November 04, 2010
Although they were designed for video games, GPUs (graphics processing units) are bringing hereforto impossible simulations within reach for chemists. The hardware responsible for bringing graphic realism to the gaming world is now making its mark in the world of chemistry. What started out as a chip in gaming components has been repurposed in the world of science for its ability to perform massively parallel calculations.
In 2006, GPU maker NVIDIA introduced Compute Unified Device Architecture (CUDA), an interface that let high-level programs run on the company's GPUs. This new GPU chip structure and programming tool kit changed the playing field, allowing scientists to access GPUs with familiar programming languages such as C, C++ and Fortran. Since then, other graphics chip vendors, such as AMD, have also come out with their own programmable GPUs.
According to an article at Chemical & Engineering News, these easier-to-use GPUs have transformed the computational field in the past three years. Author Lauren Wolf explains that when GPUs are incorporated into supercomputers, more complex calculations become possible. For example, chemists are now using the power of graphics chips to carry out classical molecular dynamics simulations on desktops. GPU computing is also enabling research on large biomolecular systems and even quantum chemical calculations with their complex algorithms are able to benefit from graphics processors.
GPUs have the potential to reduce calculation time for certain problems from years to months, but the time saved depends on the computer systems being compared as well on the size and complexity of the simulation. Generally, the speedup values are typically reported in the 10 to 100 range. According to Klaus Schulten, director of the Theoretical & Computational Biophysics Group at the University of Illinois, Urbana-Champaign, the best aspect of the GPUs' speed advantage is in the trial-and-error process that is such a key part of science. He says that by making simulations faster, "you can be more playful. Before, you could hardly accomplish one long simulation. Now you can do several."
Full story at Chemical & Engineering News
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?