“It’s alive!” was the cry issued this week by the Solaris grunts at Sun Microsystems. For the first time, the engineers managed to get a true 64-bit kernel up and running on an Opteron box. This is a key milestone in Sun’s ambitious plan to make Solaris a preferred operating system in the x86 world.
Sun staff give birth to 64-bit Solaris on Opteron
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
for Solaris for AMD64. Good luck to Sun engineers, but they are a little late in the game.
If you want an easy to use 64-bit OS on your workstation right now, just install Mandrakelinux for AMD64, or SuSE for AMD64, or Fedora for AMD64, or… well, you get the idea.
This could be very good for SUN. A real 64-bit Solaris kernel running on commodity hardware… Well maybe not as “real” as SPARC, but still a very nice achievement that is sure to attract support from vendors and customers alike…
You forgot to mention Windows Server 2003-64 (beta), FreeBSD and NetBSD.
I can’t wait for Solaris x86_64.
Well I don’t think there is another true UNIX(C) available for x86-64 and only the laughable SCO-UNIX for x86. I’d say this is big news.
Also, I wouldn’t call a “plan to make Solaris a preferred operating system in the x86 world” “ambitious”.
Let’s see what we could call such a plan:
– impossible?
– a little overoptimistic, perhaps?
– Schwartzish?
– rosy?
– unrealistic?
– dumb?
OK, I am sure some Sun employee is going to post something in this thread to the effect that Solaris is already “a preferred operating system in the x86 world”, but perhaps we could get the real figures about how much revenue Sun derives from Solaris on x86 first, eh?
That is partly because x86 Solaris has been on again, off again. People have been reluctant to embrace a platform when the status is basically “we are going to release it on x86 but only because you guys won’t quit nagging us about it, so don’t expect the next version to support x86”
If Solaris actually embraces x86-64 and sticks with it, we may actually see it gain some popularity.
I personally like Solaris, packages are easy to install and it has more of a complete product feel to me than Linux. It is also very stable. Companies don’t have to support 40 different incompatible versions of Solaris in a 6 month period of time, and it shows.
CDE takes some getting used to, but is not that bad. I prefer the UNIX feel to CDE than the cheap wanna-be XP feel of KDE. I hate KDE and nearly everything about it.
With KDE and RPM hell being most peoples first impression of Linux it hardly surprises me that so many defunct Windows users go running back to MS.
I know there are other DE’s, I have tried most of them. The one I like most is probably Red Hat’s Bluecurve, but Fedora is too rough around the edges for my taste.
The zealots should consider another platform to advocate, becasue Linux is running out of cool points.
Also, I wouldn’t call a “plan to make Solaris a preferred operating system in the x86 world” “ambitious”.
I think you are misunderstanding the phrase ” a preferred” and ” the preferred“
While this is cool news, having a true UNIX on x86. and your right its questionable if they will truly support x86 solaris. Theres still a good chance especially name regonition. People equate sun with enterprise and that might get them customers.
As for Linux. while the HYPE is slowing down and people are realizing its not perfect, id have to say its still Good. Linux is Unix for the Masses as Id like to put it. technically its not really unix but its unix-like enough to most people to think of it as a unix replacement. Its Cheap(commodity), has Good commodity hardware support(as in its getting recognition from other companies like NVIDIA) , its multiplatform(meaning it doesnt need to run on x86 and in fact runs on a ps2 also ;-). Its MORE secure than Windows(well defaults anyways. Especially since most Joe users dont need apace or web servers installed). and its graphics are improving(most people wont be complaining that KDE is pro-GUI as people now a days besides unix geeks really love CLI that much ;-).
This is laughable.. Solaris is dead. The best example is that they are thinking to open-source it (it means they don’t want to develop it anymore).
Solaris has a long reputation of being insecure, inefficient, and now they show that they’re months (years?) ahead the open-source kernel which is already 64-bit for opteron for long.
