Top Supercomputers- India

January2021

Following is the ranking of the systems in terms of Rmax (Linpack Benchmark performance).

C - CPUs, G-GPUs, ICO - Intel Co-processors. Cores (4th column) that are not qualified with C/G/ICO refer to CPUs.
OEM - Original Equipment Manufacturer, SI - System Integrator.

Rank	Site	System	Core/Processor/Socket/Nodes	Rmax (TFlops)	Rpeak (TFlops)
1	Centre for Development of Advanced Computing(C-DAC),Pune	NVIDIA DGX A100, AMD EPYC 7742 64C 2.25GHz, NVIDIA A100, Mellanox HDR Infiniband .consists of 42 nos of NVIDIA DGX A100 nodes, Dual socket populated with AMD EPYC 7742@2.25GHz 64Core (128 Cores/Node) CPU , Mellanox HDR Infiniband More Info OEM:ATOS under NSM initiative, Bidder:	41664/42/236 compute nodes each with 40cores .	4619	5267.14
2	Indian Institute of Tropical Meteorology(IITM),Pune	Cray XC-40 class system with 3315 CPU-only (Intel Xeon Broadwell E5-2695 v4 CPU ) nodes with Cray Linux environment as OS, and connected by Cray Aries interconnect. Total storage of 10.686 PB using Lustre Parallel file system ( Sonexion ) and 30 PB archive tier.Intel Xeon Broadwell E5-2695 v4 CPU ) nodes with Cray Linux environment as OS, and connected by Cray Aries interconnect. More Info OEM:Cray, Bidder:Cray	119232//Cray XC-40 class system with 3315 CPU-only (Intel Xeon Broadwell E5-2695 v4 CPU ) .	3763.9	4006.19
3	National Centre for Medium Range Weather Forecasting (NCMRWF),Noida	Intel Xeon Broadwell E5-2695 v4 CPU ) nodes with Cray Linux environment as OS, and connected by Cray Aries interconnect More Info OEM:Cray, Bidder:Cray	83592/2/Cray XC-40 class system with 2322 CPU-only (Intel Xeon Broadwell E5-2695 v4 CPU ) .	2570.4	2808.7
4	Indian Institute of Technology (IITK),Kharagpur	The supercomputer PARAM Shakti is based on a heterogeneous and hybrid configuration of Intel Xeon Skylake(6148, 20C, 2.4Ghz) processors, and NVIDIA Tesla V100. The system was designed and implemented by HPC Technologies team, Centre for Development of Advanced Computing (C-DAC) with total peak computing capacity of 1.66 (CPU+GPU) PFLOPS performance. The system uses the Lustre parallel file system (primary, storage 1.5 PiB usable with 50 GB/Sec write throughput. Archival Storage 500TiB based on GPFS)\|based on a heterogeneous and hybrid configuration of Intel Xeon Skylake(6148, 20C, 2.4Ghz) processors, and NVIDIA Tesla V100. The system was designed and implemented by HPC Technologies team, Centre for Development of Advanced Computing (C-DAC) with total peak computing capacity of 1.66 (CPU+GPU) PFLOPS performance. The system uses the Lustre parallel file system (primary, storage 1.5 PiB usable with 50 GB/Sec write throughput. Archival Storage 500TiB based on GPFS) More Info OEM:Atos, Bidder:Atos	17280//432 Nodes total of 17280 .	935	1290.2
5	Sankhya Sutra Labs Ltd,Bangalore	AMD Rome based Cluster256-Node Compute,AMD Rome based Cluster, AMD EPYC 7742 64-Core Processor with 128core,2 Master nodes,2 login nodes and Mellanox HDR 100 interconnect More Info OEM:HPE, Bidder:Cray	32768//256 compute nodes	922.48	1179.65
6	National Atmospheric Research Laboratory(NARL),Tirupati	The Supercomputer PARAM AMBAR is based on a heterogeneous and hybrid configuration of Intel Xeon Cascade Lake processors, NVIDIA Tesla V100 with nvlink, 100Gbps Intel OPA 100% fully non-blocking based with C-DAC software stack. \|based on a heterogeneous and hybrid configuration of Intel Xeon Cascade Lake processors,NVIDIA Tesla V100 with nvlink, Intel OPA 100Gbps 100% non-blocking.The system was designed and implemented by HPC Technologies team, Centre for Development of Advanced Computing (C-DAC) with C-DAC software stack More Info OEM:Tyrone, Bidder:Netweb	18816//196 compute nodes constituting a total of 18816 CPU cores .	919.61	1384.85
