Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Thursday November 20 2014, @08:20PM   Printer-friendly
from the waiting-for-the-supercomputer-on-the-desktop dept.

For years, Linux has dominated supercomputing. The November 2014 Top 500 supercomputer ranking found 485 out of the world's fastest 500 computers running Linux. That's 97 percent for those of you without a calculator at hand.

I became a little curious about what distro supercomputers run, and ran across a distro targeted directly at them: Rocks. The fastest supercomputer in the world today, Tianhe-2, runs a distro called Kylin which interestingly, used to be based on FreeBSD but is now Linux-based.

[Ed's note: See our earlier story: Top-500 Supercomputer Race Goes Cold.]

Related Stories

Top-500 Supercomputer Race Goes Cold 20 comments

Once a seething cauldron of competition, the twice-yearly Top500 listing of the world’s most powerful supercomputers has grown nearly stagnant of late.

In the most recent Top500 compilation, released Monday ( http://top500.org/lists/2014/11/ ), the Chinese National University of Defense Technology’s Tianhe-2 has retained its position as the world’s fastest system, for the fourth time in a row.

Tianhe-2 is no more powerful than when it debuted atop the Top500 in June 2013 ( http://www.computerworld.com/article/2497811/computer-hardware/china-trounces-us-in-top500-supercomputer-race.html ): In a Linpack benchmark, it steamed along offering 33.86 petaflop/s of computing power. A petaflop is one quadrillion floating point operations.

Only one new entrant debuted in the top 10 in this Top500—a system built by Cray for an unnamed U.S. government agency, which brought 3.57 petaflop/s to the table.

http://www.pcworld.com/article/2848812/top500-supercomputer-race-loses-momentum.html

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2) by tynin on Thursday November 20 2014, @09:44PM

    by tynin (2013) on Thursday November 20 2014, @09:44PM (#118259) Journal

    The shop I work at has ~9k CentOS 6 servers (none on 7 yet... because I haven't had time to re-write our inits and everything else to support systemd, but sadly I know one day in the next year we'll get there), and another ~6k servers on Gentoo. We have a couple Ubuntu's in the mix, but nothing of note.

    • (Score: 2) by tibman on Thursday November 20 2014, @10:08PM

      by tibman (134) Subscriber Badge on Thursday November 20 2014, @10:08PM (#118273)

      how do you do your Gentoo emerging/upgrading? distcc? build on one server and deploy binaries to the others? I've seen how you can cache portage but i've never seen a way to mass emerge -vu world.

      --
      SN won't survive on lurkers alone. Write comments.
      • (Score: 2) by tynin on Thursday November 20 2014, @10:23PM

        by tynin (2013) on Thursday November 20 2014, @10:23PM (#118283) Journal

        Sorry, I don't deal with any of the Gentoo nodes. I support their PXE infrastructure (and the other CentOS servers), but the patching and maintaining of the Gentoo side of the house has been left to others.

    • (Score: 1) by maxim on Thursday November 20 2014, @10:11PM

      by maxim (2543) <maximlevitsky@gmail.com> on Thursday November 20 2014, @10:11PM (#118275)

      See, systemd pollutes everything :-(

  • (Score: 2) by novak on Thursday November 20 2014, @10:30PM

    by novak (4683) on Thursday November 20 2014, @10:30PM (#118286) Homepage

    The school I went to uses Rocks on their HPC cluster. At least, on the cluster I got runtime on.

    --
    novak
  • (Score: 1) by dime on Friday November 21 2014, @12:13AM

    by dime (1163) on Friday November 21 2014, @12:13AM (#118330)

    I built (in a small team of 3) 1 system that was officially ranked in the top 100 a few years back. It's probably the smallest of a handful or so hpc clusters I {manage now, have access to}. If I had to guess, I'd say about 10^5 cores.

    The compute on these clusters are all EL. Usually Centos.

    I think one of the main reasons is that all of these systems pxe boot and don't do local storage except for scratch, so various clusterwares or "golden images" are used. In this case, fundamentals are above all. Centos provides that best.

    • (Score: 2) by tynin on Friday November 21 2014, @03:46AM

      by tynin (2013) on Friday November 21 2014, @03:46AM (#118381) Journal

      What are you doing for your remote storage with regards to filesystem, network, and OS?

      Lately I've been involved in clusters on Centos using Infiniband fdr with remote storage using Lustre. Lustre's performance is quite simply amazing as a distributed fs, as is Infiniband for the ICs. The downsides for us has been with Lustre you need someone in your group that can read and understand what the source code is doing to understand some of the failure modes, and even then we often have to quiz one of the devs (though they are helpful). With Infiniband, they have a support model for their finances. Sure the driver is open source, but everything else you pay through the nose for.

  • (Score: 0) by Anonymous Coward on Friday November 21 2014, @04:35PM

    by Anonymous Coward on Friday November 21 2014, @04:35PM (#118524)

    I wonder how many run windows

    • (Score: 0) by Anonymous Coward on Friday November 21 2014, @10:54PM

      by Anonymous Coward on Friday November 21 2014, @10:54PM (#118623)

      Currently, just one. [top500.org] Assuming that is accurate.

    • (Score: 0) by Anonymous Coward on Saturday November 22 2014, @03:06AM

      by Anonymous Coward on Saturday November 22 2014, @03:06AM (#118680)

      0x00 of them (run Windoze). Any MACs in the mix? Nah, those are for single station desktop fanbois.

      I looked up Rocks and it is based on CentOS.