IBM Cell door to slam open

IBM Cell door to slam open

Summary: Some people still think this is intended just as a gamesengine, but it's not. As the presentation noted above puts it:"Intended to be the next generation standardarchitecture." And it will be too, it has the securityfeatures of RISC, the hardware partitioning IBMloyalists demand, and awesome performance potential.

TOPICS: Hardware
Last week IBM announced that it would open source some Cell libraries. A bit earlier it had provided an extensive overview of the Cell processor architecture.

In brief the standard unit has one primary processor:

  • 64bit, Power Architecture
  • Two way, statically controlled, superscalar pipeline
  • 64bit ALU, 64bit double precision FMAC
  • 128Bit VMX (aka "Altivec")
  • 32KBI +32KB D L1 Cache
  • 512KB L2 Cache
  • Two way hardware multi-threaded
  • Supports logical partitioning
and eight SPEs - or Synergistic Processing Elements.

The SPEs are extremely powerful devices in themselves, implementing a general purpose SIMD core and communicating with each other and the primary processor via hardware memory management, a 256KB DMA buffer, and a shared "cache coherent SMP bus."

This thing, a high end grid reduced to a single chip, is incredible. Promised at 3.2Ghz, it's expected to reach 3.9Ghz almost immediately and delivers an estimated 25.4GB/Second in throughput. With the right code this thing will run better than ten times faster than Xeon, double that on purely floating point applications.

Some people still think this is intended just as a games engine, but it's not. As the presentation noted above puts it: "Intended to be the next generation standard architecture." And it will be too, it has the security features of RISC, the hardware partitioning IBM loyalists demand, and awesome performance potential.

In fact IBM engineers privately demonstrated a blade server made up from Cell processors during the 2005 E3 conference in Los Angeles recently. Here's a bit from the report on the the nikkeibp English language site:

The prototype, called the Cell Processor Based Blade Server, measured approximately 23 x 43 cm. Each board featured two Cell processors, two 512 Mb XDR DRAM chips and two South Bridge LSIs. The Cell processors were demonstrated running at 2.4-2.8 GHz. "We are driving the Cell processors at higher rates in the laboratory," said thengineer. "If operated at 3 GHz, Cell's theoretical performance reaches about 200 GFLOPS, which works out to about 400 GFLOPS per board," he added. IBM plans to release a rack product capable of storing seven of these boards.

Put seven of these in a rack, add Linux, and you have a serious supercomputer. As far as I know there's just one catch: it's easy to port Linux applications to it so they'll amble along, it's apparently quite difficult for people who don't already work in GRID style supercomputing to adapt to the full programming model so they'll run fast.

Topic: Hardware

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Not only Linux but also Apple

    Last week we had the discussion whether or not Apple will switch to Intel. With the Cell architecture, I don't see any reason for a switch. The Cell processor is faster than any Intel architecture to date, with the SIMD units perfectly capable for high bandwith applications like video and maybe functional speech recognition and therefore perfect for the living room. Scaling effects will drive down the price and will pave the way for Apple consumer electronics. Toghether with Sony. Maybe. If they can manage to get a decent (Linux-based?) system running.
    • Yes, Cell may offer a route for Apple but...

      The cell master processor has Apple's Altivec short
      array processor - something that's not necessary in the basic design and therefore suggests that IBM, at least, wants this to happen. (I wonder if they crossed their legal ts on this? i.e. if Apple/freescale gave up the rights in toto or if a decision by Apple not to use it could trigger either a new cell design or a big lawsuit? - I'm glad I'm not a lawyer, but I wish I understood those issues better...)

      On the other hand Apple's graphics would not work on cell without a third party board (or lots of software work), BSD/darwin would work, but most of Apple's media libraries would not (again, I mean without lots of work)and the looseness of the prgramming model doesn't fit Apple very well.

      Equally importantly, I can't see Apple using an architecture that IBM also uses to build commodity PC gear - there'd be no way to control cloning and Apple would slide further into IBM's waiting hands.

      I think Cell will astound the world in terms of performance and cost - and quite possibly pair with Linux to take over from Wintel as everyone's favorite monopoly. However, I don't think it's good news for Apple, and I think/hope Apple will see the sense of staying well clear of it.
      • They still have the BIOS/MicroCode

        Apple can still control that and lock out any chances for clones.

        Apple has already began opening itself up for cloning with the use of more common items such as IDE interfaces, DDR memory, and other common interconnects. Granted the PC industry refrains from using better technologies such as Firewire and Firewire 2, but this doesn't mean that Apple should lower their standards as they have in present times. I remember when Apples had only SCSI drives. I have one at home.

        I hope that IBM opens the platform for Desktop PCs that rival Apple without knocking Apple out of the market. Intel does need to be taken out of play, unfortunately, I seem AMD falling with them.