Seriously, who wants to use solaris on an opteron today? Sun is going to die for real if they are unable to face the reality of Unix future as it is.
“”I know there are other DE’s, I have tried most of them. The one I like most is probably Red Hat’s Bluecurve, but Fedora is too rough around the edges for my taste. “”
Evidentally you do not know that much, as Bluecurve is a THEME for GNOME, not a DE. Also, if you try any distro with GNOME 2.6 and a decent theme installed(covers all things, mostly icons), and automaticlaly detects most things, you would not say it is “rough around the edges” in fact Fedora Core 2 is less “Rough Around the Edges” than Windows XP.(and also performs better, with a more appealing GUI)
Maybe you should focus on Solaris(the topic) and not bash something you obviousely havn’t tried lately.
This is very good news for AMD.
A lot of companies swear by Sun, and I can’t blame them really, they’ve always very nice hardware — sun fire v880, yummy — and Solaris isn’t as bad as gc (IP: —.urbanet.ch) would have you believe.
This way companies will be able to get — probably less expencive! — Opteron machines and still run Solaris and all their Solaris software on it. That means no re-training staff!
Having installed many flavours of Linux distros in the recent past, I thought I’d try out Solaris for x86, but was utterly dismayed that it didn’t understand extended partitions, meaning that unless I bought a second hard disk or completely wiped my first one, I couldn’t install it. Maybe whilst they’re cooing and clucking over x86_64 Solaris, they could improve the installer (which I thought was quite user-hostile – almost Slackware/Debian-like in its awkwardness) while they’re at it?
> Solaris has a long reputation of being insecure
Dude, are you really that dumb or you just pretend to be? Your post is laughable in tune with the subject for you post. Solaris sure as hell has much better security track record than Linux and the majority of other Unix flavors out there. Don’t splurge about something have absolutely no clue about. And if vanilla Solaris is not secure enough for your, there is always Trusted Solaris (something that NSA, US Army/Navy/Air Force use pretty extensively).
This is great news for the x86…
It is difficult to understand why anyone can say that Solaris is dead, or that it is insecure, just look in the server rooms of any bank or military establishment. And to criticise the efforts of something that can only enhance computing? Maybe ambitious, but Sun are the ones to make it happen.
It will be good to see proper mature 64 bit unix available to people that want it without having to buy expensive UNIX hardware. Linux doesnt cut it for me, it is mostly an amateur effort and it shows –
I totally agree with IanR (IP: —.bridgeway.co.uk) and Jim (IP: —.bflony.adelphia.net). It turns out that nowadays Linux is playing for the masses, but unfortunately not everyone has had the chance to deal with stable rocksolid OS/OE Server as Sun Microsystems products. Now, all we hear is Solaris is dead.. Solaris x86..bla bla bla.. even Andrew (IP: —.fbx.proxad.net) said: or Solaris for AMD64. Good luck to Sun engineers, but they are a little late in the game. If you want an easy to use 64-bit OS on your workstation right now, just install Mandrakelinux for AMD64, or SuSE for AMD64, or Fedora for AMD64, or… well, you get the idea.
See… he argues about a workstation OS while the people who really uses Solaris would never care about Solaris running on a AMD64 workstation (besides CAD/CAM and so..)
Weird..weird world…
KUDOS TO SUNS ENGS!!!
Also, regarding the workstation comments, I would say that we should wait till SUN has released a Opteron workstation with a nice 3D Labs Wildcat card. If they make a good splash, it will stop the leakage.
If customers decide to move to x86, they can also keep with the same operating system. That is not only a win for the customer, but a win for SUN Microsystems.
What will be the deciding factor, however, is how well they market their line, and unfortunately they do a really crap job. SUN is like Digital, a company run by engineers. Sounds great, but when it comes to the touchy-touchy feely-feely work of marketing, engineers are the last group you rely on to hype the products to the PHB’s of the world.
“I think you are misunderstanding the phrase ” a preferred” and ” the preferred”
Not really: “a preferred” is just impossible, dumb, Schwartzy; “the preferred” would be ridiculous, dumberer, Schwatzier – or a typo.
