c0t0d0s0.org

Perhaps you wondered a little bit about the fact, that i didn´t wrote about the reports about the delayed Rock processor. I wrote an article, but this contained some internal information, as but i´m still not an official leak (i should talk with Jonathan about that ), i just deleted it.

The paper for the Rock presentation (held on Monday at the ISSCC) is now online. A good read ...

I assume, we will give more informations about Rock at the ISSCC´08 on Monday. In the advance program of the event you will find two sessions about our upcoming processor:

A Third-Generation 65nm 16-Core 32-Thread Plus 32-Scout-Thread CMT SPARC Processor at 01:30 pm

A 2.3GHz 396mm2 16-core 32-thread processor with deep out-of-order retirement and transactional memory is fabricated in 65nm CMOS. The logical and physical design of this high-end microprocessor enable high throughput, high single-thread performance, mainframe-class reliability, availability and serviceability (RAS), hardware transactional memory, and linear scalability.

Implementation of a Third-Generation 16-Core 32-Thread CMT SPARC
Processor at 02:00 pm

A 65nm third-generation chip-multithreading SPARC® processor operates at 2.3GHz and is targeted for high-performance servers. It is optimized for both single- and multi-threaded applications. Circuit innovations in memory arrays, register files, and floating-point hardware boost performance and circuit robustness with minimum area overhead.

The Register reports (speculates) about the upcoming releases of Niagara II and later Rock:Sun preps 2048-thread monster. It´s hard to keep thinks secret, when you write and open-source the code to support your new gear. I won´t comment the speculations of the new systems. But they will be available on time

As both Rock and Niagara 2 use FB-DIMMS this are really good news: FB-DIMMs have an addtional circuit called Advanced Memory Buffer. The higher consumption of FB-DIMMS can be credited to this AMB. IDT now annouced a 25 percent more efficient circuit. Okay, thats only between 1 and 2 Watt per DIMM, but you have 16, maybe 32 DIMMs in your server, and thus 16 to 32 AMBs. And from a system view, this really counts.

The Unix Guardian is in heavy speculation mode regarding UltraSPARC RK. I won´t comment what´s right and what´s not, but read it yourself

This is Rock!

The first pictures of the silicon in Jonathans Blog

Another interesting stat... for those fond of in-memory databases, Rock will allow us to build coherent systems supporting 256 terabytes in a single software domain (plain vanilla OpenSolaris, no less). That is an awful lot of RAM in a single system (and given the cost of memory nowadays, you'd want to post an armed guard next to that machine).

Competition, live in despair

This article by Ashlee Vance sheds light on something, that kept my mind busy the last month: How to explain the dense pack of IV+ 2.1 GHz, OPL and Rock. To make one thing clear, the conclusions are mine, not the one of Sun.

Good summary at IT Jungle: Sun Details Server Chip Roadmaps at Analyst Summit.
In my opinion the heartening message behind all the processor news is: When you are able to to boot up the processor in such a short time after you got the silicon from your foundry , there isn´t a huge gap between concept and reality, your construction plan and the physical gear. They showed it with Niagara I, they showed it with Niagara II and the show it with Victoria Falls. So i have an incredible confidence, that Rock will appear in time and in specification.

There seems to be some confusion about the thread count of Rock. I could say something about that, but i´m not authorized to be a leak;). You know ... this stuff with "up to the termination of the contract" and the tattoo "scoundrel" on your forehead. Sometimes this is really bad, when your are enthusiastic about something new.
Actually i wanted so say, that Rock gets more and more press coverage. Not bad for a processor that isn´t planed for the next few month to appear in buyable systems.

Another article (this time at eWeek) about the Rock tapeout, the Niagara speedbump and the neptune Networking chip.

We´ve taped out rock at the very beginning of this year. The mediocre article at the Inquirer forgets Niagara II completly, but nevertheless: Rock is reaching the press.

Den Artikel zu Rock, den ich schon in 2008 - the year UltraSPARC finally regained the top verlinkt habe, gibt es jetzt auch in einer deutschen Version bei ZDNet.

Nebenbei angemerkt: Die Antwort is ein definitives "JA" (klassisches "wenn ihr wüsstest, was ich weiss". Wirklich zum Heulen, das man darueber noch nichts sagen darf). Die Frage müsste übrigens anders gestellt werden. Die richtigere Frage muesste eigentlich lauten: "Bringen 16 Kerne den Umschwung für Sun SPARCs in Servern". Denn sonst ist er schon da. Und selbst so ist die Frage nicht richtig gestellt, denn Niagara 1 also known as UltraSPARC T1 ist im Serverbereich ziemlich gut eingeschlagen.

Mika pointed me to two insightful postings regarding Oracle running on Sun CMT Systems. The first is about Oracle on niagara. This explains that the marketing phrases "Niagara is not a database machine" is largly bullshit.
The second speculates about Oracle on Rock. The author thinks in this article, that Rock will obliterate the need for RAC besides of cluster failover acceleration

News about Rock at news.com: Sun puts 16 cores on its 'Rock' chip. Several new facts about features like out-of-order-retirement (a kind of hardware Undo-Button). I think that 2008 will be the year, UltraSPARC will be on the top of the processor world again ... i know, a bold statement ...