Devana is the latest HPC cluster procured by NSCC/SAS and put into operation in Q2 of 2023. It consists of 148 compute nodes, totaling 9,472 compute cores, 38 TB RAM and 32 GPGPU accelerators. The system's theoretical peak performance is about 800 TFLOP/s.
All nodes are interconnected through a fully non-blocking fat-tree InfiniBand HDR network, and are equipped with Intel® Xeon® Ice Lake architecture processors. There is an accelerated module composed of 8 nodes, each augmented with four NVIDIA Volta™ A100 accelerators. Read more in Compute Nodes section.
The cluster runs with CentOS Linux 7 operating system (Red Hat Linux family). A large variety of utilities, scientific applications and programming libraries (see sections Software and Compilation for more details) maintained by HPC team is available to its user community. The user software stack is generated using Easybuild and is accessible via environment modules. Read more in Compilation and Environment sections.
There are two shared filesystems on the cluster nodes available to users. Persistent user and project data mounted on
/home mount point and temporary calculation data reside on
/scratch mount point. Read Storage section for details.
Open source software Slurm is used as the resource and job management system. It provides a detailed computing resources allocation and job execution. For more information see section Job Submission.