Texas Tech
University
High Performance
Computing Center
August/September
2007 Newsletter
--------------------------------------------------------------------
1.
User Profiles: Bill Hase & Yu
Zhuang
2.
Usage of High Memory System
--------------------------------------------------------------------
1.
User Profiles: Bill Hase & Yu
Zhuang
--------------------------------------------------------------------
Bill Hase, Department of
Chemistry and Biochemistry, is
PI and Yu Zhuang, Department of
Computer Science, is co-PI of a
$2,500,000 NSF Partnerships for
International Research and
Education (PIRE) grant awarded
to Texas Tech University to fund
a project entitled "Simulations
of Electronic Non-Adiabatic
Dynamics for Reactions with
Organic Macromolecules, Liquids,
and Surfaces.
The grant provides $500,000/year
support for five years to
develop algorithms and software
for the simulations and to apply
them to a range of problems
including reactions of energetic
oxygen atoms with polymer
surfaces on spacecraft and
radiation-induced chemical
bond-breaking in biomolecules
such as DNA. Their research
collaborators for this grant are
at Iowa State University, Yale
University, University of
Santiago de Compostela (Spain),
University of Pisa (Italy) and
University of Vienna (Austria).
--------------------------------------------------------------------
2.
Usage of High Memory System on
Hrothgar
--------------------------------------------------------------------
Swapnil Kohale, a graduate
student in the department of
Chemical Engineering working in
Dr. Rajesh Khare’s group, has
been doing molecular dynamics
simulations with systems
containing as many as 60k atoms.
The simulations are typically
run for around 33 million steps
and take about 3-4 weeks to
finish. In the future, the
department plans to work with
systems containing around 240k
atoms. Such huge systems take
more than 13GB of memory. The
largest serial job that can run
on general purpose nodes is less
than 12GB.
The inclusion of High Memory (HiMem)
queue on Hrothgar, which
currently contains a single
machine with 64GB of memory, has
enabled Dr. Khare’s group to run
these simulations using existing
serial code. It has been of
great benefit as it helps in
studying larger systems that are
required to investigate physical
phenomena that take place on
longer time and length scales.
Srirangam Addepalli, Support
Specialist for the High
Performance Computing Center,
has been working with the
queuing system and with problems
related to the cluster. Contact
him if you have a serial problem
with large memory requirements.