Semantics apart, I don’t see how getting an alpha AMD64 kernel version running in a Sun lab is going to make the 32-bit x86 Solaris version any less unpopular than it is now.
Solaris on SPARC is certainly not dead and has a healthy, wealthy ($$$) future ahead, but both Solaris on x86 and Solaris on AMD64 having a significant market share seems like a very long shot for my crystal ball here. OTOH, besides the FUD that Micro$haft and now Sun have been throwing around, I see real Linux market growth at present and in the future.
But again, if you have any figures to show for the contrary, I am all ears…
I agree, it’s unfortunate that most have never used a robust os like solaris, with top notch support. We installed a linux oracle cluster, it was our first and may be our last. Compared to our sun boxes it is complete crap. Despite having experienced linux admins, tons of support from oracle, redhat and dell, it still took 2 months before it stopped core dumping, and that was just the first problem. Not to mention that cores with linux are not like a solaris core. The Solaris dumps are automatically written to disk and you can send it off the same day to Sun engineers who will find the root of the problem within 24 hours and tell you how to fix it. Added to the fact that if it is a hardware problem, they don’t just point at a vendor, they tell you the exact piece that caused the problem and send you a firmware update or a replacement the same day.
I doubt x86 64 on opteron is a big deal, since solaris users run sparc, but who knows, it definitely gives sun and amd more options. If they can support it like they do thier sparc platforms, and drive the cost down at the same time.. I’m in. Until you have worked with a single hardware/software vendor like sun, you don’t realize how good it is. No finger pointing, no problems with scaling, just a simple “we can get it done” or a “yes there was a problem, we’ll fix it, hardware or software”. Plus thier support staff is skilled and can actually fix your problem…its amazing, almost old world like, you know early 90s.
Semantics apart, I don’t see how getting an alpha AMD64 kernel version running in a Sun lab is going to make the 32-bit x86 Solaris version any less unpopular than it is now.
Solaris x86 is unpopular now becuase Sun wasn’t putting much resources in to it. Now with a new line of Opteron systems and thier kealia acquisition there is no question that Sun is 100% behind it. That is what will make the difference.
It is not a alpha version, opteron support is supposed to be in Solaris 10. When Solaris 10 is ready opteron support won’t be Alpha quality.
Solaris on SPARC is certainly not dead and has a healthy, wealthy ($$$) future ahead, but both Solaris on x86 and Solaris on AMD64 having a significant market share seems like a very long shot for my crystal ball here.
That is is problem with crystal balls the only predict the future, unreliably, and wihtout context. I think you shoud give up supersition and rely on facts. Solaris x86 had enough mindhare that Cutomers/Users urged Sun to revive it when they cancelled it. What makes you think that those exact same people and more won’t be thrilled to use it, fully supported by Sun and hundred percent backed by Sun?
http://www.save-solaris.org/estrin-2003-09-04.html
http://my.execpc.com/~keithp/bdlogsol.htm
Sun is now selling x86 boxes and I would expect that Sun to fully support it like thier SPARC boces. Sun will eventually release workstations and server upto 8 way based on Opteron. So opteron is not just a lab experiment it is real. Sun’s x86 boxes are lower in price than the competition, even Dell. Sun also sells linux on these by the way. But if Solaris 10 is what it is promised to be with all the features and also linux binary compatibily. I think Sun has a very compelling story for Solaris to become a preferred platform on x86. Ecspecially if they open source Solaris that would leave all the linux zealots in the dust with thier Open Source == freedom argument. That would level the playing field so to speak, linux and Solaris will compete on merits and not religious methodologies.
