2006-2007 Research Computing Support
Dedicated support staff for research computing assist in Rice's
research initiatives that require advanced computational resources,
such as computer-based modeling, simulation, visualization, massive
digital content, and data analysis and mining. These support
specialists work with individual faculty, centers, and institutes to
provide a robust computing infrastructure to meet campus-wide
research needs. Current services include systems administration,
applications support, user support, data management, and support for
research clusters and systems at Rice.
The goal of the Research Computing Support Group is to educate
researchers on how to effectively run efficient code on secure,
robust, highly available systems so the researcher can focus their
efforts on conducting research rather than on the details of running
computers. Any Rice faculty or staff member has access to shared
large-scale high productivity computing power at no cost on the Cray
AMD Opteron-based computer (called Ada) and on the HP Itanium-based
Rice Terascale Cluster (referred to as the RTC). These terascale
computers were purchased by CITI with two separate NSF MRI awards.
For more information see: http://www.rice.edu/it/research/
2006- 2007 Research Computing Utilization
The number of user accounts increased steadily over the entire
academic year, almost doubling in number, and vastly outdistancing
the goal of a 10% increase.
Utilization also increased steadily with a small dip during the
Christmas holidays, almost achieving the theoretical maximum of 85%.
Ten to fifteen percent of the CPU time on all clusters cannot be
utilized due to scheduling wait time. This occurs naturally as
the system holds CPUs idle as jobs finish in order to aggregate them
for large-CPU requests. The best indicator of the effectiveness of
the shared research computing resources is that at no time during the
entire year was the job queue empty.
Total Number of Users by Month
Resource Utilization
