The installations of SGI‘s latest NUMA beast, the x86 based Altix UV, are starting to trickle in. Launched back in November, the Altix UV finally took SGI’s scalable NUMA solution to a cheaper (and hopefully more profitable) x86 base, using Nehalem EX processors instead of Itaniums.

The Institute of Cancer Research have announced that they are the latest to deploy an Altix UV. Although the press release is short on specs of the machine, it looks like large Single System Image (SSI) machines are still in demand, despite the overwhelming dominance of clusters in HPC.

Altix UV will provide the ICR with a massively scalable shared memory system to process its growing data requirements, including hundreds of terabytes of data for biological networks, MRI imaging, mass-spectrometry, phenotyping, genetics and deep-sequencing information across thousands of CPUs.

SGI needed some lower cost NUMA SSI machines – some would argue since the disastrous migration away from MIPS – and the Nehalem base Altix UV has been a long time coming. Hopefully this will mark an upturn in SGI’s fortunes. The full press release can be viewed here.

News comes from SGI about the launch of their new Channel Network Partner program, a global program for ‘a limited set of specialized partners’. The program benefits are fairly standard tools for resellers:

To support the new program, the company announced the SGI Channel Network Portal, a single point of web access to a comprehensive set of value-add sales and marketing tools for program partners, including a quote configurator, promotions, demand generation, training modules and opportunity registration.

What’s more interesting is the hint that SGI will be moving to making custom solutions available only through selected partners (presumably with some special value adding secret sauce):

SGI partners serve customers across vertical segments, including government, energy, manufacturing, Internet and cloud, financial services, digital media, and research and education. The SGI Channel Network Program offers increased support to address these high-value markets and meet the ever-growing customer requirements for HPC and data center solutions. SGI partners have access to the entire portfolio of SGI products, including new channel-oriented solutions for which partners will be the primary route to market.

Emphasis in the last sentence is mine. What new HPC toys can we expect to see from this new route to market?

The full press release from SGI can be found here, and the new SGI Channel Network portal can be found here.

SGI have announced that the Tasmanian Partnership for Advanced Computing (TPAC) at the University of Tasmania’s (UTAS) supercomputing facility have just installed a chunky 64 node Altix ICE compute cluster.

‘Katabatic’ has a total of 512 processors and a terabyte of RAM, and will be used for Antartic and South Pacific climate modelling.

Thirty full-time TPAC users and more than 100 university researchers in the Antarctic Climate and Ecosystems Cooperative Research Centre (ACE CRC), the Australian Integrated Marine Observing System (IMOS), the School of Chemistry, the School of Maths and Physics, and the Menzies Research Institute share access to Katabatic every day.

The Altix ICE cluster also boasts over 70TB of disk space, and over 500TB of mirrored tape storage. More on SGI‘s press release here.

Verari, makers of large scale, energy efficient blade systems, appear to have either gone into administration, or a very involved re-organisation. With key figures leaving the company, including co-founder and CTO Dave Driggers, there’s a lot of uncertainty floating about as to what will happen to customers with support agreements in place.

SGI have been the first to announced that they will happily cover Verari customer’s support deals – from the press release:

In an effort to provide Verari customers confidence in their business continuity and an ability to map their technology futures, SGI is now adding Verari Systems to our multi-vendor service offerings for both Customer Service and Professional Service,” said David Yoffie, senior vice president of services at SGI. “Verari global customers will now benefit from our worldwide service infrastructure and our experience supporting existing Verari customers in Europe. Customers will be able to work with SGI to define custom support solutions that address their specific needs.

Back in 2008 Verari did a deal with SGI, allowing SGI to provide support to Verari customers throughout the EMEA region. Looks like with the current uncertainty around the company, SGI are nicely positioned to step in and take over all support commitments.

As expected, SGI have used the SC09 show to launch their latest single system image NUMA machine – the Altix UV. The specs are impressive – not only because SGI have dropped in Nehalem EX processors (as expected) – but the improvements in core density and NUMAlink bandwidth are also impressive.

As with the previous Origin and Altix machines, the Altix UV is available in two models. The Altix UV 100 is the familiar 3u ‘building block’ system, allowing you to scale up as needed. Altix UV 100 scales to 96 sockets (768 cores) and 6TB of shared memory in two racks for up to a claimed 7.0 Tflops of performance.

Altix UV 1000 is the big daddy, scaling up to 256 sockets (giving 2,048 processor cores) and 16TB of shared memory in four racks, for up to a claimed 18.6 Tflops of performance. Interestingly, the 16TB memory limit is imposed by the Nehalem EX architecture.

The next generation of NUMAlink offers a staggering 15 GB/sec transfer rate. The new hub chip has been designed to offload MPI communication. Instead of the CPUs having to handle the packaging and transmission of MPI messages, the UV hub now takes that load. This clears the CPUs to do pure number crunching, but still enabling the level of fast MPI communications that’s needed in such a large NUMA system.

Speaking of large systems, the UV design allows individual 4 rack systems to be hooked together in an 8×8 torus. The theoretical limit of the UV hub could provision over 32,000 cores. The UV hub also has some FB-DIMMs to cache directory information, which not only speeds up operations but also helps with the scalability of the solution.

The design of the processor board is interesting. SGI have used Intel’s Boxboro chipset to handle I/O, with the UV hub plugged directly into both Nehalem CPUs via the QPI interconnect.

The I/O risers mean that, since it’s a single system image, any processor core can access any I/O device anywhere in the system. With the potential for so much I/O throughput, it would be interesting to see what a large Altix UV system packed with Tesla GPUs could achieve.

The Altix UV is an evolution, rather than an evolution, of the flexible NUMA design that first appeared in the Origin 3000. Despite all the press about clusters, big single system image machines still remain the most efficient for many problems. The problems that needed solutions like the original Origin 2000 – pre- and post-processing tasks, very large data problems, I/O and memory intensive apps – have, if anything, gotten more complex and demanding over time, and SGI still have the technology to solve them.

The Nehalem EX won’t be formally launched by Intel until Q2 2010, so SGI aren’t releasing any performance figures. However SGI have announced four initial customers, who will be taking delivering once the processors start volume shipment.

The customers announced at launch are the University of Tennessee (1024 cores, 4TB memory), the North German Supercomputing Alliance (HLRN) (two systems totalling 4352 cores, 18TB of memory, to plug into their existing ICE installation), CALcul en MIdi-Pyrénées/Computations in Midi-Pyrénées (CALMIP) based at the University of Toulouse in France (128 cores and 1 TB of memory), and the Institute of Low Temperature Science at Hokkaido University in Japan (180 cores, 360 GB of memory).

With the Altix UV no longer requiring customers to recompile for Itanium, SGI now have a real chance to push these machines – not just for HPC, but also in business data centres, where Sun and HP have been very successful selling large machines like the F25k and Superdome.