Sun has already signed up with more than 20 OEMs for x86 platforms. That is quiet a jump from where they were before.
http://www.linuxworld.com/story/44960.htm
I doubt x86 64 on opteron is a big deal, since solaris users run sparc, but who knows, it definitely gives sun and amd more options. If they can support it like they do thier sparc platforms, and drive the cost down at the same time.. I’m in. Until you have worked with a single hardware/software vendor like sun, you don’t realize how good it is. No finger pointing, no problems with scaling, just a simple “we can get it done” or a “yes there was a problem, we’ll fix it, hardware or software”. Plus thier support staff is skilled and can actually fix your problem…its amazing, almost old world like, you know early 90s.
I’d say it is a big deal, there are already companies looking at replacing their Windows and Linux servers with Opteron/Solaris/Java Enterprise System.
The ability to utilise existing Solaris knowledgable staff whilst being able to move to a cheaper platform is the main attraction, and with the focus on developing Quad and Dual Opteron for the low end market will enable SUN to compete with HP/IBM and Dell, without compromising on quality or the “unified product of software and hardware”.
Just a small addendum to my previous post.
You only have to look at the signatories in the petition to revive solaris to realize that a lot of key IT people in very major organizations have signed.
http://www.save-solaris.org/signatories.html
1320 signatories from all over the world and important organizations like world bank, harvard, Motorola, Cisco system ………..
I would say that is significant mindhare and now they can make thier decisions with full confidence that Sun will fully back x86. nuff said
The zealots should consider another platform to advocate, becasue Linux is running out of cool points.
Well, if I did that I wouldn’t be a very good zealot, would I?
BTW, the coolest point about Linux is its licensed under the GNU General Public License. And that’s why I’m proud to be a Linux zealot. hehe
But since we’re all talking about Solaris X86, I was wondering, how does it compare to Solaris Sparc? Are all those Solaris zealot comments about stability, security, performance and NSA certification talking about the X86 product or the Sparc product? Are these two based on the same kernel, most of the same code in the kernel, or almost none of the same code? I heard the Solaris Sparc kernel was very modular and well designed. Does that apply to the Solaris X86 kernel also? I believe, from the perspective of a zealot, that these points are important to clarify.
But since we’re all talking about Solaris X86, I was wondering, how does it compare to Solaris Sparc? re all those Solaris zealot comments about stability, security, performance and NSA certification talking about the X86 product or the Sparc product? Are these two based on the same kernel, most of the same code in the kernel, or almost none of the same code?
As it turns out, we used this incredibly advanced technology called a compiler to generate binaries for different architectures from the same source code. (Apologies for the (bitter) sarcasm, we just get very sick of hearing this question/concern.) So yes, they’re the same source base. In fact, pretty much all of us in kernel development do all of our development from either our x86-based laptops, or our Opteron-based workstations. (And a handful of lucky ones have Opteron-based laptops.) We think of the two ISAs as being equal — there is not a first-class citizen — but if one were forced to distinguish between them, the edge would have to go to Solaris x86…
“Solaris x86 is unpopular now becuase Sun wasn’t putting much resources in to it. Now with a new line of Opteron systems and thier kealia acquisition there is no question that Sun is 100% behind it. That is what will make the difference.”
You do realize, of course, that x86 is an old 32-bit architecture, whereas AMD64 is a different, new, 64-bit architecture?
So how does the fact that Sun is taking the first steps to implement Solaris on AMD64 rehabilitate the fiasco of Solaris on x86?
“It is not a alpha version, opteron support is supposed to be in Solaris 10. When Solaris 10 is ready opteron support won’t be Alpha quality.”
Read the article. The author himself is slightly skeptical that Solaris 10 will have support for AMD64 at launch time (next January if I understood it correctly). So right now the AMD64 Solaris kernel they have running in a lab is an alpha version.
“That is is problem with crystal balls the only predict the future, unreliably, and wihtout context. I think you shoud give up supersition and rely on facts.”
I guess my irony was lost on you…
“Sun is now selling x86 boxes…”
OK, but how many are selling with Solaris x86? What is the actual revenue Sun is deriving from Solaris x86?
If you have any numbers I would be willing to consider your arguments but until now all I have seen is heavily biased opinion.
