Computer Studies Bldg. Room 209
Towards Heterogeneous Hardware Acceleration of Patient-Specific Modeling Applications
Patient-specific modeling, as an emerging area in the study of biomedical
sciences, has started to gain increasing attention because it offers us a
potential ability to \emph{virtually} reproduce the subject-specific
pathophysiological conditions of the living organisms, an ability could
potentially transform the current practice of medicine. However, these
patient-specific modeling applications are highly compute intensive and
take unreasonable amounts of time that make their inclusion in the medical
environment unfeasible. Computationally speaking, data-assimilation
modeling applications perform linear algebra computations over large data
sets in the shape of matrices, source of high parallelism. They also
encompass different computation phases that communicate with each other
through data transfers. We will present a heterogeneous (CPU, GPU, and
FPGA) hardware solution to this type of applications that can dramatically
reduce the execution time, to the point where these applications become
practical in a medical setting.
BIO: Professor Sonia Lopez Alarcon was born in Madrid, Spain. She received a
Bachelor of Physic degree in Physic and Master degree in Electronic in
2002 from the University Complutense of Madrid. In her latest college
years she worked at Lucent Technologies, Madrid, and Fundetel at
Polytechnic University of Madrid, were she became familiar with the design
and fabrication process of integrated circuits. In 2003 she started
working toward a PhD degree in Computer Engineering at the University
Complutense of Madrid, working on cache hierarchy in simultaneous
multithreaded architectures. In 2004 she started her cooperative research
with Professor David H. Albonesi, at the University of Rochester and,
later on, at Cornell University. She graduated in 2009, and she joined the
Department of Computer Engineering at the Rochester Institute of
Technology in the fall of 2009. Her current research interest is on cache
optimization, GPU architecture, and heterogeneous hardware solutions.
Refreshments will be provided at 10:30AM