Institute for Advanced Computing Applications and Technologies

Resources

Accelerator Cluster

Contact Information

For more information about the accelerator cluster, contact:

Mike Showerman
mshow@ncsa.illinois.edu
217-244-5478

This 32-node cluster combines both GPU (graphics processing units) and FPGA (field-programmable gate array) technology in order to explore the potential of these novel architectures to accelerate scientific computing.

The 32 compute nodes feature:

Two dual-core 2.4 GHz AMD Opterons, 8 GB of memory
One NVIDIA Tesla S1070 containing 4 GT200 GPUs, each with 4 GB of memory
Nallatech H101-PCIX FPGA accelerator, 16 MB SRAM, 512 MB SDRAM
2GB/sec Infiniband connection. (QDR hardware connected to generation 1 PCI-E)
Two compute nodes contain 16GB of memory

The FPGA accelerators were donated by Xilinx. NVIDIA donated the GPUs to the research team led by Wen-mei Hwu, the leader of IACAT's research theme on the simulation of natural and engineered systems. Hwu is also a principal investigator on the Blue Waters project to build the world's first sustained-petascale computing system for open scientific research.

The cluster is used for Electrical and Computer Engineering courses, for training workshops, as a platform to explore the potential for GPUs to accelerate applications to the petascale and beyond, and as a resource for science and engineering researchers.

Staff at the National Center for Supercomputing Applications (NCSA) assist with deployment and support for the accelerator cluster. NCSA explore innovative architectures and techniques to accelerate scientific computing through its Innovative Systems Laboratory.

Additional details on AC are available in this conference paper.

CUDA software downloads

Download the CUDA wrapper library

Download the CUDA memory tester