Comments Locked

13 Comments

Back to Article

  • bubblyboo - Saturday, August 26, 2017 - link

    So did it sell well?
  • Spoelie - Saturday, August 26, 2017 - link

    Direct quote from the article:

    "The products in question are the Xeon Phi 7220A, 7220P and 7240P coprocessors, which were used for software development by various close partners of Intel, but were never released as commercial products"
  • ImSpartacus - Saturday, August 26, 2017 - link

    It didn't sell at all.

    The whole point of this article is that this is being discontinued earlier than is typical and there's no need for several years of EOL support because it's not being used by anyone.
  • Elstar - Saturday, August 26, 2017 - link

    I don't think they were ever offered for sale. Intel was preparing to do so, and then canceled their plans at the last minute.
  • Samus - Sunday, August 27, 2017 - link

    It's so grossly uncompetitive with AMD and NVidia, they made the right move not putting this on the market. Intel is years behind in parallel performance, it's actually pretty embarrassing when you consider they nix'd their mobile (SoC) development as well.

    A company this large and powerful should have been able to compete by now with NVidia and Qualcomm.
  • Strunf - Monday, August 28, 2017 - link

    Intel is $160 Billion, NVIDIA $97 Billion and Qualcomm $76 Billion... they aren't that small compared to Intel to be easy to be handled, not to mention they are specialized on markets where Intel has little to no expertise.
  • tipoo - Monday, August 28, 2017 - link

    Market cap is one way to look at it, but look at employee count - Intel dwarfs Nvidia, and to a lesser degree Qualcomm.

    https://www.wolframalpha.com/input/?i=Intel+Nvidia...
  • tipoo - Monday, August 28, 2017 - link


    I feel like their post Itanium "x86 or bust" mentality hurts them. An x86 decode and ucode ROM block per core when they're getting to GPU-like core counts that has to hurt die area and efficiency.
  • Elstar - Tuesday, August 29, 2017 - link

    The die cost of the x86 instruction set architecture is fairly trivial these days. Most of the die is take up by cache, I/O logic (PCIe, memory, QPI, OmniPath, on-die ring/mesh network, etc), and vector units.

    For a processor like Xeon Phi, the design assumes that memory bandwidth is the biggest bottleneck for code – not instruction decoding, branch prediction, or single-threadedness. Therefore Xeon Phi gets away with a far simpler pipeline than traditional CPU designs which work hard to make single-threaded code run halfway decent. This tradeoff would be just as true if another instruction set architecture were used.
  • HStewart - Saturday, August 26, 2017 - link

    This is not for the processors themselves - but the PCIe add on cards - but the cpu's are still available. It appears these PCIe cards were only initially use for development purpose and not general distribution

    http://ark.intel.com/products/series/92650/Intel-X...
  • Elstar - Saturday, August 26, 2017 - link

    I don't think they canceled the coprocessor cards because of I/O, density, or peak compute performance, but due to lack of RAM (which is arguably an I/O problem, but not the networking/cluster kind).

    In other words, if your compute heavy problem fits in 16 GB of RAM, then GPUs are still your best bet. But if your workload requires hundreds of gigabyes of RAM, then the [non-coprocessor] Xeon Phi is basically your only option.
  • CiccioB - Saturday, August 26, 2017 - link

    They have probably have hit the PCI latency/bandwidth limitations which nvidia already hit and passed with the use of nvlink.
    So the option was to create a better bus for allowing faster and more powerful acclerators to be added to the motherboards, allowing also other to do so, or exploiting their advantage on being the only one able to make direct I/O towards system RAM through the CPU socket.
    So they scrapped the accelerator option and moved to the socket solution, though this way they cannot count on density anymore. They in fact call themselves off the HPC market, at least that of the biggest servers, just to go towards the market of smaller computing servers.

    nvidia may have marked a strong victory in HPC market through GP100 becoming almost impossible to beat with their monstrous GV100 which promises to be even much better in efficiency and absolute power.

    Question is what will happen to the partners that chose Intel accelerators for their servers: weren't two of the future biggest servers meant to use Intel Phi accelerators for the exascale era?
  • IntelUser2000 - Monday, August 28, 2017 - link

    They likely just replaced it with a bunch of Omni-Path connected *socketed* Xeon Phi chips.

Log in

Don't have an account? Sign up now