HIGH PERFORMANCE COMPUTING ENVIRONMENT The High Performance Computing
- Slides: 14
HIGH PERFORMANCE COMPUTING ENVIRONMENT The High Performance Computing environment consists of high-end systems used for executing complex number crunching applications for research it has two such machines and they are, v LIBRA (GPU) Cluster. v VIRGO Super cluster.
IBM's VIRGO SUPERCLUSTER
VIRGO FDR 10 Infiniband Inter Processor Communication Network 298 IBM System x i. Data. Plex dx 360 M 4 Servers SAN Switch - 1 PFS - 1 SAN Switch - 2 PFS - 2 NFS 1 Gbps Ethernet x. CAT Management Network
System Configuration Ø Ø Ø 292 Compute Nodes 2 Master Nodes 4 Storage Nodes Total Compute Power 97 TFlops IBM System x i. Data. Plex dx 360 M 4 Highly Optimized Servers for HPC Ø Populated with 2 X Intel E 5 -2670 8 C 2. 6 GHz Processor Ø A total of 64 GB RAM per node with 8 X 8 GB 1600 MHz DIMM connected in a fully balanced mode. Ø Low powered Mezzanine Adapter for FDR 10 Infiniband based Inter processor communication
Storage Configuration Ø Ø Ø 2 Storage Subsystem for PFS and 1 Storage Subsystem for NAS. Fine Tuned Redundant SAN Switch IBM DS 3500 Series of Storage System Total PFS Capacity 160 TB Total NAS Capacity 50 TB
Virgo Highlights Ø Ø Ø Ranked 4 th in India Ranked 224 th in the world Fastest in Indian Academic Institutions 1 st in India 5 th in the world Rmax of 91. 126 TF Rpeak of 97. 843 TF Efficiency is 932 MFlop/Watt
Software Stack Novell Su. SE Enterprise Linux with SP 2 Extreme Cluster Administration Toolkit (x. CAT) IBM General Parallel File System IBM Load Leveler Intel Cluster Studio IBM Tivoli Storage Manager Vtunes and Intel Cluster Studio XE Commercial Software Namd 2. 9 Gromacs 4. 0. 7, 4. 5. 5 Ansys 11, 12 Abaqus 6. 10, . 6. 11 Fluent 6. 3 Open. Foam 2. 1. 1 Comsol 4. 3 Mathematica 8, 9 Matlab 2012 a, 2012 b Lammps (parallel) Gaussian 9 version A. 02 Allinea
Text File Editor Ø Gvim/VI editor Ø Emacs Ø Gedit Compilers Ø Ø Ø Gnu compilers 4. 3. 4 Intel Compilers 13. 00 Javac 1. 7. 0 Python 2. 7/2. 7. 3 Cmake 2. 8. 10. 1 Perl 5. 10. 0 Scientific Libraries Ø Ø Ø Ø FFTW 3. 3. 2 HDF 5 MKL GNU Scientific Library BLAS LAPACK ++ Parallel computing Frameworks Ø Open. MPI Ø MPIch Ø Intel-MPI Interpreters and Runtime Environments Ø Java 1. 7. 0 Ø Python 2. 7/2. 7. 3 Ø Numpy 1. 6. 1 Open Source Software Ø Open foam Ø Scilab Visualization Software Ø Gnuplot 4. 0 Debugger Ø Gnu gdb Ø Intel idb
Libra Cluster Ø 1 Head Node on HP Proliant DL 380 G 7 servers with Dual Processors, Six-Core Intel Xeon 5670 series processors with 24 GB RAM and 146 gb of SAS Hard disk. Ø 8 nodes based on HP Proliant SL 390 s server of Dual processor, Six-core, Intel Xeon X 5675 Processors with 3 Telsa M 2070 GPU card and 146 gb of SAS Hard disk in each node.
GPU Schematic
GPU Hardware The GPU Cluster consist of One head node and 8 compute nodes and storage which achieves a performance of 6 TFLOPS. Each node entails 3 Tesla GPU's, Intel Xeon with 12 CPU's and 10 TB of shared storage.
- Sand: towards high-performance serverless computing
- Maui high performance computing center
- Laptops for high performance computing
- Mttf
- High performance computing modernization program
- Big purple nyu
- High performance computing linux
- High performance computing modernization program
- High performance spaceflight computing
- Matlab high performance computing
- High performance embedded computing
- High performance embedded computing
- Army high performance computing research center
- Financial environment of business
- Statistical computing environment