> kernel, or almost none of the same code? I heard the
> Solaris Sparc kernel was very modular and well designed
> Does that apply to the Solaris X86 kernel also? I believe,
For quite a few Solaris releases now (since I think
Solaris 2.4), the x86 version has been built from the
*exact* same source code as the SPARC version. Yes, the
Solaris kernel is modular and well designed and that’s
true with the x86 version as well. The primary difference
between the versions feature wise had to do with the
differences in the hardware and driver support. By the
vast amount of source code was generic and the same
functionality was and is shipped for both platforms.
Someone earlier mentioned Trusted Solaris – this also is
available for both SPARC and x86 platforms.
“As it turns out, we used this incredibly advanced technology called a compiler to generate binaries for different architectures from the same source code.”
Misplaced sarcasm and also incorrect. The kernel source code cannot just be recompiled for a new architecture.
I would be suspicious if any Sun engineer made such a statement, so I can only imagine that you are faking your IP address. Either that or Sun is in worse shape than I thought.
N.B.: another factual mistake is claiming that some of the Sun engineers are using Opteron laptops. AMD64 laptops are all based on the 754-pin Athlon64 chip, which is not an Opteron (but has the same instruction set, obviously).
You do realize, of course, that x86 is an old 32-bit architecture, whereas AMD64 is a different, new, 64-bit architecture?
The term x86 has now come to encompass any chip that implements the intel IA32(x86) instruction set. Opteron fully implements the IA32(x86) instruction set, any binary compiled for x86 should run unmodified on any opteron-based chip. AMD64 likewise is a 64-bit instruction set extension to the x86 instruction set not a new architecture.
Solaris x86 binaries should run on an AMD chip unmodified in 32-bit mode, infact Sun ships Solaris on the v20z which is an opteron based server.
Architecturally, Pentium 4s and Pentium Ms are very different but they are x86 cpus nonetheless. AMD Athlon and Athlon64 are architecturally different but they too are x86 cpus.
Don’t confuse Architecure and Instruction sets.
Misplaced sarcasm and also incorrect. The kernel source code cannot just be recompiled for a new architecture.
Of course not — there are obviously components that are instruction specific that need to be ported to a new architecture. Once that port is complete, however, Solaris is Solaris — regardless of platform. The vast (vast, vast!) majority of work in Solaris is ISA-neutral. Hence my sarcasm about the compiler: people tend to forget how much code can just be recompiled for a new ISA…
I appreciate your sarcasm.
And having a Linux background I understand how much of the kernel can be recompiled for another platform. But your comments about the Solaris kernel are insightful. However you did not completely answer my questions. Some of them pertained to stability. Can you shed some light on the testing and QA that went into the AMD64 port for us lazy sys admins?
From my experiences with Linux I would suspect there are some problems with some Linux AMD64 systems in a production environment because its still relatively new. I wonder if Sun will be nearly as stable as Linux (or possibly more stable, depending on hardware and QA) or if it will take a few months/years to find out.
“Don’t confuse Architecure and Instruction sets.” (sic)
I didn’t. The confusion seems to be within Sun: porting Solaris to AMD64 is not going to help Solaris x86, no matter how much corporate wishful thinking goes into it. They are two distinct products, one of them still in development (Solaris AMD64) and the other a market failure (Solaris x86).
Only very flawed logic could make the revival of Solaris x86 dependent on the hypothetical success of Solaris AMD64.
But with so many Sun engineers posting here on OSNews, I wonder why you haven’t yet answered my simple question about the revenue generated by Solaris x86: let’s talk numbers, eh?
They are two distinct products, one of them still in development (Solaris AMD64) and the other a market failure (Solaris x86).
Why is Solaris with AMD64 support not the same as Solaris x86?
Solaris SPARC runs on Sun’s UltraSPARC hardware and Fujitsu’s SPARC hardware.
http://www.fujitsu.com/services/computing/server/unix/os/
I would hazard a very educated guess that there will only be one Solaris x86 version that also supports AMD64 (whenever that is).
