To carry out scientific computations, clusters and high-end computing servers based on Alpha, Intel (Xeon, Itanium) and Power 5+ platforms are available for scientific computing. These high-end servers are accessible from desktops over RRCATNet. Two clusters and one storage server are operational as part of DAEGrid.

All the centralized computing resources are accessible over RRCATNet on 24x7 basis. Network File System is configured on computing servers. It allows to access files on remote hosts in exactly the same way as accessing local files.

Network Information System is configured, which facilitates users to login to any of the computing servers with same login ID & password. Users are able to work in single home directory for all the servers and they can use resources of all the servers. Virtual Network Computing (a simple protocol for remote access to graphical user interfaces) is implemented, which facilitates graphical output on user's desktops.

• High Performance Computing Cluster - Aryabhatta
  

  Xeon based high performance computing cluster comprising of 32 nodes/ 64 processors/ 128 cores named as 'Aryabhatta', is commissioned for computing applications of RRCAT. The peak computing power of Aryabhatta is 450 Gigaflops. Operating systems of Aryabhatta is 64 bit Red Hat Enterprise Linux AS4. 'Torque' and 'Maui' are used as resource manager and scheduler on this cluster. Each node contains two dual core Xeon processors (@ 3.73 GHz) with 4 MB L2 cache per processor, 8 GB memory and Gigabit network interface. Intel Fortran compiler (ifort), GNU Fortran 77 compiler (g77), GNU Fortran 90/95 compiler (gfortran), Intel C compiler (icc), GNU C & C++ compiler (gcc & g++) are available on this cluster. Parallel processing libraries - MPICH2-1.0.5 for g77, gfortran & ifort; MPICH-1.2.7 for g77, gfortran & ifort; and Parallel Numerical Libraries - Intel Math Kernel Library, LAPACK, CBLAS, f77BLAS, ATLAS are made available for parallel environment on Aryabhatta cluster. Parallel application WIEN97 [computation of electronic structure of solids within density functional using Linearized Augmented Plane Wave (LAPW) method] and CPMD (Car-Parrinello Molecular Dynamics - Electronic Structure and Molecular Dynamics Program) are successfully ported and running on 'Aryabhatta'.
  

  
  
  32 nodes (64 Processors) High Performance Computing Cluster - Aryabhatta

  
  
• High Throughput Cluster - Nalanda
  

  32 node cluster with CONDOR clustering software is being used for parallel and sequential jobs and it is named as 'Nalanda'. Each node has 2.8 GHz Pentium IV processor with 2 GB RAM. Intel Fortran, lf95, Absoft Fortran 90/95 compilers and parallel library MPICH for each of these compilers, are configured on 'Nalanda'. Math kernel library, SCALPACK, BLAS, CBLAS libraries are also configured on this cluster. Many application programs like ADF (Amsterdam Density Functional - a Fortran program for calculations on atoms and molecules), WIEN97 & WIEN2K [computation of electronic structure of solids within density functional using Linearized Augmented Plane Wave (LAPW) method], CPMD (Car-Parrinello Molecular Dynamics - Electronic Structure and Molecular Dynamics Program) are successfully ported and running on 'Nalanda'. Transport software package (a computer program used to design charged particle beam transport systems) of version 1_5a for linux was configured on 32-node cluster with check-pointing facility.
  

  
  
  32 nodes High Throughput Cluster - Nalanda

• High Performance Computing Clusters as part of DAEGrid
  

  DAEGrid is operational between BARC, RRCAT, IGCAR & VECC to share computing resources Two clusters and one storage server are made operational as part of DAEGrid. The operating system is Scientific Linux version 4.4 on both the clusters. Intel Fortran V 9.0, gcc, gfortran compilers, Math Kernal library v 8.0, parallel processing library MPICH-1.2.7 are installed on both the clusters. Genius based grid-portal on Scientific Linux has been installed on these clusters. Users from RRCAT can submit their parallel and sequential applications on the resources available under DAEGrid, using grid-portal made operational in User Hall of Computer Centre.
  
