Language Selection

English French German Italian Portuguese Spanish

Second Blue Gene system rises up the ranks

Filed under
Hardware

The system, called Blue Gene Watson and located at IBM's Thomas Watson Research Center, performed 91.3 trillion calculations per second, or 91.3 teraflops. That means it's second only to IBM's original Blue Gene/L system, the fastest supercomputer in the world, located at Lawrence Livermore National Laboratory and clocked at 135 teraflops.

IBM's Watson announcement will come two weeks before an expected bevy of such results, when researchers plan to release the latest list of the world's 500 fastest machines. The list is unveiled twice yearly at supercomputing conferences.

The higher speed, measured with an algebraic calculation test called Linpack, puts IBM another notch ahead of Silicon Graphics, whose Columbia system at NASA reached 51.9 teraflops last fall.

Big Blue has a healthy lead in the list and expects to stay on top. The Livermore machine is in line for another doubling in size and, roughly, performance.

IBM sells the Blue Gene systems for about $2 million per rack; each rack has 1,024 processors, and Blue Gene Watson has 20 racks. IBM also rents out access to the machines.

Blue Gene Watson has been used for protein simulations that tie into biology and drug development research.

Source.

More in Tux Machines

Android Leftovers

Leftovers: OSS

Ubuntu 16.04 Review: What’s New for Desktop Users

Ubuntu is a tricky distribution. As much as I love it on my home server, my desktop is a different ballgame. In my experience, releases between LTS versions have many new technologies that may or may not survive in the next LTS. There were many technologies or features that Canonical thought were ambitious -- HUD, experimenting with menus, online dash search, Ubuntu Software Center, etc. -- but they were abandoned. So, if I were to use Ubuntu on my desktop, I would still choose LTS. Read more

Workflow and efficiency geek talks Drush and Drupal

I started using Drupal because I needed an open source content management system (CMS) to use in several community projects. One of the projects I was involved with was just getting started and had narrowed its CMS selection down to either Drupal or Joomla. At the time I was using a different framework, but I had considered Drupal in the past and knew that I liked it a lot better than Joomla. I convinced them to go with the new Drupal 6 release and converted all of my other projects for consistency. I started working with Drush because I wanted a unified mechanism to work with local and remote sites. My first major contribution to Drush was site aliases and sql-sync in Drush 3. Read more