Nvidia’s Tesla GPU’s are known for its super high computational power and have had a hand in some of the fastest supercomputers including Titan at Oak Ridge National Laboratory and Tianhe-1A, in Tianjin, China. Now, its GPU Tesla platform are also powering Indian Institute of Technology (IITD) Delhi’s supercomputer which has featured as the 166th fastest world’s supercomputer in the list of TOP 500.
The supercomputer is the newest entry in the global list from India and is also the 4th fastest supercomputer in India. With this, the TOP 500 list comprises of 70 systems that are Nvidia Tesla-based systems which demonstrates the significance of accelerated systems in shaping the future of high performance computing.
This 860 teraflops centralized compute system is the largest High Performance Computing (HPC) system across all IITs and is available to developers, end-users and researchers. Ease of programming using CUDA and OpenACC coupled with great performance and lesser infrastructure costs made NVIDIA Tesla computing platform a preferred choice. Nvidia experts worked closely for about two years with the IIT Delhi team to ensure the success of this strategic initiative.
IIT Delhi, associate head computer services center, Dr. Subodh Kumar said, “We are happy to have commissioned the largest GPU-centric HPC cluster in the nation and among the top few in the world. The system is under active use by a variety of researchers. We are sure the capability multiplication that the system provides will enable them to reach the next level in their research and help solve challenging problems in Atmospheric Science, Molecular Systems, Biology, Nano-systems, and many other areas.”
“This is another milestone in our growth journey and validates the capabilities of our Tesla platform which is accelerating numerous supercomputing sites worldwide”, said Nvidia, MD, South Asia, Vishal Dhupar. “Ready availability of 370+ certified HPC applications in different domains will help in teaching and research work in the areas of Data Analytics , Deep Learning Computational Physics, Chemistry, Bioinformatics, Computation Fluid Dynamics and Material Science immensely.”
This list that captures the 500 most powerful commercially available computer systems known globally, was unveiled at SC15, the International Conference for High Performance Computing, Networking, Storage and Analysis, held recently at Austin, Texas.