You'll be lucky....
..to get anything useful out of Fujitsu outside of Japan. Products are great, organisation is not.
The relationship between Oracle, acquirer of the Sun Microsystems server business, and Fujitsu, licenser of the Sparc instruction set and maker of its own Sparc iron – you can't even call it clone iron since Sun and then Oracle have been selling the Sparc Enterprise M systems for years – has been particularly inscrutable for the …
"HP has secretly contracted with Intel to keep churning out Itaniums so that HP can maintain the appearance that a dead microprocessor is alive," Oracle wrote. "The whole thing is a remake of Weekend at Bernie's."
Now the real question is when will Oracle stop support for SPARC64 now that they confirmed it is already dead.
back to the beach
IBM has been showing me this chart for awhile. The chart was created by an ex-sun M9000 technical sales person so you can count on it being accurate. Just because a chart says 1-64 sockets does not mean it will ever have 64 sockets or that the 8, 16, 32 sockets will come out within 12 months of the chip. Just ask HP about their 64 socket superdome2 or the year it took to make the 32 available although they claimed two 16 socket boxes next to each other with a single mgmt console somehow should qualify. Fujitsu is out of the SPARC business except for that exotic chip the government paid they to create.
Off to Jamaica for the break
"....IBM has been showing me this chart for awhile. The chart was created by an ex-sun M9000 technical sales person so you can count on it being accurate...."
Yes, that was really accurate. Someone told a guy at IBM that ... and thus, the information that IBM presents about a competitor is accurate and trust worthy. Not that IBM has ever FUDed before, right? It would not surprise me if IBM made all of this up out of the BLUE.
IBM has a famous track record for stating things that are not... accurate. I have posted some of that IBM "trust worthy" and "accurate" information earlier:
For instance, how can one z196 Mainframe with 24 slow z196 cpus, replace 1.500 much faster x86 servers? If you ask yourself this question, then it is clear that the information that IBM provides, is not always "accurate".
Thus, I would not be surprised if this road map presented by IBM, is made out of the blue. Their source? Some unidentified ex-sun guy. Great.
On the other hand, I can identify IBM executives that say that AIX is going to be killed. Now that information comes directly from IBM. Not "I heard someone that knew a guinea that heard that IBM is killing AIX". No, it is not that kind of "information". It comes directly from IBM:
maybe you should try refuting the facts vs. pointing people to an article from 2003 which was a misquote and corrected in the same article
Bowen, the executive in charge of AIX, emphasised that IBM's Unix isn't being replaced by Linux on any product plans. "We've got people now who are building chips for 2007 systems. If we had any belief that AIX was going to fall down and stumble, we wouldn't be doing that," Bowen said. In particular, a major revamp of AIX is due in early 2004. And though Bowen wouldn't provide specifics, he said the AIX development team is somewhere between two and four times as large as the 250 people IBM employs to improve Linux at its Linux Technology Center.
How was the article a misquote? They say AIX will not be killed soon, but it will happen.
At least I am pointing to credible articles (articles from IBM are credible yes?) which proves I do not make anything up, nor lie, nor FUD. There is always substance in my claims.
In your claims, however, there are no substance at all. Your claims are indistinguishable from pure FUD and lies, because you never can back up your claims with any links. "I heard that someone said that a guinea pig said that an ex-sun had done this picture" - yes, very credible indeed. Or pure FUD.
But according to wikipedia, the definition of FUD stems from IBM doing it systematically, all their employees did it. With such a master, I am not surprised that you, Allison Park, also FUD.
This is the type of information that Allison would drool over, with the lack of
critical thinking that is.
"T3 was 6 months late"? T3 was earlier than most analysts thought it would
be. IBM says that P7+ will be out any day now... at least Oracle had an excuse
after having just acquired Sun. T4 was WAY early and they expect T5 to be
much earlier than expected as well.
"Very short chassis lifetime"?-- Just hyperbole. No one knows except Oracle.
