On the Overhead of Topology Discovery for Locality-Aware Scheduling in HPC

The increasing complexity of parallel computing platforms requires a deep knowledge of the hardware and of the application needs. Locality a key criteria for performance optimization. It involves software tools to expose information about the hardware topology to high performance runtime libraries. We show that the overhead of gathering such information from the operating system is significant on large computing nodes that run Linux. This overhead also increases more than linearly with the number of processes that perform it simultaneously. We then study the actual needs of the HPC software ecosystem in terms of topology information. We propose some ways to avoid multiple expensive topology discovery and to share topology information between components such as the resource manager or the runtime libraries.