Characterizing the Performance of Executing Many-tasks on Summit

TL;DR

This paper evaluates the performance of executing many independent tasks on the Summit supercomputer using RADICAL-Pilot with JSM and PRRTE, highlighting scalability, overheads, and resource utilization improvements.

Contribution

It provides the first comprehensive performance characterization of RADICAL-Pilot with JSM and PRRTE on Summit for large-scale task workloads.

Findings

PRRTE scales better than JSM for over 1000 tasks

PRRTE overheads are negligible

Resource utilization reaches 63% with 16,000 tasks on 404 nodes

Abstract

Many scientific workloads are comprised of many tasks, where each task is an independent simulation or analysis of data. The execution of millions of tasks on heterogeneous HPC platforms requires scalable dynamic resource management and multi-level scheduling. RADICAL-Pilot (RP) -- an implementation of the Pilot abstraction, addresses these challenges and serves as an effective runtime system to execute workloads comprised of many tasks. In this paper, we characterize the performance of executing many tasks using RP when interfaced with JSM and PRRTE on Summit: RP is responsible for resource management and task scheduling on acquired resource; JSM or PRRTE enact the placement of launching of scheduled tasks. Our experiments provide lower bounds on the performance of RP when integrated with JSM and PRRTE. Specifically, for workloads comprised of homogeneous single-core, 15 minutes-long tasks we find that: PRRTE scales better than JSM for > O(1000) tasks; PRRTE overheads are negligible; and PRRTE supports optimizations that lower the impact of overheads and enable resource utilization of 63% when executing O(16K), 1-core tasks over 404 compute nodes.