Sun Microsystems may have dropped a bit of weight by the time Oracle officially acquires the company. According to two people briefed on Sun’s plans, the company has cancelled its Rock chip project, putting an end to one of its biggest revitalization bets. Sun has been working on the Rock project for more than five years, hoping to create a chip with many cores that would trounce competing server chips from IBM. and Intel. The company has talked about Rock in the loftiest of terms and built it up as a game-changing product. In April 2007, Jonathan Schwartz, the chief executive of Sun, bragged about receiving the first test versions of Rock. But the two people familiar with Sun’s plans say Rock has met with an unceremonious end. The people requested anonymity, as they are not authorized to speak with the press about Sun’s plans. Michelle Parkinson, a Sun spokeswoman, said the company had no comment.
Isn’t because Andy Bechtolsheim left the company ?
Wasn’t he the dude who drived the project?
SPARC and Power still have a niche in the mainframe-like Unixes (Sun M series, p590, etc)… really high-end systems. But that’s all.
Nowadays, with x86-64 systems with up to 48 CPUs (8 socket 6-core Xeons) with tons of RAM and SAN storage… It’s really hard to sell propietary unix hardware. x86 servers completely eaten the entry and even the mid-range servers (“thanks” to VMware ESX).
It’s really sad, cause x86 servers are SH*T… I have to deal with stupid hardware failures every day… faulty power supplies, faulty system boards and HBAs… it’s a freakin shame, belive me.
With that line you are contradicting yourself, there are companies that are able to pay for something without those problems, so there is a market, but not mass market
Let me tell you, we see the exact same hardware faults with Sun proprietary hardware. I’m unclear how you can associate problems with faulty hardware with the x86 architecture.
If you’re having that many problems then maybe you should go and find a new vendor – in the 10 years I have been in the IT industry I have never seen that many problems from hardware. I’ve got three computers at home, not a single problem outside that of replacing the DVD drive in my iMac.
I’ve sold many servers loaded with a variety of operating systems and never seen the level of faults you are; it either says of two things; you’re a terrible system administrator or the hardware vendor you purchased if off is based on the ‘lowest bid wins’ rather than selecting it based on the best balance between performance versus price versus quality.
Edited 2009-06-17 03:12 UTC
I’ve worked with hardware from many different Unix vendors in my career. SGI mips, DEC Alpha, Sun SPARC and they’ve all had hardware failures. RAID cards, hard drives, power supplies, motherboards, they all die eventually no matter how much you paid for your hardware.
There is nothing about x86 that requires cheap or bad hardware. Buy a top of the line x86 server from a reputable vendor and be willing to pay the premium and you’ll get basically exactly the same hardware quality as you’d get from an IBM Power or Sun SPARC.
I want to agree with you… but reality is 100% different.
Propietary Unix hardware has higher quality standard because it’s a high-margin market. x86 servers are low-margin products, vendors compete only on price… so they cut cost in every possible way.
Vendors use cheaper components in his x86 line than in their propietary line. For example, even low-range System Ps have MUCH better components than high-end System Xs.
Hey, I’m not saying that x86 architecture is the problem… vendors are the problem!
x86 architecture can be a problem. Usually, loads of cash from Intel and AMD can make up for this.
Also, vendors compete on profits, not price. Its true that price contributes to sales, but it isn’t the only factor.
I’ve run many X86 servers in my day and I’ve never had that many problems daily, not even when running on cheap whitebox hardware.
Either you don’t know what you’re doing or you’re buying from some truly awful vendor.
Rest assured that Sun’s Rock is not the last big chip failure for the industry. Get ready to witness Intel’s Larrabee and AMD’s Fusion projects come crashing down like so many Hindenburgs.
Anybody who thinks that last century’s multithreading CPU and GPU technologies will survive in the age of massive parallelism is delusional, in my opinion. When the pain becomes unbearable (it’s all about money), it will suddenly dawn on everybody in the industry that it is finally time to force the baby boomers (the Turing Machine worhsippers) into retirement so that we can boldly break away from the flawed and failed computing models of the last century.
