Available Facility

Pegasus:

Head Node

2 X HPE DL 380, 2 X 16 Cores Intel IceLake, 512 GB RAM, Infiniband EDR

Login/Utility Node

LN1-LN2 - 2 X Lenovo SR 650, 2 X 16 Cores Intel SkyLake, 384 GB RAM, Infiniband EDR

LN3-LN4 - 2 X HPE DL 380, 2 X 16 Cores Intel IceLake, 512 GB RAM, Infiniband EDR

Graphics Node

2 X Lenovo SR 650, 2 X 16 Cores Intel SkyLake, 384 GB RAM, Infiniband EDR, NVIDIA Tesla P40 GPU

GPU nodes

4 X HPE DL 380, 2 X 16 Cores Intel SkyLake, 384 GB RAM, Infiniband EDR, NVIDIA Tesla P100 GPU

Compute Nodes

hpc001-hpc060 - Lenovo - 2 X 16 Cores Intel SkyLake, 384 GB RAM, Infiniband EDR

hpc061-hpc080 - HPE - 2 X 16 Cores Intel IceLake, 512 GB RAM, Infiniband EDR

CPU

80 nodes * 32 cores =  2560 cores

Interconnect

Infiniband EDR

Scheduler/Management

Portable Batch System (PBS)

 

Perseus (Decommissioned):

Head Node

Two DL380G7 , 2x3.06Ghz , Hex Core X5675, 96GB RAM, 10K SAS HDD, 2x146GB QDR IB HCA, RPS, DVD ROM

Utility node

Two DL580G7, 4x2.0Ghz, 8 core E7-4820, 128GB, 3x146GB 10K SAS, 4x450GB 10k SAS HDD, QDR IB HCA, 10 G Adapter, RPS, DVD ROM

Compute Nodes hpc001 - hpc094
CPU

94 nodes * 2 CPUs * 8 cores = 1504 cores SL230s compute nodes with 2 CPU (2.6GHz), 8 core E5-2670, 128 GB memory, 1 x 500 GB 7200 rpm SATA disk, QDR IB (Sandy bridge chip)

RAM 94 * 128 GB = 12032 GB = 12 TB
Interconnect InfiniBand
Storage

1 X DDN SFA10000 Hardware RAID array with 750 TB usable storage. 50 TB for LIGO data, 150 TB for NFS, and TB for luster file system and 250 TB for archival totaling 750 TB

Scheduler/Management Loading Sharing Facility (LSF)
Web portal Custom made web portal for submission of jobs and resource monitoring by TCS click here

 

Hyades (Decommissioned): 48 cores GP-GPU Cluster

Head node hyades.ac.iucaa.ernet.in (cn001), 192.168.11.244
Compute Nodes cn001 - cn004
CPU

dual Intel (R) Xeon (R) X5670 @ 2.93 GHz (1 Headnode and 3 Compute nodes)

(Total 8 CPUs * 6 cores = 48 cores)

RAM 24 GB RAM on each node (Total: 24*4=96 GB)
Interconnect 4X QDR Infiniband
Storage 500 GB on each node, 2TB /data1, 20TB /app
Scheduler/Management LSF (Load Sharing Facility)
GPU Card 12 Tesla M 2050 GPU  (4* 3 cards)