Library and Technology Services and the HPC Steering Committee have purchased a test-bed server for researchers on campus to gain experience with GPU computing and to learn how this type of hardware can help satisfy their computational needs..

		Return to Newsletter

Heterogeneous Computing Coming to Lehigh

Pendulum swings are a common trend in computer technology, for example centralized computing to distributed and back. Another example is homogeneous vs. heterogeneous processing.

At present the pendulum swings toward heterogeneous computing -- computers that use a variety of computational units. General Purpose Graphics Processing Units (GPGPU) or GPUs are hundreds or even thousands of small computational cores that are great at doing small, mathematical calculations very quickly. GPUs were originally designed to offload the processing required for display operations from the Central Processing Units (CPUs) of desktop computers.

This processing load is particularly significant in the gaming world, where the object is always to squeeze as many pixels (picture elements) as possible onto a screen as quickly as possible as a gamer flies down a race course in a sports car or drives a tank on a battlefield. The High Performance Computing (HPC) community realized that these graphics co-processing cores could be harnessed to do scientific research that can often require millions of arithmetic operations in large matrix manipulations in fields like molecular dynamics (MD - think molecules vibrating) and medical image processing.

Library and Technology Services and the HPC Steering Committee have purchased a test-bed server for researchers on campus to gain experience with GPU computing and to learn how this type of hardware can help satisfy their computational needs. The six core Intel Xeon (X5650 Westmere 2.66 GHz) CPU based server, with 24 GB of DDR3 memory, comes with 2 nVidia “Fermi” Tesla C2050 GPUs.

Each one of those GPUs has 3 GB of GDDR5 memory and supports single and double precision floating point operations (with a Peak Performance rated at 515 GFLOPS double precision floating point and 1.03 TFLOPS single precision floating point performance) across 448 cores. The machine can be expanded in the future to accommodate another CPU, 2 more GPUs, and twice as much RAM. Access to this Linux-based server will be open to users with Enhanced Level 2 HPC accounts .

These users may employ the hardware in several ways: to load programs that are GPU-aware already, such as MD packages like GROMACS or LAMMPS; to access higher level mathematical subroutine libraries for packages that are GPU-aware, like MATLAB from MathWorks and Jacket from Accelereyes; and to program the cores directly using programming language library extensions in C/C++/Fortran with CUDA or OpenCL.

Questions on this hardware and on access may be directed to Brandon Leeds (byleeds@lehigh.edu), tel: 84905.

-- Brandon Leeds

Article posted September, 2010

Return to Newsletter