So please get it into your head that opterons are x86 cpus and Solaris x86 will work and fully support them (64-bit) eventually, Just as Solaris SPARC supports a variety of SPARC cpus.
As others have pointed out majority of the Solaris kernel is platform independant. To educate your self read Solaris Internals By Jim Mauro and Rich McDougall.
Only very flawed logic could make the revival of Solaris x86 dependent on the hypothetical success of Solaris AMD64.
As I mentioned above, I would hazard a guess that they are one and the same thing and one is not mutually exclusive of the other.
N.B:- I don’t have information about product releases and Marketting names. Things could very well be different at time of release. My guess is based on how we have done things in the past.
DISCLAIMER:- I don’t speak for my employer Sun Microsystems Inc. The view expressed here are my own.
N.B.: another factual mistake is claiming that some of the Sun engineers are using Opteron laptops. AMD64 laptops are all based on the 754-pin Athlon64 chip, which is not an Opteron (but has the same instruction set, obviously).
Are you just being Anal?
The opterons and Athlon 64s are not only instruction set compatible they use the same core(pipleline).
The only differentiator among the Athlon 64, Athlon 64 FX and the Opteron are the intergrated memory controllers for various price points and markets. The pin count difference of 185 seems about right given that the Athlon 64 has a single channel memory controller. Infact, there are new 128-bit 939 Athlon 64s.
So for all intents and purposes all Athlon 64 chips were derived from the opteron since it debuted first. They all share the smae K8 cores (Hammer).
Opteron and Athlon are marketing names and most of these are essentially the same CPU as visible to operating system software.
Is there any plan to have people test Solaris (64-bit kernel and userland) on Opteron, beside Sun engineers ? That would be an easy way to weed out bugs before next year.
@ AR:
“Why is Solaris with AMD64 support not the same as Solaris x86?”
For the very same (obvious) reason that Solaris with x86 support is not the same as Solaris SPARC.
“So please get it into your head that opterons are x86 cpus…”
You have a strange way of arguing. AMD64 CPUs are backwards compatible with x86 CPUs, that we all know. But (obviously again) a 64-bit version of Solaris running on AMD64 CPUs is quite a different product from Solaris x86 – clearly just as different as Solaris SPARC.
But please, instead of your guesswork, can you give us the numbers: how much revenue does Sun derive from the sale of Solaris x86? What is the market share of Solaris x86? And how would that (disastrous) situation improve with a (too late) release of Solaris AMD64 sometime in 2005?
@ Warren:
They just got an alpha kernel running. You’ll have to wait until they reach beta stage for a test version. Don’t hold your breath, though.
You have a strange way of arguing. AMD64 CPUs are backwards compatible with x86 CPUs, that we all know. But (obviously again) a 64-bit version of Solaris running on AMD64 CPUs is quite a different product from Solaris x86 – clearly just as different as Solaris SPARC.
Bullshit AMD64 is not bacjwards compatible with x86, it is x86 with extensions. All you have showed in your arguments on this forums is ignorance for anything technical.
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_93…
Only a cursory glance at the FAQ should tell you that AMD64 is x86 extended to 64-bit.
“Q: What is AMD64?
A: AMD64 is AMD’s 64-bit platform that extends the industry-standard x86 instruction set architecture and defines a new class of computing. The AMD64 platform is the first designed to deliver full compatibility with existing x86 solutions and world-class 64-bit performance.
Q: What is the AMD64 instruction set architecture?
A: The AMD64 ISA is an AMD innovation that extends x86, the industry’s most widely supported instruction set, and is designed to enable 64-bit computing while remaining compatible with the vast x86 infrastructure and industry experience. ”
AMD64 chips are chiefly x86 cpus with 64 bit extensions.
“Bullshit AMD64 is not bacjwards compatible with x86” (sic)
Wrong.
