IBM Collaboration Opportunities Allen D Malony malony cs uoregon

IBM Collaboration Opportunities Allen D. Malony malony@cs. uoregon. edu http: //www. cs. uoregon. edu/research/tau Performance Research Laboratory (PRL) Neuroinformatics Center (NIC) Department of Computer and Information Science University of Oregon

Outline r TAU parallel performance system ¦ ¦ ¦ r UO Neuroinformatics Center ¦ ¦ ¦ r r Overview Projects IBM involvement Overview Projects ICONIC Grid Computational science Oregon high-performance computing and data grid IBM Visit 2006 IBM Collaboration Opportunities 2

TAU Performance System r r Tuning and Analysis Utilities (14+ year project effort) Performance system framework for HPC systems ¦ r Integrated, scalable, flexible, and parallel Open source toolkit for performance problem solving ¦ ¦ Instrumentation, measurement, analysis, and visualization All computing platforms, processor architectures, OSs Ø IBM, ¦ HP, Sun, Linux, SGI, Cray, Apple, Windows, … Many programming languages and compilers Ø C, C++, Fortran 90/95, UPC, HPF, Java, Open. MP, Python Ø IBM, Intel, PGI, GNU, Sun, Microsoft, SGI, Cray, HP, … r Partners: LLNL, ANL, Research Center Jülich, LANL IBM Visit 2006 IBM Collaboration Opportunities 3

TAU Projects r TAU performance system ¦ ¦ ¦ r Para. Prof: parallel profile analysis and visualization Perf. DMF: performance data management Perf. Explorer: performance data mining KTAU: Linux kernel performance monitoring Integration with the Eclipse Platform Open Trace Format (OTF) for scalable parallel tracing Performance analysis of large-scale application ¦ ¦ ¦ CCSM / ESMF / WRF climate/earth/weather simulation National laboratory applications DOE Sci. DAC and Do. D PET IBM Visit 2006 IBM Collaboration Opportunities 4

IBM Involvement r TAU runs on all IBM platforms and processors ¦ ¦ ¦ r IBM Petascale Tools Workshop (‘ 05 and ‘ 06) ¦ ¦ r IBM T. J. Watson Research Center (David Klepacki) Student participation (K. Huck, A. Nataraj) IBM HPCS project ¦ r p. Series, x. Series, BG/L (64 K system at LLNL) All national laboratory IBM systems run TAU Porting to Cell processor TAU integration in Eclipse Parallel Tools Platform (PTP) Open Software Development Lab ¦ University of Oregon membership IBM Visit 2006 IBM Collaboration Opportunities 5

Support Acknowledgements Department of Energy (DOE) ¦ Office of Science Ø MICS, Argonne National Lab ¦ ASC/NNSA Ø University of Utah ASC/NNSA Level 1 Ø ASC/NNSA, Lawrence Livermore National Lab r Department of Defense (Do. D) ¦ HPC Modernization Office (HPCMO) ¦ Programming Environment and Training (PET) r NSF Software and Tools for High-End Computing r Research Centre Juelich r Los Alamos National Laboratory r Para. Tools, Inc. r IBM Visit 2006 IBM Collaboration Opportunities 6

Acknowledgements r r r r r Dr. Sameer Shende, Senior Scientist Alan Morris, Senior Software Engineer Wyatt Spear, PRL staff Scott Biersdorff, PRL staff Robert Yelle, PRL staff Kevin Huck, Ph. D. student Aroon Nataraj, Ph. D. student Kai Li, Ph. D. student Li Li, Ph. D. student Suravee Suthikulpanit, M. S. student IBM Visit 2006 IBM Collaboration Opportunities 7

Neuroinformatics Center (NIC) r UO Brain, Biology, and Machine Initiative (BBMI) ¦ ¦ r Computational science applied to human neuroscience ¦ ¦ ¦ r r r Interdisciplinary research in cognitive neuroscience, biology, physics, and computer science Human neuroscience focus Tools to help understand dynamic brain function Tools to help diagnosis brain-related disorders HPC simulation, large-scale data analysis, visualization Integration of neuroimaging methods and technology Interaction with Electrical Geodesics, Inc. Technology transfer to Cerebral Data Systems IBM Visit 2006 IBM Collaboration Opportunities 8

NIC Projects r Advanced statistical signal analysis (PCA, ICA) ¦ Hi. Per. SAT Ø High-performance signal analysis toolkit Ø Parallelize ICA, wavelet, and other algorithms ¦ APECS Ø Automated r Computational human head / brain modeling ¦ ¦ r r protocol for EEG component separation Conductivity modeling Source localization modeling Image segmentation and brain structure extraction GEMINI IBM Visit 2006 IBM Collaboration Opportunities 9

Electrical Geodesics Inc. (EGI) r r EGI Geodesics Sensor Net Dense-array sensor technology ¦ r 256 -channel geodesics sensor net ¦ ¦ r r Ag. Cl plastic electrodes Carbon fiber leads Net Station ¦ r 64/128/256 channels Advanced EEG/ERP data analysis Stereotactic EEG sensor registration Research and medical services ¦ Epilepsy diagnosis, pre-surgical planning IBM Visit 2006 IBM Collaboration Opportunities 10

Dipole Sources in the Cortex r r Scalp EEG is generated in the cortex Interested in dipole location, orientation, and magnitude ¦ ¦ r Need to capture convoluted geometry in 3 D mesh ¦ r Cortical sheet gives possible dipole locations Orientation is normal to cortical surface From segmented MRI/CT Linear superposition IBM Visit 2006 IBM Collaboration Opportunities 11

Building Computational Brain Models r r MRI segmentation of brain tissues Conductivity model ¦ ¦ Measure head tissue conductivity Electrical impedance tomography Ø small currents are injected between electrode pair Ø resulting potential measured at remaining electrodes ¦ r Finite element forward solution Source inverse modeling ¦ ¦ Explicit and implicit methods Bayesian methodology IBM Visit 2006 IBM Collaboration Opportunities 12

UO ICONIC Grid r NSF Major Research Instrumentation (MRI) proposal “Acquisition of the Oregon ICONIC Grid for Integrated COgnitive Neuroscience Informatics and Computation” IBM Visit 2006 IBM Collaboration Opportunities 13

GEMINI r r r “Grid-based Environments and Methods for Integrated Neuroimaging” Dynamic neuroimaging algorithms and visualization Grid-based integration (processing and data sharing) High-end tool integration and environments Neuroinformatics data ontologies IBM Visit 2006 IBM Collaboration Opportunities 14

Leveraging Internet, HPC, and Grid Computing r Telemedicine imaging and neurology ¦ ¦ r r r Distributed EEG and MRI measurement and analysis Neurological medical services Shared brain data repositories Remote and rural imaging capabilities Build on emerging web services and grid technology Leverage HPC compute and data centers Create institutional and industry partnerships ¦ ¦ ¦ Electrical Geodesics, Inc. Cerebral Data Systems (UO partnership with EGI) Looking for other industrial partnerships IBM Visit 2006 IBM Collaboration Opportunities 15