The Seven Fastest Supercomputers in U.S. Academia

Print Friendly, PDF & Email

ca1882eacd44027b15d24b57f7dea42cOver at Education Dive, Keith Button has posted a list of the top seven academic supercomputers in the USA.

The updated TOP500 supercomputer rankings will be released at ISC’14 on June 23rd, but until then, these seven university-owned supercomputers in the U.S. are currently ranked within the list’s top 100.

  1. Stampede, Texas Advanced Computing Center/University of Texas
  2. Conte, Purdue University
  3. Kraken​, National Institute for Computational Sciences/University of Tennessee
  4. AMOS, Rensselaer Polytechnic Institute
  5. Big Red II, Indiana University
  6. HPCC, University of Southern California
  7. Palmetto2, Clemson University

Read the Full Story.

Comments

  1. It’s worth mentioning that this is based on the Top 500 list, which Blue Waters (https://bluewaters.ncsa.illinois.edu/) does not participate in.

  2. Sharan Kalwani says

    How come the NCSA machine is not included? In several public events, they mentioned it was the fastest machine on any academic campus….

    • NCSA refuses to submit LINPACK numbers to TOP500, so they are not ranked. At the same time, they do go around quoting their peak flops, which is ridiculous.

  3. Sharan Kalwani says

    (1) IMHO, restricting one to just the top500 list gives only 1 particular aspect or view of the title ‘academic supercomputing’.
    (2) If you check out https://bluewaters.ncsa.illinois.edu/aboutallocations;jsessionid=E24624A69CA817642E3071583D78CBA6, there are several institutes in academia that are accessing and using Blue Waters. Not my particular response is NOT limited to Blue Waters, but more towards perhaps an article or sharing information which should be a little more evenyl balanced.
    (3) Compute Power based on top500 alone is just 1 way. There are several others ways in which one can measure this: # of original publications which originate out of this, or perhaps # of research teams, or the # of compute hours made available? Maybe a system listed on this at #7 actually gives more compute capacity or has the maximum number of academic teams engaged…..
    (4) I would have been more excited to read about what different science they are tackling/accomplishing, as opposed to the machoFLOPS they have at their disposal – somehow we cannot seem to get away from this.

  4. Jim Fonseca says

    Kraken was decommisioned in April/May of 2014