That’s exactly the main point of running an AMD64 CPU in 64-bit mode: you get the advantages of a 64-bit ISA with full backwards compatibility with your 32-bit applications.
“All you have showed in your arguments on this forums is ignorance for anything technical.”
Not really.
OTOH I am still waiting for any numbers that would show how much revenue Sun derives from Solaris x86. But you can’t come up with those, can you?
N.B.: you should use a grammar checker.
How typical of those who have run out of arguments to start picking on grammar and spelling…
When Solaris is ported to as many hardware platforms as NetBSD or even Linux then they should talk. Unitl then, choosing Solaris still leads to platform lock-in (albeit two platforms now). Somehow, I can’t imagine Solaris ever running on a Power5 machine from IBM. Might be interesting to compare AIX and Solaris at that point. Have they even got Solaris running on Itanium2? Then we could compare Solaris with HP-UX. These would be true apples-to-apples comparisons- one commercial unix vs another on the same hardware.
There was a project at Sun to port Solaris to IA64 (Itanium), back when Itanium was still called Mercedes.
AFAIK it was cancelled and Sun decided to continue SPARC development with TI as its foundry.
But (obviously again) a 64-bit version of Solaris running on AMD64 CPUs is quite a different product from Solaris x86 – clearly just as different as Solaris SPARC.
This is a ridiculous statement. If you’re going to compare it to anything, try AMD64 is to IA32 as SPARCV9 is to SPARCV7. They are both extensions of an existing architecture, which means that the majority of the code do not need to be modified. Do you really think that there are two versions of everything in Solaris, one for sparcv7 and one for sparcv9? Why should it be any different for amd64? You can’t even argue that they have to port to the new world of 64-bit computing; they’ve already done it once for SPARCV9 so all their code is probably 64-bit clean already. Saying that amd64 is as different from x86 as it is from SPARC just shows a lack of technical expertise on your part.
I work at a military base, and Solaris is used in places where Linux and Windows are not approved. Insecure? Out of the box normal Solaris may not be all that secure, but Trusted Solaris is the shit. Great stuff.
“AMD64 is to IA32 as SPARCV9 is to SPARCV7”
I don’t agree with you here, and since you base your entire argument on this assumption, then I think there is nothing else to discuss.
Also you should re-read my phrase: I was talking about the Solaris product that Sun is hypothetically going to sell.
Of course Sun can decide to market Solaris for AMD64 as just an extended version of Solaris x86. Who cares?
That is not sun’s idea. It’s AMD’s. But of course you know better than them. You are a genius and the rest of us are all really stupid.. What was your name again? Bobby Fischer?
Bullshit AMD64 is not bacjwards compatible with x86″ (sic)
Wrong.
May be you should learn to post things in context. I said “it is x86 with extensions”. Backward compatibility is not a feature of AMD64 that will be dropped in the future. The term “Backward compatibility” implies that something may not be in the future becuase the some thing will replace it.
AMD’s own FAQ carefully never mentions “backward” and “compatibility”. The mention that thier new architecture is compatible with existing stuff, which is a completely different thing.
The term “Backward compatibility” implies that something may not be in the future becuase the some thing will replace it.
Before the grammar nazi (a person who has no real arguments) complains.
The term “Backward compatibility” implies that something may not be available in the future becuase the some thing else will eventually replace it.
AMD64 is not backward compatible with x86. It is fully compatible with x86, AMD says so.
http://iforce.sun.com/partners/opteron/faqs.html
Q. Do I have to rewrite, recompile, or repackage my existing Solaris x86 applications to run on the AMD Opteron processor-based systems from Sun?
A. No. Most existing Solaris x86 binaries should simply run on these systems, and in many cases will experience a performance increase due to the AMD Opteron processor’s innovative architecture which eliminates many memory and I/O bottlenecks. These 32-bit applications should work whether these systems are running the current 32-bit Solaris x86 operating system, or the forthcoming 64-bit version of Solaris x86.
I think this should quell the discussion that Solaris x86 will work in 64-bit mode and there won’t be a Solaris AMD64.
