, often, the purchase of new hardware gets tied up in red tape, with less-than-qualified people making decisions they shouldn’t be making. I believe some reasons this happens is because these people want to feel more important, and also they want the boss to pat them on the head for choosing the cheapest solution.