AWS Parallel Computing Service (PCS) now supports Slurm v25.05. You can now create AWS PCS clusters running the newer Slurm v25.05.
The release of Slurm v25.05 in PCS provides new Slurm functionalities including enhanced multi-cluster sackd configuration and improved requeue behavior for instance launch failures. With this release, login nodes can now control multiple clusters without requiring sackd reconfiguration or restart. This enables administrators to pre-configure access to multiple clusters for their users. The new requeue behavior enables more resilient job scheduling by automatically retrying failed instance launches during capacity shortages, thus increasing overall cluster reliability.
AWS PCS is a managed service that makes it easier for you to run and scale your high performance computing (HPC) workloads on AWS using Slurm. To learn more about PCS, refer to the service documentation and AWS Region Table.
Categories: marketing:marchitecture/compute
Source: Amazon Web Services
