Using Runtime Energy Optimizations to Improve Energy
Efficiency in High Performance Computing
SessionDoctoral Showcase Session 2
Presenter
Event Type
Doctoral Showcase
Applications
Architectures
Energy
Heterogeneous Systems
OS and Runtime Systems
Runtime Systems
TimeTuesday, November 14th2:14pm -
2:36pm
Location210-212
DescriptionEnergy efficiency in high performance computing (HPC)
will be critical to limit operating costs and carbon
footprints in future supercomputing centers. With both
hardware and software factors affecting energy usage
there exists a need for dynamic power regulation. This
dissertation highlights an adaptive runtime framework
that can allow processors capable of per-core specific
power control to reduce power with little performance
impact by dynamically adapting to workload
characteristics. Monitoring of performance and power
regulation is done transparently within MPI runtime and
no code changes are required in the underlying
application. In the presence of workload imbalance, the
runtime reduces the frequency on cores not on the
critical path thereby reducing power without
deteriorating performance. This is shown to reduce
run-to-run performance variation and improve performance
in certain scenarios. For applications plagued by memory
related issues, new memory metrics are identified that
facilitate lowering power without adversely impacting
performance.
