August 26, 2013
Researchers at the NASA Advanced Supercomputing Division have a lot of processing power at their disposal, particularly with Pleiades, the 9th fastest supercomputer in the United States and 19th in the world. But when they need to explore data in different ways, they turn to hyperwall-2, a powerful collection of 128 LCD monitors that can show complex physical events in very high resolution.
The hyperwall-2 is, to put it bluntly, every gamer's fantasy display setup. Powered by 1,024 AMD Opteron cores and 128 NVIDIA GeForce GPUs connected to 208 GB of graphics memory, this 28-ft. by 10-ft. wall of displays can render a quarter billion pixels. But this monster's Call of Duty isn't winning first-person shooters. Rather, it's being used to drive visualizations for actual science.
NASA's supercomputing center recently gave journalists a tour of the hyperwall-2, which is located at the organizations facility at Moffett Field in Mountain View, California. The 128 LCD monitors that compose the hyperwall-2 can be used to display individual images in their own "cells," or, more impressively, be used to display a single image across all the screens.
The folks at Design News were there, and took a video of the hyperwall-2 being used to display visualizations of the birth of the Milky Way galaxy, of the flow of energy in the earth's oceans and atmosphere, and a simulation of NASA's new Heavy Lift Vehicle.
The visualizations that run on hyperwall-2 are critical to furthering NASA's scientific understanding of the universe. "If you don't have a high resolution visualization system like this, if you're looking at them on a laptop, if you're looking at terabytes of data, you're going to see a very low resolution version of these kinds of results," said an unidentified NASA employee in the video.
The hyperwall-2, and is connected directly to Pleiades' Lustre filesystem over InfiniBand network connections. "This direct connection means that our visualization experts can read data directly from the Pleiades filesystem for pre- and post-processing on the hyperwall-2, saving many hours of time in copying very large data files for production visualization work," NASA says on its website.
The big wall of pixels features 9 teraflops of its own computing power, to go along with 1.5 petabytes of storage. Custom built for NASA by Colfax International, the system is governed by a Suse Linux operating system, controlled with the PBS job scheduler, and programmed using C and Intel Fortran compilers.
The scientific visualizations available for display on the hyperwall-2 will get even better, thanks to an upgrade to Pleiades that is occurring right now. The SGI-based cluster is getting an additional 3,312 Intel Ivy Bridge nodes, which will significantly boost its processing power.
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?