SGI births smaller baby super

Resurrecting the ghost of Octane


Silicon Graphics, like Sun Microsystems, got its start as a supplier of technical workstations for nerds, and the new SGI - a combination of the old, bankrupt SGI and the niche server maker Rackable Systems - wants to do what Sun hasn't: get back into the workstation game in a serious way.

Today, such a machine is not called a workstation, but rather a personal supercomputer. The idea behind both boxes is fundamentally the same, despite 25 years of history: get a powerful machine in the hands of a single user with a budget they can afford and without requiring the approval of an IT department.

In August, shortly after closing the acquisition of SGI's assets and putting the SGI name on the combined company, SGI rolled out its combined product roadmap and then rejiggered Rackable's CloudRack cookie sheet servers with a CloudRack X2 variant that was shrunken version of the CloudRack machines, which came in half-rack and full-rack variants. The CloudRackX2, says Geoffrey Noer, senior director of product marketing at the new SGI, was aimed at departments that needed a workgroup cluster. The CloudRack X2s, which launched in early August, offered up to 216 X64 cores with some two switches and three power supplies in a 14U rack on wheels. This is not exactly perceived as a workstation, and its price was surely well above the discretionary budget level. If the CloudRack X2 was to bridge the gap between a two-socket, dual video card PC-style workstation and a rack of clustered servers, it was a bridge too far. The real bridge, it turns out, is the Octane III machine announced today by SGI.

The Octane III is being billed as the personal supercomputer, and it comes to market 16 years after SGI's popular MIPS/Irix Indigo workstations rolled out; the original two-socket Octanes and their funky tower cases came out in 1996, and were beefed up with the Octane2 in 2000. The interesting thing about the Octanes was that they had a crossbar switch architecture, called Crossbow, instead of a system bus, a kind of point-to-point interconnect that was more widely commercialized by Advanced Micro Devices as HyperTransport. It is ironic, in that sad IT definition of irony, that SGI had the right ideas so long ago, ideas that Intel is only getting to market with the QuickPath Interconnect architecture of its Nehalem family of chips this year. The Octanes ran both Irix and Linux.

The Octane III chassis is 12.5 inches wide, 27.5 inches high, and 26.1 inches deep. The system has been designed to be quiet enough to use in an office environment, although it is hard to imagine how this can be the case in an 80-core system. But, that's what Noer said it was, so if you get one and it is loud, you know who to send the complaints to.

You can get the Octane III personal supercomputer in three different configurations, and the first is not substantially different from a two-socket workstation you might build yourself or buy from Hewlett-Packard or Dell. The OC3-TY11 is set up as a high-end graphics workstation that can has a motherboard mounted vertically in the chassis that supports two of Intel's quad-core Xeon 5500 "Nehalem EP" processors.

This machine has 18 DDR3 memory slots, for a maximum of 144 GB of main memory. The board has twin Gigabit Ethernet NICs and fast PCI-Express x16 slots for sporting two graphics cards from nVidia (the Quadro FX1800, FX3800, FX4800, and FX5800 are all supported) or up to two Tesla C1060 graphics co-processors. (You can do one of each, of course, but picking two Tesla co-processors restricts you to the crap on-board graphics on the mobo).

For this workstation to be interesting, you really want more than two x16 slots. Anyway, the graphics workstation configuration of the Octane III can have four 3.5-inch SATA disks. Pricing was not available because the new SGI has inherited some bad habits from Rackable about not having list prices for its products.

The personal supercomputer configuration that is probably going to be most appealing to HPC customers is the OC3-10TY12, which puts CloudRack cookie sheet trays with ten two-socket Xeon 5500 servers plus a Gigabit Ethernet or InfiniBand switch (either dual or quad data rate speeds) into the chassis. Because of the power and cooling issues involved with cramming ten two-socket servers into the pedestal chassis, SGI can only use the Xeon L5520 processors, which have a 60 watt power envelope and which run at 2.26 GHz. If you don't fully load the machine, you can use faster and hotter 95 watt Xeon 5500 parts in the boards.

Next page: Mucho Mobo

Broader topics


Other stories you might like

  • Germany to host Europe's first exascale supercomputer
    Jupiter added to HPC solar system

    Germany will be the host of the first publicly known European exascale supercomputer, along with four other EU sites getting smaller but still powerful systems, the European High Performance Computing Joint Undertaking (EuroHPC JU) announced this week.

    Germany will be the home of Jupiter, the "Joint Undertaking Pioneer for Innovative and Transformative Exascale Research." It should be switched on next year in a specially designed building on the campus of the Forschungszentrum Jülich research centre and operated by the Jülich Supercomputing Centre (JSC), alongside the existing Juwels and Jureca supercomputers.

    The four mid-range systems are: Daedalus, hosted by the National Infrastructures for Research and Technology in Greece; Levente at the Governmental Agency for IT Development in Hungary; Caspir at the National University of Ireland Galway in Ireland; and EHPCPL at the Academic Computer Centre CYFRONET in Poland.

    Continue reading
  • AMD nearly doubles Top500 supercomputer hardware share
    Intel loses out as Instinct GPUs power the world’s fastest big-iron system

    Analysis In a sign of how meteoric AMD's resurgence in high performance computing has become, the latest list of the world's 500 fastest publicly known supercomputers shows the chip designer has become a darling among organizations deploying x86-based HPC clusters.

    The most eye-catching bit of AMD news among the supercomputing set is that the announcement of the Frontier supercomputer at the US Department of Energy's Oak Ridge National Laboratory, which displaced Japan's Arm-based Fugaku cluster for the No. 1 spot on the Top500 list of the world's most-powerful publicly known systems.

    Top500 updates its list twice a year and published its most recent update on Monday.

    Continue reading
  • All-AMD US Frontier supercomputer ousts Japan's Fugaku as No. 1 in Top500
    Exascale beast's test system also claims top spot in the Green500

    The land of the rising sun has fallen to the United States’ supercomputing might. Oak Ridge National Laboratory’s (ORNL) newly minted Frontier supercomputer has ousted Japan’s Arm-based Fugaku for the top spot on the Top500 rankings of the world's most-powerful publicly known systems.

    Frontier’s lead over Japan’s A64X-based Fujitsu machine is by no means a narrow one either. The cluster achieved peak performance of 1.1 exaflops according to the Linpack benchmark, which has been the standard by which supercomputers have been ranked since the mid-1990s.

    Frontier marks the first publicly benchmarked exascale computer by quite a margin. The ORNL system is well ahead of Fugaku’s 442 petaflops of performance, which was a strong enough showing to keep Fugaku in the top spot for two years.

    Continue reading

Biting the hand that feeds IT © 1998–2022