c0t0d0s0.org

There is an interesting rumour about Niagara 3 at Register: Sun's Niagara 3 will have 16-cores and 16 threads per core.

Sun Microsystems looks poised to lead the "mainstream" multi-core race for at least a couple more years. By late 2009, the server maker should deliver a third major revision of its Niagara processor which will have 16 cores and an astonishing 16 threads per core, The Register has learned.

As usual i don´t comment the level of truth in such rumours, but just two things: At first, despite to the opinion in the comments ... we don´t kill of Rock. Period. At second: Ashlee doesn´t know half

One of the favorite misconception about our UltraSPARC T1/2 systems is the "Each core has 1.2 GHz, so divide it by four threads, each core has 300 MHz"-meme. This meme was introduces by our beloved market companions like HP or IBM. At first, this sounds quite intutive, but do you really think that such a processor would be the fundation of the fastest single socket SAP system? Denis Sheahan explains in Lesons learned from T1 the misconception behind this FUD:

This line of argument doesn't hold because most commercial code chases pointers and is constantly loading data structures. On average a commercial application stalls every 100 instructions for a variety of reasons such as TLB miss, I cache miss, Level 2 cache miss etc. When a thread stalls it is usually delayed for many cycles, an Icache miss for instance is 23 cycles. So even though a thread is running at 1.2GHz it usually spends 70% of its time stalled. This is why major processor manufacturers create ever deeper out-of-order pipelines in an effort to avoid this stall.
All this stalling is perfect for CMT. The hardware automatically switches out a thread when it stalls and shares its cycles amongst the other 3 threads on the pipeline masking the stall. With this technique we can utilize the pipeline 75% - 80 of the time provided there are enough threads to absorb the stall

Even this explanation leaves out some cases: The cryptographic units work largely parallel to the pipelines. Thus the computational power of one core is even higher than you would assume from the pure frequency. To say it a little bit simplified: For cryptographic workloads you can see this processor as an 64+8 threads and 8+8 cores system, albeit the 8 cryptographic cores are specialized ones ...

Rick Hetherington was in Munich today ... he gave a good presentation into the future of CMT SPARC and what we will see in the near and the middle term. Many people have the opinion, that Sparc will see it´s end soon, especially after the announcement of TI to stop investments in new process technologies. The direct opposite is true, more than ever. I was quite impressed. Unfortunately it´s not up to me to disclose things.

The support for logical domains made it into the 2.6.23 linux kernel. Wow, this was really fast ...

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.

There is a slight but extremely important change in the licensing modell of Oracle. The Multicore FAQ states:

Oracle Database Standard Edition can only be licensed on servers that have a maximum capacity of 4 sockets. When using RAC to cluster this limitation is mirrored in the cluster. It may be licensed on a single cluster of servers supporting up to a maximum of four sockets.

What does this means: Our workhorse server V490 is now usable with Standard Edition. So you can save a vast amount of money. Instead of 3 Licenses (each core is counted as 0.75) Enterprise Edition you need only three licenses Standard.

Or you can build an incredible powerful OLTP RAC-Cluster with 4 T2000 servers. Although they have 32 threads there is only one single socket per Server. 128 parallel threads = 4 UltraSPARC T1 = 4 Sockets = within the limitation Standard Edition. Standard Edition = RAC included.

And now the kicker: Niagara 2 equals 64 threads per socket. Means 256 threads per 4-node cluster. But: Still only 4 sockets. And you surely know it: Still Standard Edition.

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.

While reading Scientias good article about AMD Q4 earnings, i thought a little bit about the situation in the x86 market:

AMD touts the quad core K8L, Intel speaks about the 45 nm Core Duo, both tells the world that they are faster than the other. And both is correct for certain timeframes. Sounds familar. Yeah, it´s leapfrogging. When two architectures are rather similar (UltraSPARC and Power, AMD x64 or Intel x64) then the companies starts to play technical leapfrogging. At one point of time you lead the competition, at an different point of time it´s the competition.

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.

Interesting customer story about the usage of T2000 and X4100 at ebay inc. It´s one of this marketing speak news, nevertheless it points to real world usages of T2000 and databases.

PS: Guiseppe already pointed to it in a comment a few days ago but i didn´t want to hide it in the comments any longer

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

The review at hardware.info starts with a comment which is certainly true: Within all the news about Intel and AMD, you might think that all other companies leaved the processor business. It´s hard to speak agains marketeers who tout the addition of 200 MHz to a cpu as the next revolution of CPU technology. But this only by the way ....
In 8 processor cores on a chip they describe and explain the T1000/T2000 in some depth. A really interesting read. And i like comments like:

No one can deny that Sun has hit a rough patch, but time and time again they don't fail to surprise the market with their well thought-out products. What Apple is for the desktop world, Sun is for networking and server solutions.

Yet another positive review of a Sun system at Network Computing: Analysis: Next-Gen Blade Servers. Although we didn´t win the shootout (you can argue about the metrics) the overall attitude and commentary to the system is very positive:

Perhaps the most impressive aspect of the Sun Blade 8000 Modular System stems from Sun's decision to use a passive midplane concept that passes the PCI Express system bus through the system, rather than using blade-based GbE, FC or InfiniBand I/O fabrics.

and

But, considering that a fully-loaded Sun Blade 8000 Modular System offers the processing equivalent of 10 conventionally racked, four-socket 4U servers, the 8000 is an impressive, brushed aluminum tower of power.

The whole system will be even more impressive, when you take into consideration, that there will be UltraSPARC T1 blades for the system real soon now (I don´t think i tell you an secret here, as the changelog for Opensolaris already revealed it). Imagine a blade server where you can mix and match the processor technologies you need to fullfil your tasks with an uniform Solaris Operating Environment as an umbrella over the system.