Latest News Space & Technology

PRATYUSH SUPERCOMPUTER : Which put India from the 300s to the 30s in the Top500

India’s supercomputing prowess moved up several notches Monday after it unveiled Pratyush, an array of computers that can deliver a peak power of 6.8 petaflops. One petaflop is a million billion floating point operations per second and is a reflection of the computing capacity of a system.

Pratyush, which was inaugurated by the Union Minister for Earth Sciences Dr.Harsh Vardhan, is located at the Indian Institute of Tropical Meteorology (IITM) in Pune.



Pratyush is the first Petaflop Supercomputer developed by India. Pratyush an array of computers that can deliver a peak power of 6.8 petaflops. Pratyush is the fourth fastest supercomputer in the world dedicated for weather and climate research, and follows machines in Japan, USA and the United Kingdom. It will also move an Indian supercomputer from the 300s to the 30s in the Top500 list, a respected international tracker of the world’s fastest supercomputers.

Pratyush is expected to add significantly to the nation’s capacity in weather forecasting and climate monitoring and takes India to the fourth position – only behind the UK, Japan, and the USA in terms of dedicated high powered computing resources for weather and climate monitoring. Pratyush, which is essentially an array of computers, draws 4 Petaflops of it’s peak power from IITM, Pune, and the remaining 2.8 Petaflops come from the National Centre for Medium-Range Weather Forecast (NCMRWF) at Noida.

Introduction of Pratyush, which cost the taxpayers over Rs. 450 crores, is expected to predict extreme weather events and droughts. Additionally, it will also be able to provide more accurate cyclone forecast as well.

“As a result of added computing power, the number of farmers who receive weather information through text messages is expected to increase from 24 million at present to 45 million by 2019,” said Union Minister Harsh Vardhan to Economic Times.

“MoES has been trying to scale down the weather forecast, which is presently given at district level, to block level. “With the help of Pratyush, the block level forecast will be possible in another one year. The forecast of next monsoon will also get better,” added the minister.


Top 10 positions of the 50th TOP500 in November 2017
Rank Rmax
Name Model Processor Interconnect Vendor Site
country, year
Operating system
1 93.015
Sunway TaihuLight Sunway MPP SW26010 Sunway NRCPC National Supercomputing Center in Wuxi
 China, 2016
Linux (Raise)
2 33.863
Tianhe-2 TH-IVB-FEP Xeon E5–2692, Xeon Phi 31S1P TH Express-2 NUDT National Supercomputing Center in Guangzhou
 China, 2013
Linux (Kylin)
3 19.590
Piz Daint Cray XC50 Xeon E5-2690v3, Tesla P100 Aries Cray Swiss National Supercomputing Centre
  Switzerland, 2016
Linux (CLE)
4 19.136
Gyoukou ZettaScaler-2.2 HPC system Xeon D-1571, PEZY-SC2 Infiniband EDR ExaScaler Japan Agency for Marine-Earth Science and Technology
 Japan, 2017
Linux (CentOS)
5 17.590
Titan Cray XK7 Opteron 6274, Tesla K20X Gemini Cray Oak Ridge National Laboratory
 United States, 2012
Linux (CLE, SLES based)
6 17.173
Sequoia Blue Gene/Q A2 Custom IBM Lawrence Livermore National Laboratory
 United States, 2013
Linux (RHEL and CNK)
7 14.137
Trinity Cray XC40 Xeon E5–2698v3, Xeon Phi Aries Cray Los Alamos National Laboratory
 United States, 2015
Linux (CLE)
8 14.015
Cori Cray XC40 Xeon Phi 7250 Aries Cray National Energy Research Scientific Computing Center
 United States, 2016
Linux (CLE)
9 13.555
Oakforest-PACS Fujitsu Xeon Phi 7250 Intel Omni-Path Fujitsu Kashiwa, Joint Center for Advanced High Performance Computing
 Japan, 2016
10 10.510
K computer Fujitsu SPARC64 VIIIfx Tofu Fujitsu Riken, Advanced Institute for Computational Science (AICS)
 Japan, 2011

Share of Supercomputers 

(161) Views


One Reply to “PRATYUSH SUPERCOMPUTER : Which put India from the 300s to the 30s in the Top500

  1. This system is an american made Cray XC40.

    “The latest supercomputer at IITM,Pune “Pratyush” is a Cray-XC40 LC [Liquid Cooled] System with 3312 nodes running Intel Xeon Broadwell E5-2695 processors with a peak performance of 4,006 TFLOPS and a total system memory of 414TB. The system is composed of 18 Compute cabinets and uses Cray’s Aries NOC with Dragonfly Interconnect network topology. In addition, the system consists of 16 Intel KNL 7210 accelerator nodes with a peak performance of 42.56TFLOPS and a total memory of 1.5TB.”

Leave a Reply

Your email address will not be published. Required fields are marked *