Sun blew it but it’s not too late. Oracle should let bygones be bygones and immediately fund another big chip project, one designed to truly rock the industry this time around and ruffle as many feathers as possible. That is, if they know what’s good for them. It would be a shame to let all that engineering talent migrate elsewhere.
How to Solve the Parallel Programming Crisis:
http://rebelscience.blogspot.com/2008/07/how-to-solve-parallel-prog…
Edited 2009-06-16 22:55 UTC
What a minute, that’s stepping way over the line. The last century of computing models have been a failure? The unparalleled progression has changed our lives more than anything in the history of mankind. If you stepped back to 1909 you would see a completely different world and a LOT more manual tasks that we all take for granted today.
The context of my comment is the parallel programming crisis. The Turing Computing Model is no help in finding a solution and is, in fact, the cause of the crisis. Sounds like failure to me.
Anyway, this is just my opinion. So don’t let it bother you too much.
You’re looking way too much into this; Sun has two processors, they have the traditional multi-core processor from Fujitsu and their own multi-core monster. I’d say that on purchase of sun, when looking at the continued development of the custom Sun multi-core processor and compared it to what Fujitsu is doing – there is little to justify continuing it.
SPARC will continue to be developed – and you might see Oracle purchase in the future the Fujitsu SPARC assets. Don’t expect pie in the sky bet the whole farm ideas coming forward; they’ll take conservative with future development but when things do improve they’ll be realistic enhancements and not ideas born out of ‘in theory’, ‘on paper’ and ‘in the lab’ that fuelled much of T1/T2 and Rock’s development.
Fujitsu will take care of SPARC. But it is too sad that Alpha and PA-RISC are gone. Not to mention Power Architecture is doing poorly in some areas.
Making the market too x86-ish is too hectic. I hope SPARC will be fine forever.
The project which became Itanium was originally referred to within HP as “Super-Parallel-PA” or “PA-Wide-Word”. It is PA-RISC: The Next Generation – PA-RISC was always intended by HP to end.
I didn’t know this. Thank you.
Edited 2009-06-18 04:19 UTC
And phase out the sparc all together. Expect with the new tanglewood Itaniums a solaris 10 port, and a copy of transitive bundled with the x86/itanium versions to run sparc exe’s on Solaris 10.
Hell if Apple, SGI, and HP could do it, why not SUN?
Down the road, expect IBM to throw in the towel for the POWER cpu… It’s had a good run, but it’s just cheaper to buy CPU’s from a CPU company.
Not everybody has to invent their own wheels.
A lack of competition is already causing problems. The interesting thing is, the lack of competition was imposed by “standards”: the x86 architecture, since most people use an OS that would only run on it. If the “peoples’ OS” had been linux or NetBSD (or really anything open source), we’d also have a lot more viable processors today, and probably better ones on the whole. True, applications would need to be cross platform as well, but its easy enough to build for different architectures under the same OS for most Unices.
Times are changing though, and ARM is slowly making its way in, so perhaps people will still be making decent workstation processors someday other than Intel and AMD. Though the competition is fierce between these two, an analogy comes to mind – think about how much flexibility a two party system gives you in a Republic.
You really have no clue what you are talking about do you? IBM IS a processor company. The Power processor is probably powering the computational tasks of most cars on the road today. It has a gigantic embedded and millitary application market. Not to mention that the Power6 drives probably the most powerful UNIX servers available today.
Yea, it would be a great idea to throw in the towel on that.
Remember that Fujitsu also develops SPARC cpus together with SUN. The new Fujitsu 45nm 8-core SPARC64 codename “Venus” will deliver 128GFlops. It is the world’s fastest CPU.
http://www.xbitlabs.com/news/cpu/display/20090515100227_Fujitsu_Dis…
http://www.theregister.co.uk/2009/05/13/fujitsu_venus_sparc64/
quad core variant 65nm:
“A 64-way SPARC Enterprise M9000 equipped with quad-core 2.52GHz SPARC64 VII processors delivered 2.023 TFLOPS on the Linpack HPC benchmark”.
SUN will offer high end servers with this Venus CPU. Also SUN will offer the lowend CPU called Niagara T2 SPARC and the new T3 Niagara (16 cores, 16 threads per core = 256 threads). The T2 has several performance world records.