AWS ParallelCluster 3.2 liberate is now usually to be had. AWS ParallelCluster is a completely supported and maintained open supply cluster control instrument that makes it simple for scientists, researchers, and IT directors to deploy and arrange Top Efficiency Computing (HPC) clusters within the AWS cloud. HPC clusters are collections of tightly coupled compute, garage, and networking assets that allow shoppers to run massive scale medical and engineering workloads.
Important characteristic improvements in AWS ParallelCluster 3.2 come with:
- More than one document programs mounts. Shoppers can now mount as much as 20 Amazon FSx and 20 Amazon EFS current document programs to a unmarried cluster.
- Strengthen for added Amazon FSx document programs. Shoppers can now mount current Amazon FSx for OpenZFS, Amazon FSx for NetApp ONTAP, and Amazon FSx for Lustre Persistent_2 document programs in your cluster. The ParallelCluster cluster configuration document now supplies further sections to configure those document programs.
- Reminiscence-aware process scheduling with Slurm. Shoppers can now leverage Slurm’s memory-based scheduling capability to specify reminiscence necessities on process submission, and regulate placement of compute cases that fit the ones reminiscence necessities.
- Extending flexibility to dynamically replace clusters. Shoppers can now dynamically replace process queue homes equivalent to AMI with no need to prevent and get started their clusters. For instance, via a easy CLI command, you’ll sizzling switch an AMI whilst controlling the biking of compute nodes.
AWS ParallelCluster is to be had at no further price, and also you pay just for the AWS assets had to run your packages. Discover ways to release an HPC cluster the usage of AWS ParallelCluster right here.
For extra element you’ll in finding the whole liberate notes for the most recent model of AWS ParallelCluster right here.