Project Overview

The Keeneland system will be delivered in two phases

Phase 1:

The first phase provides an experimental platform called the Keeneland Initial Delivery System (KIDS), which was delivered in October 2010. It is composed of an HP SL-390 (Ariston) cluster with Intel Westmere hex-core CPUs, NVIDIA 6GB Fermi GPUs, and a Qlogic QDR InfiniBand interconnect. The system has 120 nodes with 240 CPUs and 360 GPUs. Keeneland work with a select group of users with this new architecture to develop programming tools and libraries in order to ensure that the platform can productively accelerate important scientific and engineering applications. In the spring of 2012 KIDS was upgraded from NVIDIA M2070 to M2090 GPUs for total peak performance of 255 TFLOPS.


Phase 2:

The Keeneland Full Scale System (KFS) was delivered in the summer of 2012 and completed acceptance in October 2012. It is a Linux cluster with an architecture similar to KIDS. Each node has two 8-core Intel Sandy Bridge (Xeon E5) processors, 32 GB of host memory, a Mellanox FDR InfiniBand interconnect, and three NVIDIA M2090 GPU accelerators. The KFS system has 264 nodes for a total of 8448 physical CPU cores and 792 GPUs. This system will be added to the Extreme Science and Engineering Discovery Environment (XSEDE) resource allocation base to allow a broad set of users to take advantage of this novel architecture. Peak performance of KFS is 615 TFLOPS.