Members of and subscribers to the ICB have access to a high-performance
scientific computing infrastructure. Several computational architectures
are available to ensure that different workloads can be run effectively;
notable assets include several large-memory, multi-CPU systems
(including a 24 processor system with 48 GB of memory) and over 2 TFlops
and nearly 600 processor cores of cluster computing power, much of it
backed by a high-speed, low latency fiber-optic interconnect.
This computational power is backed by a large, 40+ TB pool of managed
storage. Users' home directories and most systems are backed-up daily.
Additionally, local mirrors of commonly used biological datasets (such
as sequence databases and the Protein Data Bank) are maintained and
continuously updated.
The ICB maintains an extensive library of both commercial and
open-source scientific and mathematical software packages. These include
applications for mathematical and statistical modeling, graphics and
data visualization, molecular and systems modeling packages, and many
others. With some caveats due to licensing restrictions or system
compatibility issues, these packages are available throughout the ICB
infrastructure, are ready to run, and are regularly updated. The ICB has
developed a software version maintenance system, so older versions of
software are maintained to provide a stable platform for research and
ensure reproducibility of results.
For software developers, the ICB maintains a full stack of tools for
development of high-performance computational applications as well as
for developing and hosting publicly accessible, web-based applications.
These tools include several flavors of relational and object-oriented
databases, a revision control system, as well as C and FORTRAN compilers
and parallel programming tools. Additionally, a full stack of tools for
web-based application development and hosting is supported, including
redundant web servers and separate application servers for those
instances where significant back-end computational resources are required.
To subscribe to the ICB infrastructure please e-mail Audrey Rivera at ajr2004@med.cornell.edu or call her at (212) 746-6361
| 
