Charm++ on Kubernetes Cloud

on YouTube ↗08m

" summary

Charm++ runtime runs on Kubernetes by reusing the MPI operator with minimal code changes and adding an MPI operator CRD changes that introduce min/max replicas and a job priority field. A controller implements elastic scheduling that shrinks lower‑priority jobs to free CPUs and later expands them. Runtime resizing is demonstrated with a 2D stencil benchmark; rescaling overhead falls below 10% for larger problem sizes. Also supports Charm4py on Kubernetes.

AI-written summary · about

" related talks

• 36mCapacity Scheduling for Elastic Resource Sharing in KubernetesQingcan Wang, Yuan Chen · KubeCon + CloudNativeCon North America 2021
• 33mFrom spreadsheet scheduling to Kubernetes: building an on-premise ML platformMarek Molisch, Petr Chmelař · DevConf.CZ 2025
• 29mScale Your Batch / Big Data / AI Workloads Beyond the K8s...Anish Asthana, Antonin Stefanutti · Cloud Native AI Day Europe 2024
• 30mBeyond Experimental: Spark on KubernetesWeiwei Yang · Kubernetes Batch + HPC Day North America 2022
• 23mPractice of Fine-grained Cgroups Resources Scheduling in KubernetesQingcan Wang, Xianlu Chen · KubeCon + CloudNativeCon North America 2020 Virtual
• 19mKeynote: Building Bridges: Cloud Native and High Performance ComputingRicardo Rocha · KubeCon + CloudNativeCon Europe 2022
• 38mChallenges of and Solutions for Migrating Spark From Legacy Hadoop Clusters To KubernetesNeha Singla, Rasik Pandey · KubeCon + CloudNativeCon Europe 2025
• 32mUnraveling the Chaos: Raising Workload Resilience on K8s with Chaos MeshThiago Finardi · DevOpsDays Florianópolis 2024