Rank-Aware Resource Scheduling for Tightly-Coupled MPI Workloads on Kubernetes

TL;DR

This paper introduces rank-aware resource scheduling on Kubernetes for MPI workloads, enabling fine-grained CPU provisioning and in-place scaling to improve cluster efficiency and performance in cloud environments.

Contribution

It proposes a novel rank-aware scheduling method with in-place CPU scaling for MPI on Kubernetes, reducing resource waste and improving performance.

Findings

Proportional CPU requests eliminate Linux CFS throttling.

Concentric decomposition with proportional CPU is 19% faster than Scotch baseline.

Proportional allocation reduces CPU provisioning by 82%, freeing scheduling resources.

Abstract

Fully provisioned Message Passing Interface (MPI) parallelism achieves near-optimal wall-clock time for Computational Fluid Dynamics (CFD) solvers. This work addresses a complementary question for shared, cloud-managed clusters: can fine-grained CPU provisioning reduce resource reservation of low-load subdomains, improving cluster packing efficiency without unacceptably degrading performance? We propose rank-aware resource scheduling on Kubernetes, mapping each MPI rank to a pod whose CPU request is proportional to its subdomain cell count. We also demonstrate In-Place Pod Vertical Scaling (Kubernetes v1.35 GA) for mid-simulation CPU adjustment without pod restart. Three findings emerge. First, hard CPU limits via the Linux CFS bandwidth controller cause 78x slowdown through cascading stalls at MPI_Allreduce barriers; requests-only allocation eliminates throttling entirely. Second, on non-burstable c5.xlarge instances, concentric decomposition with equal CPU is 19% faster than the Scotch baseline, while adding proportional CPU yields a further 3% improvement. Third, at 16 MPI ranks on 101K-cell meshes, proportional allocation is 20% faster than equal allocation while reducing sparse-subdomain provisioned CPU by 82%, freeing 6.5 vCPU of scheduling headroom. Experiments are conducted on AWS EC2 c5.xlarge clusters (4-16 ranks) running k3s v1.35. All scripts and data are released as open source.