Visualization

Cluster technology generally focuses on new ways to use existing hardware to help solve tasks in parallel. One such development that has advanced high-performance computing has been the advent of super-high-performance graphic processors or GPUs. Many of the video technologies used in the lucrative video gaming industry exploit the availability of low-cost high-performance clusters. This has given raise to the use of GPGPUs within the realm of HPC. GPGPU stands for General Purpose Graphics Processing Unit. GPGPU computing is the use of a GPU (graphics processing unit) to do general purpose calculations ranging from scientific and engineering computing to financial analysis.

The model for GPGPU computing is to use a CPU and GPU together as a heterogeneous computing model. The sequential part of the application runs on the CPU and the computationally-intensive part runs on the GPU. GPGPU application do require software developer has to modify their application to take the compute-intensive kernels and map them to the GPGPU. The rest of the application remains on the CPU. GPGPU computing is enabled by the massively parallel architecture of NVIDIA’s GPUs and AMD's GPUs. Each GPU can consists of 100s and soon thousands of processor cores that operate together to crunch through the data set in the application.

GPGPU's can handle extensive parallel floating point operations, programming to take advantage of GPU technology could provide a cost-effective increase in performance per node. Current GPU systems use high-speed bus technologies like PCI-Express x16 (capable of 32GB/second of bidirectional DMA transfer), and combine this massive memory access with parallel high-accuracy floating point processing. Modern GPGPU systems are designed to perform large volumes of matrix and vector calculations, useful in many scientific and engineering fields.