Lotus Cluster

Introduction

IIT Tirupati has set up a High Performance Computing Cluster (HPC), Lotus, in 2019. The Lotus cluster provides advanced computational resources for IIT Tirupati faculty and students to perform computational research on campus. The system provides high performance, big data, and graphical processing unit (GPU) computing for research groups spanning multiple disciplines.

System Configuration

The Lotus cluster is made up of 2 master nodes, 24 compute nodes and 2 GPU nodes running on RHEL 7 that take and process jobs submitted from the master node using the PBS Pro job scheduler.

Each node has Intel(R) Xeon(R) Gold 6126 CPU @ 2.60GHz with 24 cores, 96GB RAM. Lotus cluster uses Lustre parallel file system for fast IO, and Infiniband network for computational interconnect.

Totally, the cluster has 624 CPU cores, 2 TB RAM, 100TB PFS (Lustre storage), and 2 x Nvidia Tesla V100 16GB GPUs.

Queues

There are 7 Queues - cpu15d, cpu7d, cpu3d, cpu2d, cpu1d, cpu1h, and gpu4d, with appropriate resources grouped together. Details are below:

Software – Compilers, Libraries, Applications

GNU compilers

C = /usr/bin/gcc C++ = /usr/bin/g++ FORTRAN = /usr/bin/gfortran, /usr/bin/f95

Intel MPI (Parallel) compilers

mpiicc = /apps/intel/parallel_studio_xe_2019.5.075/compilers_and_libraries_2019/linux/mpi/intel64/bin/mpiicc mpiicpc = /apps/intel/parallel_studio_xe_2019.5.075/compilers_and_libraries_2019/linux/mpi/intel64/bin/mpiicpc mpiifort = /apps/intel/parallel_studio_xe_2019.5.075/compilers_and_libraries_2019/linux/mpi/intel64/bin/mpiifort

Applications:

Utilization

View Resource Utilization