back to article HPC processor project tosses architectural designs on desk of the European Commission

After just six months, the EU-funded project attempting to build home-grown CPUs that would power future European supercomputers has delivered its first architectural designs. The European Processor Initiative (EPI) has sent blueprints for its RISC-based chip to the European Commission. The body said it expects to ship the …

  1. Anonymous Coward
    Anonymous Coward

    Just what we need

    €77bn of willy-waving to redo work already done by the US and the UK, just because we can't be seen to be beholden to 'foreigners'. Trump would be proud.

    1. devTrail

      Re: Just what we need

      Please inform yourself better before posting such comments.

      The €77bn refers to the entire Horizon2020 program which funds thousands of research projects, it manages all the research supported by the EU and this project is just a small bit compared to projects like the brain simulation for AI and the body simulation for medical research.

      BTW while the Britons as usual get the chance to insult the EU they benefit hugely from it. Notwithstanding Brexit the UK is one of the main beneficiaries of the program.

      1. Ian Bush
        Go

        Re: Just what we need

        Just to add some numbers over the period 2014-2016 UK researchers and Innovators received 15.2% of the Horizon 2020 funding, for a total of just under €3.8bn funding. This was the second highest in Europe, just behind Germany (16.7%): https://ec.europa.eu/unitedkingdom/news/horizon-2020-three-years-uk-tops-league-participants_en

        1. Phil O'Sophical Silver badge
          Trollface

          Re: Just what we need

          over the period 2014-2016 UK researchers and Innovators received 15.2% of the Horizon 2020 funding, for a total of just under €3.8bn funding.

          That's nice, €1.3bn/year. Over the same 3-year period we send around £13bn gross (after the rebate) to the EU each year, of course (Germany sends €20bn/year). They're paying us with our own money, and making it seem like we're getting something, which is pretty smart.

          1. Ian Bush

            Re: Just what we need

            Icon noted, so let's just note that while Horizon 2020 is not by any means the only way we get money back from the EU it does constitute almost 20% of our academic research budget.

            1. Lars Silver badge
              Happy

              Re: Just what we need

              "the only way we get money back from the EU ".

              All you have to do is to wreck your economy and join countries in need of help like Poland and others.

              And, voila, your dream will be fullfilled and you can once again live on the work of others.

              Meanwhile I hope my country will remain rich enough to afford to take part in financing the project(s).

  2. Peter Gathercole Silver badge
    Headmaster

    Pedantic terminology alert.

    Um,

    "based on the Reduced Instruction Set Computing (RISC) architecture, the same architecture successfully employed by Arm"

    I'm sure you don't mean this. Arm is an architecture (specifically things like Arm7 etc.) IBM POWER is an architecture. Intel x86 is an architecture.

    RISC is a style of processor design, typically one that trades off a smaller number of different instructions against a faster execution speed. Of course, this definition is waaaaaay to simplistic now.

    The only place that confusion arises is with the RISC-V processor type, where the architecture is RISC-V (but not just RISC).

  3. TDog

    Lots of women there then!

    Well, one in front on my left and the bearded one in the second row. Truly a case of jobs for for the boys!

  4. Anonymous Coward
    Anonymous Coward

    1. Super-collider. 2. Super-computer. 3. Super-state.

  5. Justthefacts

    This type of project is why you shouldn’t give your money to the EU.

    So, the EU want to grow a European CPU manufacturer. I’m not entirely against that, if they could really make it happen.

    They have a choice of unfortunately rather few key players with real experience and muscle. The ones that spring to mind are - ARM, Nordic, Dialog, NXP has European sites, Qualcomm/CSR. Swiss u-blox acquired a lovely little VLIW expansion of ARM with great horsepower per watt, widely sold commercially in GPS chips, which should work a treat for the massive vector workloads relevant to supercomputing.

    And who did they pick to design the European CPU? Ummm.....a company called Kalray!

    I play in this space myself, and no I hadn’t heard of them either, so I Googled. A company with 70 engineers, 60 of whom are PhDs - already, you know where this is going. Zero actual product in the field, basically just two PoC, and that at 28nm, but with Very Big PowerPoint Plans. Requires programming In A Weird Language to get the manycoreness to work (yeah, we’ve all seen these before) and no real tools expertise. A “roadmap” that shows them outperforming the Google TPU by 2021.

    This is pork and incompetence of the highest calibre.

    It just doesn’t matter how much we give the EU, how much they spend on Horizon2020 and whether U.K. companies get their fair share. What matters is whether the EU spends the resources in ways that benefit us or indeed anybody. And they don’t. They spend money on weird shit like this, on pork companies that never succeed in the real world. Every damn time.

    Have a look at what happened to the Human Brain Project. A billion euros with the strategic spend so poorly allocated to weird inappropriate stuff that the neuroscience community itself has been signing petitions begging to stop the funding for their area.

    1. Louis Schreurs Bronze badge

      Re: This type of project is why you shouldn’t give your money to the EU.

      Its not EU, but general political decision making

      1. Justthefacts

        Re: This type of project is why you shouldn’t give your money to the EU.

        Your whatabouttery is incorrect.

        One of the Trump cards of the EU among the neoliberal consensus is that “yeah the EU is wasteful and corrupt, but so is all government”. It isn’t.

        I have had the pleasure of managing some half dozen contracts in the multi-million range in each of the EU and U.K. research and governmental sectors. It is night and day.

        The U.K. officials are technically astute, ask the right questions, have sensible strategic outlooks, and are broadly hands-off. They have usually worked extensively in the commercial industrial sector themselves and see broadly. They can however be commercially naive in writing contracts.

        The EU officials have almost never worked in a truly commercial environment; but usually are Ecole Poly or equivalent, via Megacorp subsidised pork projects, to manage the self-same projects. They make decisions exclusively on pork basis. They are smart politically, meh technically, and dumb strategically.

        What has been *your* personal experience of managing these research projects that you would like to tell the class?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2020