7	Supercomputer Education and Research Centre (SERC), Indian Institute of Science (IISc),Bangalore	The Cray XC-40 System installed at SERC, IISc.(1468 Intel Xeon E5-2680 v3 @ 2.5 GHz dual twelve-core processor CPU-only nodes, 48 [Intel Xeon E5-2695v2 @ 2.4 Ghz single twelve-core processor+Intel Xeon Phi 5120D] Xeon-phi nodes, 44 [Intel Xeon E5-2695v2 @ 2.4 Ghz single twelve-core processor+NVIDIA K40 GPUs] GPU nodes) w/ Cray Aries Interconnect. HPL run on only 1296 CPU-only nodes. More Info OEM:Cray, Bidder:Cray	36336C + 2880ICO + 126720G/1 CPU + 1 GPU/CPU-only nodes: Dual Intel Xeon E5-2680 v3 (Hawell) twelve-core processor at 2.5 GHz, 128 GB memory per node with Cray Linux environment as the OS. Xeon Phi nodes: one Intel Xeon E5-2695 v2 (Ivybridge) twelve-core processor CPU at 2.4 GHz, one Intel Xeon Phi 5120D containing 60 xeon phi cores, 64 GB memory per node. GPU nodes: one Intel Xeon E5-2695 v2 (Ivybridge) twelve-core processor CPU at 2.4 GHz, one NVIDIA K40 GPU containing 2880 GPU cores, 64 GB memory per node.	901.51	1244
8	Indian Institute of Technology(IIT),Kanpur	The PARAM SANGANAK is a heterogeneous and hybrid configuration of Intel Xeon Cascade Lake processors, NVIDIA Tesla V100 with NVLink, Mellanox HDR interconnect with C-DAC HPC Software stack having peak capacity of 1.6 Peta Flops. This system is designed and implemented by the HPC Technologies Group, C-DAC at IIT Kanpur under National Super Computing Mission (NSM). is a heterogeneous and hybrid configuration of Intel Xeon Cascade Lake processors, NVIDIA Tesla V100 with NVLink, Mellanox HDR interconnect with C-DAC HPC Software stack having peak capacity of 1.6 Petaflops. This system is designed and implemented by the HPC Technologies Group, C-DAC at IIT Kanpur under National Super Computing Mission (NSM) More Info OEM:ATOS, Bidder:ATOS	13834//292	851.3	1350
9	Indian Institute of Tropical Meteorology,Pune	Aaditya HPC is built on IBM P6 575 nodes totaling 117 numbers including 4 GPFS nodes and 2 Login nodes. Each node is populated with 32 cores of IBM P 6 CPU running at 4.7 G Hz.The total number of cores works to 3744 cores.The Aaditya HPC has Interconnectivity using Infiniband Switches and Ethernet switches for Management purposes. Total of 200 TB of Storage including Online, Nearline and Archival Storgae.GPFS and other Managment SW.IBM/Lenovo System X iDataPle DX360M4, Xeon E5-2670 8C 2.6 GHz, Infiniband FDR More Info OEM:IBM, Bidder:IBM	38016/16/117	719.2	790.7
10	Indian Lattice Gauge Theory Initiative, Tata Institute of Fundamental Research (TIFR),Hyderabad	Cray XC-30 LC with Nvidia K20x GPUs consisting of 476 nodes. The CPUs are based on Intel Xeon Processor E5-2680 v2, and each node has a NVIDIA K20x GPU. The nodes are interconnected by Cray Proprietary - Aries interconnect. The total storage is 1.1 PB based on Lustre Parallel Filesystem.476 Intel Xeon Processor E5-2680 v2, 10Core 2.8 GHz plus 476 NVIDIA K20x GPU More Info OEM:Cray, Bidder:Cray	4760C + 1279488G/1 CPU plus 1 GPU/Intel Xeon Processor E5-2680 v2 ten-core processor at 2.8 GHz with Cray Linux environment, 32 GB RAM.	558.7	730

,