Demonstrating improved application performance using dynamic monitoring and task mapping
Authors:
- 1 September 2014
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 408-415
Abstract
This work demonstrates the integration of monitoring, analysis, and feedback to perform application-to-resource mapping that adapts to both static architecture features and dynamic resource state. In particular, we present a framework for mapping MPI tasks to compute resources based on run-time analysis of system-wide network data, architecture-specific routing algorithms, and application communication patterns. We address several challenges. Within each node, we collect local utilization data. We consolidate that information to form a global view of system performance, accounting for system-wide factors including competing applications. We provide an interface for applications to query the global information. Then we exploit the system information to change the mapping of tasks to nodes so that system bottlenecks are avoided. We demonstrate the benefit of this monitoring and feedback by remapping MPI tasks based on route-length, bandwidth, and credit-stalls metrics for a parallel sparse matrix-vector multiplication kernel. In the best case, remapping based on dynamic network information in a congested environment recovered 48.9% of the time lost to congestion, reducing matrix-vector multiplication time by 7.8%. Our experiments focus on the Cray XE/XK platform, but the integration concepts are generally applicable to any platform for which applicable metrics and route knowledge can be obtained.Keywords
This publication has 14 references indexed in Scilit:
- Exploiting Geometric Partitioning in Task Mapping for Parallel Computers2014
- REDUCING THE BULK IN THE BULK SYNCHRONOUS PARALLEL MODELParallel Processing Letters, 2013
- There goes the neighborhood2013
- Benefits of Topology Aware Mapping for Mesh InterconnectsParallel Processing Letters, 2008
- Topology-aware task mapping for reducing communication contention on large parallel machines2006
- An overview of the Trilinos projectACM Transactions on Mathematical Software, 2005
- Communication-Aware Processor Allocation for Supercomputers2005
- The ganglia distributed monitoring system: design, implementation, and experienceParallel Computing, 2004
- Multilevel mesh partitioning for heterogeneous communication networksFuture Generation Computer Systems, 2001
- Scotch: A software package for static mapping by dual recursive bipartitioning of process and architecture graphs1996