Slurm Bridge: Slurm Scheduling Superpowers in Kubernetes - Alan Mutschelknaus & Tim Wickberg

Don't miss out! Join us at our next Flagship Conference: KubeCon + CloudNativeCon events in Amsterdam, The Netherlands (23-26 March, 2026). Connect with our current graduated, incubating, and sandbox projects as the community gathers to further the education and advancement of cloud native computing. Learn more at https://kubecon.io Slurm Bridge: Slurm Scheduling Superpowers in Kubernetes - Alan Mutschelknaus & Tim Wickberg, SchedMD If you are looking to run multi-node AI inference, AI training, or High Performance Computing workloads, scheduling resources efficiently is imperative. Kubernetes was created to manage microservices which requires reliability, but does not natively handle optimal workload placement. Slurm scheduling does account for the minutiae of node resources, such as node hardware topology, and other desired features such as workload planning, fair use, and batch type scheduling. With Slurm’s scheduling combined with fine-grained resource control offered by DRA drivers for CPUs, NICS, and GPUs, Kubernetes will be leveled up to support large-scale, granular resource scheduling. We demonstrate how the recently released Slurm Bridge, one of the projects within SlinkyProject, which uses the Kubernetes Scheduling Framework and the Slurm scheduler to drive multi-node workload placement decisions.