Param Vikram-1000
The Param Vikram-1000 is a High Performance Computing (HPC) Cluster setup in June 2023. The Param Vikram-1000 provides theoretical peak performance (Rpeak) 1395.63 TeraFlop/s (TF) and Maximal LINPACK performance (Rmax ) 956.34 TF . This is a centrally established High Performance Computing facility to cater the high-end computational requirements of PRL scientific and technical fraternity.
The Param Vikram-1000 is the 14th fastest supercomputer in India.
The Param Vikram-1000 has 108 computing nodes dispense 7296 CPU cores, 2,76,480 GPU Cores, 74TB of RAM and 1 PB of high performance Lustre parallel filesystem. To perform image analysis and run related scientific tools like Mathematica, a separate Visualization Node with two 32 core Intel Icelake CPUs having clock speed of 2.60 GHz and two Nvidia A40 GPU Cards is also available. The system has been setup using open source tools/libraries like Lustre, OpenHPC, Slurm, Ganglia, Intel oneAPI etc..
The Param Vikram-1000 is the 14th fastest supercomputer in India.
The Param Vikram-1000 has 108 computing nodes dispense 7296 CPU cores, 2,76,480 GPU Cores, 74TB of RAM and 1 PB of high performance Lustre parallel filesystem. To perform image analysis and run related scientific tools like Mathematica, a separate Visualization Node with two 32 core Intel Icelake CPUs having clock speed of 2.60 GHz and two Nvidia A40 GPU Cards is also available. The system has been setup using open source tools/libraries like Lustre, OpenHPC, Slurm, Ganglia, Intel oneAPI etc..