Bengaluru: Hewlett Packard Enterprise (HPE) sees a larger adoption of high-performance computing (HPC) in India, said the Satsangi Som State Manager director.
Growth, he said, was driven in part by cloud-based offers.
HPE in India focuses on HPC solutions in defense, weather and climate, academic research, semiconductor industry, automotive industry, and oil & gas.
Ashok Leyland has deployed HPC at the R & D design center for design, Boeing and Airbus activities that have spread it in their design centers in Bengaluru, and TV motorbikes have mobilized HPC in their R & D center in Hosur.
HPC – or supercomputing as before – is the ability to process data and do complex calculations with very high speeds, at least 1 teraflop (operation of one trillion floating-point per second).
HPC uses thousands of computing nodes that work together to complete one task or more, a process called parallel processing.
Traditionally, only big businesses are able to pay HPC.
HPC in the premise is also complicated, and requires a lot of physical space and maintenance of experts.
Cloud has changed dramatically by giving HPC as a service.
A much smaller organization is now able to buy it.
HPE so far is the largest HPC vendor in the world, with a market share of 37%, followed by Dell and Lenovo.
HPE consolidated its position in the segment with the acquisition of SGI (Silicon Graphics International) in 2016 and Cray in 2019.
Both are supercomputer vendors.
In India, said Satsangi, HPE has a 46% share in the HPC market.
He said the company saw an increase in traction for HPC as a service, which offered the Pay-As-You-Go option for offers that bundle software, storage and networking.
This pre-bundle offer comes in small, medium and large choices.
HPE also has a specially built option to run simulation modeling and workload.
HPC as the service is activated with the HPE Greenlake Cloud service.
“This will not only help the company, but also SMEs to bring distinguished applications utilizing AI and ML,” Satsangi said.
Technology such as AI / ML, IoT, and 3D imaging involves a large number of data, and HPC is very important to deal with it.
HPC is also needed to process data in real time during direct streaming of sports events, or analyzing stock trends, tracking developing storms, or testing new products.
“In accident simulations, latency can be proven dangerous if we cannot retrieve data on time.
Whether it’s weather forecasting or molecular modeling, you need a strong system and HPC in the middle,” Satsangi said.
The Ministry of Sciences (Moes) in India has deployed the XC40 Cray SuperComputer in the two divisions – Institute of Tropical Indian Meteorology (IITM) in Pune, and the National Center for Medium Weather Estimates (NCMRWF) in Noida.
What in IITM will be used to conduct research on weather and climate estimates.