TY - GEN
T1 - ParaLearn
T2 - 24th ACM International Conference on Supercomputing, ICS'10
AU - Bani Asadi, Narges
AU - Fletcher, Christopher W.
AU - Gibeling, Greg
AU - Wawrzynek, John
AU - Wong, Wing H.
AU - Nolan, Garry P.
AU - Glass, Eric N.
AU - Sachs, Karen
AU - Burke, Daniel
AU - Zhou, Zoey
PY - 2010
Y1 - 2010
N2 - ParaLearn is a scalable, parallel FPGA-based system for learning interaction networks using Bayesian statistics. ParaLearn includes problem specific parallel/scalable algorithms, system software and hardware architecture to address this complex problem. Learning interaction networks from data uncovers causal relationships and allows scientists to predict and explain a system's behavior. Interaction networks have applications in many fields, though we will discuss them particularly in the field of personalized medicine where state of the art high-throughput experiments generate extensive data on gene expression, DNA sequencing and protein abundance. In this paper we demonstrate how ParaLearn models Signaling Networks in human T-Cells. We show greater than 2000 fold speedup on a Field Programmable Gate Array when compared to a baseline conventional implementation on a General Purpose Processor (GPP), a 2.38 fold speedup compared to a heavily optimized parallel GPP implementation, and between 2.74 and 6.15 fold power savings over the optimized GPP. Through using current generation FPGA technology and caching optimizations, we further project speedups of up to 8.15 fold, relative to the optimized GPP. Compared to software approaches, ParaLearn is faster, more power efficient, and can support novel learning algorithms that substantially improve the precision and robustness of the results.
AB - ParaLearn is a scalable, parallel FPGA-based system for learning interaction networks using Bayesian statistics. ParaLearn includes problem specific parallel/scalable algorithms, system software and hardware architecture to address this complex problem. Learning interaction networks from data uncovers causal relationships and allows scientists to predict and explain a system's behavior. Interaction networks have applications in many fields, though we will discuss them particularly in the field of personalized medicine where state of the art high-throughput experiments generate extensive data on gene expression, DNA sequencing and protein abundance. In this paper we demonstrate how ParaLearn models Signaling Networks in human T-Cells. We show greater than 2000 fold speedup on a Field Programmable Gate Array when compared to a baseline conventional implementation on a General Purpose Processor (GPP), a 2.38 fold speedup compared to a heavily optimized parallel GPP implementation, and between 2.74 and 6.15 fold power savings over the optimized GPP. Through using current generation FPGA technology and caching optimizations, we further project speedups of up to 8.15 fold, relative to the optimized GPP. Compared to software approaches, ParaLearn is faster, more power efficient, and can support novel learning algorithms that substantially improve the precision and robustness of the results.
KW - Bayesian networks
KW - FPGA
KW - Markov chain Monte Carlo
KW - reconfigurable computing
KW - signal transduction networks
UR - http://www.scopus.com/inward/record.url?scp=77954756015&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77954756015&partnerID=8YFLogxK
U2 - 10.1145/1810085.1810100
DO - 10.1145/1810085.1810100
M3 - Conference contribution
AN - SCOPUS:77954756015
SN - 9781450300186
T3 - Proceedings of the International Conference on Supercomputing
SP - 83
EP - 94
BT - ICS'10 - 2010 International Conference on Supercomputing
Y2 - 2 June 2010 through 4 June 2010
ER -