NNIN/C equipment, logging in

The main NNIN/C computer facilities consist of twin AMD Opteron cluster. Each cluster contains 28 units, each unit with four processors sharing 4 GB or RAM (cluster #1) or 6 GB of RAM (cluster#2). Thus there are 112 processors on each cluster. On one of the clusters, 14 of the units are connected with Infiniband. The remaining 42 units use gigabit ethernet. In addition, there is a SUN large memory suite: 4 units of 4-way Opterons from SUN Microsystems, two with 24 GB memory, two with 32 GB memory for a total of 112 GB RAM.

The NNIN/C cluster is accessed via ssh to nnininn.seas.harvard.edu.

GPU Cluster
Graphical Processing Units (or GPUs) are dedicated graphics devices whose ancestors date to the 1980s and the graphical chips at the heart of Atari Game consoles and others. In order to render computer graphics, GPUs are optimized to perform various linear algebra functions, such as matrix multiplication with a high degree of parallelism.
Recently it has been discovered that the highly parallel functionality of GPUs makes them potentially useful for accelerating various kinds of scientific calculation and simulation. The Harvard node of NNIN/C has purchased a set of Nvidia Tesla S1070 computers with associated host CPUs. NNIN/C joined together with the NSF-funded Cyber Discovery Institute at Harvard University to propose an NVIDIA Center for Excellence in GPU Computing which has since been awarded (in January 2009).
The principal research computer for the Center will consist of a cluster of S1070 GPUs obtained with support from Nvidia Corporation. NNIN/C is providing funds for a portion of that cluster, which will be accessible to the NNIN/C community. The cluster has been ordered from Microway Inc. and is expected to be installed in the mid to late spring, 2009. At that time, NNIN/C users will be given information on hardware access.