Representative Paths Analysis
SessionPerformance Evaluation Tools
Event Type
Paper
Applications
Performance
TimeWednesday, November 15th2pm -
2:30pm
Location402-403-404
DescriptionRepresentative paths analysis generalizes and improves
MPI critical path analysis. To improve diagnostic
insight, we sample the distribution of program path
costs and retain k representative paths. We describe
scalable algorithms to collect representative paths and
path profiles. To collect full paths efficiently, we
introduce path pruning that reduces permanent space
requirements from a trace (proportional to ranks) to
path length (the minimum). To make space requirements
independent of ranks and execution time --- a small
constant in practice --- we profile program paths.
Avoiding the limitations of prior path profiling
approaches, we dynamically discover tasks and attribute
costs in high resolution. We evaluate our algorithms on
seven applications scaled up to 7000 MPI ranks. Full
program paths use as little as 0.01% the permanent space
of current methods; profiles require a nearly constant
100--1000 KB. Execution overhead is under 5% when
synchronization intervals are sufficiently large (a few
milliseconds).
Download PDF:
here