That’s the problem with ignorant people and their ridiculous arguments.
Also, Sun’s current revenue from Solaris x86 is irelevant. Sun does not charge for Solaris licences for below 8 way systems even on SPARC systems. Also Sun is just starting to agressively ramp up thier Solaris x86 solutions, remember they cancelled thier x86 version, the petition to revive it was posted. So it is premeutre to use revenue data.
Solaris x86 was a so called “failure” (define failure, the petition signatories are admins from big companies/orgs) was because Sun never really backed it as a product, Sun is now.
“Sun’s current revenue from Solaris x86 is irelevant.”
I would say Solaris x86 is irrelevant. Probably Solaris x86 with 64-bit extensions, whenever that’s ready, will be irrelevant too. It’s too little, too late.
The hard truth is, Sun needs to ship their Opteron servers with a true 64-bit OS, and they still cannot face the fact that 99% of their users presently install Windows or Linux on their x86 hardware.
Funny things is, Micro$haft will probably force Sun to install and support Windows Server 64 bits, at some point in the future.
When you derive zero revenue from a product and Bill Gates puts a $1.95 billion check on the table, well, it changes your perspective on things, doesn’t it?
N.B. : I am not a “grammar Nazi”. But like any well-educated person, I would say your constant littering of your posts with both grammar and spelling mistakes is, to a certain point, irritating. Also you have a tendency to denigrate people, as shown on various threads here at OSNews.
I would say Solaris x86 is irrelevant. Probably Solaris x86 with 64-bit extensions, whenever that’s ready, will be irrelevant too. It’s too little, too late.
That’s your perspective, fine. I believe in having facts before making judgement. When speaking of Solaris x86, even Scott Mcnealy thinks they didn’t back the product enough. But you claim that it was a failure. How can it be a failure if enough effort was never put into it.
The hard truth is, Sun needs to ship their Opteron servers with a true 64-bit OS, and they still cannot face the fact that 99% of their users presently install Windows or Linux on their x86 hardware.
Incase you haven’t noticed that is what this whole article is about. They are preparing to ship a 64-bit Solaris x86. Sun supports and sells linux on thier x86 boxes and also certifies them. Critics of Sun claimed that Sun wasn’t shipping linux and now that they have, they claim that they aren’t shipping a 64-bit version of Solaris.
Funny things is, Micro$haft will probably force Sun to install and support Windows Server 64 bits, at some point in the future.
No Sun’s customers might. That is a very naive view of things to say the least.
When you derive zero revenue from a product and Bill Gates puts a $1.95 billion check on the table, well, it changes your perspective on things, doesn’t it?
Sun derives revenue from hardware sales and services, ecspecially for low-end boxes. Since opterons are low-end boxes (less that 8 cpus) how does the Solaris revenue matter? Sun would make money with the hardware sales and services even if the boxes ran linux.
N.B. : I am not a “grammar Nazi”. But like any well-educated person, I would say your constant littering of your posts with both grammar and spelling mistakes is, to a certain point, irritating.
Have you heard off the word “typo”. “Bacjward” the “j” key is very close to the “k” key and in a hurry it is very easy to hit it.
Also you have a tendency to denigrate people, as shown on various threads here at OSNews.
I also concede points where I am wrong and say it, However, just go back through our own discussion and coun the number of times you conceded a point.
Furthermore, People living in glass houses….. Weren’t you the one who claime d Bryan Cantrill was not qualified to make his statement. I think Bryan knows more about Operating Systems in his sleep than you would wide awake and fully alert.
Ignorance and argonance to admit it, irritates me and I can get nasty if people are not reasonable. Like right now. You have been blatantly worng on almost everything you have said and I and others have ripped apart every thing based on facts. Since you can’t win an agrument based on reason and fact you pick points outside the scope (like revenue, grammer) of said discussion.
You were the one who denigrates people when you can’t win an argumet. Your response to Bryan was a prime example.