Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards an Adaptive Runtime System for Cloud-Native HPC

Aditya Bhosale, Advait Tahilyani, Laxmikant Kale, Sara Kokkila-Schumacher

Abstract

The ongoing convergence of HPC and cloud computing presents a fundamental challenge: HPC applications, designed for static and homogeneous supercomputers, are ill-suited for the dynamic, heterogeneous, and volatile nature of the cloud. Traditional parallel programming models like MPI struggle to leverage key cloud advantages, such as resource elasticity and low-cost spot instances, while also failing to address challenges like performance variability and processor heterogeneity. This paper demonstrates how the asynchronous, message-driven paradigm of the Charm++ parallel runtime system can bridge this gap. We present a set of tools and strategies that enable HPC applications to run efficiently and resiliently on dynamic cloud infrastructure across both CPU and GPU resources. Our work makes two key contributions. First, we demonstrate that rate-aware load balancing in Charm++ improves performance for applications running on heterogeneous CPU and GPU instances on the cloud. We further demonstrate how core Charm++ principles mitigate performance degradation from common cloud challenges like network contention and processor performance variability, which are exacerbated by the tightly coupled, globally synchronized nature of many science and engineering applications. Second, we extend an existing resource management framework to support GPU and CPU spot instances with minimal interruption overhead. Together, these contributions provide a robust framework for adapting HPC applications to achieve efficient, resilient, and cost-effective performance on the cloud.

Towards an Adaptive Runtime System for Cloud-Native HPC

Abstract

The ongoing convergence of HPC and cloud computing presents a fundamental challenge: HPC applications, designed for static and homogeneous supercomputers, are ill-suited for the dynamic, heterogeneous, and volatile nature of the cloud. Traditional parallel programming models like MPI struggle to leverage key cloud advantages, such as resource elasticity and low-cost spot instances, while also failing to address challenges like performance variability and processor heterogeneity. This paper demonstrates how the asynchronous, message-driven paradigm of the Charm++ parallel runtime system can bridge this gap. We present a set of tools and strategies that enable HPC applications to run efficiently and resiliently on dynamic cloud infrastructure across both CPU and GPU resources. Our work makes two key contributions. First, we demonstrate that rate-aware load balancing in Charm++ improves performance for applications running on heterogeneous CPU and GPU instances on the cloud. We further demonstrate how core Charm++ principles mitigate performance degradation from common cloud challenges like network contention and processor performance variability, which are exacerbated by the tightly coupled, globally synchronized nature of many science and engineering applications. Second, we extend an existing resource management framework to support GPU and CPU spot instances with minimal interruption overhead. Together, these contributions provide a robust framework for adapting HPC applications to achieve efficient, resilient, and cost-effective performance on the cloud.
Paper Structure (21 sections, 8 figures, 3 tables)

This paper contains 21 sections, 8 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Charm++
  4. Overdecomposition
  5. Migratable Objects
  6. Resource Elasticity in Charm++
  7. Performance Challenges of HPC on Cloud
  8. Poor Network Performance
  9. Experiments
  10. Processor Heterogeneity and Variability
  11. Experiments
  12. HPC on Spot Instances
  13. GPU Support
  14. Capacity Rebalancing
  15. Experiments
  16. ...and 6 more sections

Figures (8)

  • Figure 1: Communication latency hiding using overdecomposition in Charm++
  • Figure 2: Weak scaling for Jacobi2D with different overdecomposition factors
  • Figure 3: Weak scaling with and without load balancing on heterogeneous CPUs and GPUs
  • Figure 4: CharmCloudManager framework with capacity rebalancing
  • Figure 5: Overhead of spot interruption on CPUs when one instance is interrupted
  • ...and 3 more figures