IBM may give you one or two generations of slot compatible, but that's about
it. The M series was released in 2007, so according to IBM it will survive 'til at
least 2012 with at least 4 generations and speed bumps.
IBM questions the use of the term M4 and M5? Why not? Fujitsu never used
those terms for their chips... It's like telling IBM to not call their chip P8
because it will use a different chassis than P7. Nonsense. SPARC is SPARC,
no matter who makes it.
IBM complains about TSMC making the chips for Oracle? Huh? I thought
that one of the benefits of IBM making Power was that they make chips for
other platforms as well (games consoles, etc). TSMC has been making the
chips for T-Series and from what I understand SPARC64 as well. So the
problem here is?
IBM then says that the merger of M-Series with T-Series is the first
"new combined, dynamic threading model?" Huh? "Power5 had
this for 10 years"? Please correct me if I am wrong here, but I could
have swore that IBM's "dynamic threading" model was that if you
wanted full performance you had to turn off threading and possibly
even cores (you had to choose single thread perf or throughput).
T4 it is truly dynamic right now, not in the future, changing behavior
depending on the needs of the code. If you have a process that
needs a full core, then it will get it. If you need more threads, then
you will get those. No turning off cores or threads. When will Power
This IBM FUD piece is laughable.
LOL! One has to applaud such a comprehensive de-FUDding! Excellent riposte, Mr Billl.
Of course, one does have to question whether T4 being "truly dynamic" will help it in the market, seeing as the single-threaded performance with today's heavy-single-thread apps is likely to be less than the Pee5 mentioned.....
Yes, the architecture specifications (and even some of the chip designs) are public and downloadable without having to pay a fee.
The use of the SPARC name/trademark, though, is licensed and requires a SPARC International membership as well as passing your implementation though the "compliance test", again as administrated by SPARC International. See "how much does it cost" paragraph on http://www.sparc.org/aboutFAQ.html
So it's kinda open-as-in-OpenJDK.
When you create and talk about the homunculus, you probably need to say that while it quacks like a duck, walks like a duck, looks like a duck and mates with ducks, it's pure chicken DNA.
Now which pocket did I stuff that testing kit into ...
IBM does upgrade their enterprise systems....p570 to p770....and P7 p795 was just a book replace in the p595.....but then again while the next two chips will be socket compatible for the current tukwila that also means the number of QPI links will never get in sync with the systems. 5 QPI's when you need 8 in the glueless 8 socket blade and only 3 of the 5 are used in the superdome2
Erm, changing "the book" is ripping the heart out of the system! All that's left is the frame. Thanks for pointing out that the next two gen of Itanium will be pin-compatible with current Tukzilla kit, which means all you do is add the new CPUs and enjoy the benefits, something that will definately NOT be the case with Pee8. And please stop the babbling about glueless sockets, it's getting very boring. Hp have now lab-tested, approved and are shipping hp-ux on the Superdome2 all the way to 128-cores, as reported here on The Reg (http://www.theregister.co.uk/2011/09/06/superdome_hp_ux_update/), despite you continually assuring us this would never happen. Were you too busy rip'n'replacing all those Pee5 books to notice?
LOL, I think you'll find that hp's service engineers will have a bit more developed process than that. For a start, there will be no need to pull off heatsinks as the CPU will come as a module with the heatsink attached, just like Xeons do now.... Please remember not all vendors have as ropey upgrade practices as It's Being Mended.
But, you have missed an interesting point - who is going to advise all those old Sun shops on upgrades from existing Sun kit? Snoreacle have severely annoyed the channel (for example, one of our old Sun-only resellers is now trying to flog us hp and CISCO instead), and they have not built up the salesforce Larry said they would, so there's currently much more of a chance that it will be hp or IBM resellers (or direct) whispering sweet promises in those ex-Sun customers' ears.
Biting the hand that feeds IT © 1998–2019