Full correlation matrix analysis of fMRI data on Intel® Xeon Phi™ coprocessors

Full correlation matrix analysis (FCMA) is an unbiased approach for exhaustively studying interactions among brain regions in functional magnetic resonance imaging (fMRI) data from human participants. In order to answer neuroscientific questions efficiently, we are developing a closed-loop analysis...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis pp. 1 - 12
Main Authors: Wang, Yida, Anderson, Michael J., Cohen, Jonathan D., Heinecke, Alexander, Li, Kai, Satish, Nadathur, Sundaram, Narayanan, Turk-Browne, Nicholas B., Willke, Theodore L.
Format: Conference Proceeding
Language:English
Published: New York, NY, USA ACM 15.11.2015
Series:ACM Conferences
Subjects:
Computer architecture
Computer systems organization > Dependable and fault-tolerant systems and networks
Computing methodologies > Machine learning > Machine learning approaches > Factorization methods > Canonical correlation analysis
Coprocessors
Correlation
General and reference > Cross-computing tools and techniques > Performance
Hardware
Mathematics of computing > Mathematical analysis > Numerical analysis
Mathematics of computing > Probability and statistics > Statistical paradigms > Regression analysis
Networks > Network performance evaluation
Neuroscience
Real-time systems
Support vector machines
Theory of computation > Design and analysis of algorithms
Intel® Xeon Phi™ coprocessor
fMRI data
ISBN:1450337236, 9781450337236
ISSN:2167-4337
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Full correlation matrix analysis (FCMA) is an unbiased approach for exhaustively studying interactions among brain regions in functional magnetic resonance imaging (fMRI) data from human participants. In order to answer neuroscientific questions efficiently, we are developing a closed-loop analysis system with FCMA on a cluster of nodes with Intel® Xeon Phi™ coprocessors. Here we propose several ideas for data-driven algorithmic modification to improve the performance on the coprocessor. Our experiments with real datasets show that the optimized single-node code runs 5x-16x faster than the baseline implementation using the well-known Intel® MKL and LibSVM libraries, and that the cluster implementation achieves near linear speedup on 5760 cores.
ISBN:1450337236
9781450337236
ISSN:2167-4337
DOI:10.1145/2807591.2807631