November 11, 2008
Yesterday, the DOE announced that the Cray XT 'Jaguar' supercomputer at Oak Ridge has been upgraded to 1.64 peak petaflops. That's quite a jump from the 260 peak teraflops* mark it achieved this past summer, and a 60-fold increase from Jaguar's original 26 teraflops in 2006.
The Jaguar upgrade was achieved by tacking on 200 Cray XT5 cabinets onto the existing 84 XT4 cabinets. The new gear uses the latest 2.3GHz quad-core Opterons and Cray's new ECOflex liquid cooling system. To keep the compute nodes fed with data, the system will have a whopping 362 terabytes of memory and a 10-petabyte file system.
Unlike the classified work destined for the Roadrunner system at Los Alamos National Laboratory, the world's first petaflop supercomputer, Jaguar is intended for open science use by the DOE and the broader scientific community. Under the department's Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program, outside researchers and even commercial users are going to get access to their first petaflop super in 2009.
Scientists have already used the upgraded Jaguar to run a superconductivity calculation that achieved a sustained performance of more than 1.3 petaflops. The system will continue to be used for big science in areas such as climate modeling, astrophysics and fusion energy, but now with 5 times the performance.
The petaflop achievement for ORNL is a bit ahead of schedule, perhaps spurred on by the IBM Roadrunner deployment at Los Alamos just a few months ago. The IBM super achieved 1.375 petaflops back in June, breaking the petaflop barrier and grabbing the number one position on the TOP500 list. Jaguar is now in line to capture the top spot on the next list, which will be announced on November 18. With a peak rating of 1.64 petaflops, Jaguar's new Linpack performance should settle in at around 1.3 petaflops, beating June's Roadrunner by 300 teraflops.
Of course, dust rarely settles on the IBM supers. An upgrade to Roadrunner, or even a surprise Blue Gene/P makeover at Argonne or somewhere else, is always possible. We'll know in a week.
Energy consumption at ORNL is going set some records too. Depending upon how efficient that new Cray liquid cooling system is, power use under compute-intensive workloads is liable to be anywhere from 5 to 10 megawatts -- enough to power a small town. One of the advantages of the Cell-based Roadrunner supercomputer is its rather high energy efficiency. For a petaflop of Linpack performance, Roadrunner draws a mere (!) 2.3 megawatts.
The new petaflop machine means more than just prestige to Cray. If ORNL signs off on the upgraded system before the end of the year, Cray will post a profit in 2008, otherwise not. System acceptance would be quite a Christmas present for the supercomputer maker. The last time Cray recorded a profitable year was 2003.
*corrected typo per reader comment.
Posted by Michael Feldman - November 10, 2008 @ 9:00 PM, Pacific Standard Time
Michael Feldman is the editor of HPCwire.
No Recent Blog Comments
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?