80 likes | 94 Views
Prophesy: Performance Analysis & Modeling of Parallel & Distributed Applications. Valerie Taylor Texas A&M University Xingfu Wu, Joseph Paris Northwestern University Rick Stevens, Ivan Judson, Mark Hereld Argonne National Laboratory.
E N D
Prophesy: Performance Analysis & Modeling of Parallel & Distributed Applications Valerie Taylor Texas A&M University Xingfu Wu, Joseph Paris Northwestern University Rick Stevens, Ivan Judson, Mark Hereld Argonne National Laboratory SPONSORS: National Science Foundation, National Aeronautics and Space Administration
Problem Statement • Given: • Performance models and analyses are critical • Requires significant development time • Distributed systems are complex • Some applications require distributed systems • Goal • Efficient execution of distributed applications • Proposed Solution • Automate as much as possible • Community involvement http://prophesy.mcs.anl.gov
Prophesy System PROPHESY GUI Profiling & Instrument. Template Database Model Builder Performance Database Actual Execution Symbolic Predictor Systems Database DATA ANALYSIS DATA COLLECTION DATABASES http://prophesy.mcs.anl.gov
Automated Instrumentation • In-line data collection • Instrument at the level of basic loops • Allow for user-specified instrumentation Profiling & Instrument. Execution Manager T=E * f; INSTRUMENTATION CODE for (I=1; I<N; I++){ V(I) = A(I) * C(I); B(I) = A(2I + 4); } INSTRUMENTATION CODE Actual Execution T=E * f; for (I=1; I<N; I++){ V(I) = A(I) * C(I); B(I) = A(2I + 4); } http://prophesy.mcs.anl.gov
Databases • Hierarchical organization • Organized into 4 areas: • Application • Executable • Run • Performance Statistics Template Database Performance Database Systems Database http://prophesy.mcs.anl.gov
Prophesy Database Application Performance Application Executable Run Modules Inputs Module_Info Function Performance Systems Compilers Functions Model Template Basic Unit Performance Function_Info Resource Connection Control Flow Data Structure Performance Model_Info Library http://prophesy.mcs.anl.gov
Automated Modeling Techniques • Utilize information in the template and system databases • Currently include three techniques • Curve fitting • Parameterization • Composition using coupling values http://prophesy.mcs.anl.gov
Modeling Example http://prophesy.mcs.anl.gov