Abstract | ||
---|---|---|
Optimizing the performance of HPC software requires a high-level understanding of communication patterns as well as their relation to source code structures. We describe an algorithm to detect communication patterns in parallel traces and show how these patterns can guide static code analysis. First, we detect patterns that identify potential bottlenecks in MPI communication traces. Next, we associate the patterns with the corresponding nodes in an abstract syntaxtree using the ROSE compiler framework. Finally we perform static analysis on the annotated control flow and system dependence graphs to guide transformations such as code motion or the automatic introduction of MPI collectives. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1007/978-3-540-69389-5_29 | ICCS (3) |
Keywords | Field | DocType |
hpc software,mpi communication patterns,mpi communication trace,abstract syntaxtree,rose compiler framework,communication pattern,source code structure,static code analysis,mpi collective,guide source code transformations,code motion,static analysis,control flow,source code,software requirements | Unreachable code,Static program analysis,Computer science,Source code,Parallel computing,Code generation,Compiler,Redundant code,KPI-driven code analysis,Distributed computing,Dead code | Conference |
Volume | ISSN | Citations |
5103 | 0302-9743 | 18 |
PageRank | References | Authors |
1.10 | 7 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Robert Preissl | 1 | 70 | 4.97 |
Martin Schulz | 2 | 2227 | 129.64 |
Dieter Kranzlmüller | 3 | 868 | 117.76 |
Bronis R. Supinski | 4 | 31 | 2.24 |
Daniel J. Quinlan | 5 | 652 | 80.13 |