  Daksha:
  

  8 node cluster named "Daksha" is operational as part of DAEGrid. Each node contains Intel dual core Xeon processor @ 3.60GHz with 2 MB cache, 4 GB memory and Gigabit network interface.

  
  
  8 nodes HPC cluster - Daksha

  Ramanujam:
  

  16 nodes (32 processors) high performance computing cluster "Ramanujam" has been installed and added to the DAEGrid. Each node comprises of two dual core Xeon processors @3.73GHz, 4 MB L2 cache per processor, 4 GB memory and Gigabit network interface. The total no of processors on this cluster is 32 and each processor is dual core with hyper threading. All the nodes are HP DL 380 G5 series servers.

  
  
  16 nodes (32 processors) HPC cluster - Ramanujam

• Storage Server for DAEGrid
  

  Storage Server of 1.6 TB capacity has been commissioned as part of DAEGrid and it can be accessed by DAEGrid portal and clusters. LCG Disk Pool Manager (DPM head node and DPM disk server) has been configured on this 64-bit Scientific Linux based Storage Server. The server is functioning with Catalogue services of DAEGrid. RRCAT users as well as users on DAEGrid can store huge files on this server, which is essentially required for computing applications.
  

  
  
  Typical listing of DAEGrid resources

• Virtual cluster based solution for OPERA Software
  

  One virtual cluster based setup has been commissioned for OPERA software, which is well known magnet designing and simulation software. OPERA version 11.02 has been procured by Indus Operations & Accelerator Physics Design Division, for which virtual cluster based solution has been implemented by Computer Centre. The setup consists of three number of HP DL 380 G5 servers (Xeon 3.73 GHz Dual core, Dual processor servers with 4, 8, 8 GB RAM). One of the servers is configured as licence and backup server (for high availability) and remaining two servers are configured as application servers with Red Hat Enterprise Linux 4.0.
  GUI based 2D and 3D applications of OPERA need high performance industry standard OpenGL run-time libraries. All the three servers are optimized for OpenGL run-time libraries.
  

  
  
  Remote execution of OPERA on application server from window desktop

  Users can run their GUI based applications remotely from their desktops on RRCATNet. Free open source software Xming is configured for running applications with OpenGL support from remote Window desktops.

• Computing Server Ganak-1 (गणक-1)
  

  Itanium2 based HP 3600 series server (Dual core, Dual processor, 1.6 GHz Itanium2, 64-bit processor server with 16 GB RAM & 18 MB L3 cache) named as 'Ganak-1' has been commissioned for scientific computing applications. This server has been configured with 64-bit Red Hat Enterprise Linux (for Itanium IA64) version 5.2 operating system. Fortran compilers - g77, gfortran, Intel fortran (ifort) and C compilers - gcc, g++, Intel C compiler (icc) are installed on this server. Intel Math Kernel Library version 10 is also made available on this server.

• Power 5+ Server
  

  IBM Power 5+ RISC processor based computing server named as 'Avantika' has been commissioned and released to the scientific computing users. This is IBM P series server (p5 550/550Q) comprising of dual core 2.1 GHz Power 5+, 64-bit RISC processor with 36 MB L3, 1.9 MB L2 cache and 8 GB RAM. This server is configured with 64-bit Red Hat Enterprise Linux version 5 operating system. Fortran compilers - g77, gfortran & IBM XL fortran compiler (xlf, xlf90, xlf95, xlf2003) and C compilers - gcc, g++ & IBM XL C/C++ compiler (xlc, xlc++) are installed on this server. BLAS and LAPACK libraries are also compiled and configured on this server for advance computing uses. This server is configured in NIS enabled centralized setup and all users having NIS enabled login-id and password can access the server with same login-id and password.

• Compilers, Scientific Libraries and Tools
  
  Following compilers, scientific libraries and tools are made available to the computing users:
  
  

  • Absoft Pro Fortran compiler version 10.2 is installed and configured on Xeon based Linux server with graphical fx3 debugger. This is a Fortran 90/95/2003 compatible compiler and supports OpenMP & all available SSE3 Instructions.

  • Absoft Fortran 90/95 compiler version 9.0 is installed and configured on IBM P5+ server 'Avantika' with graphical fx2 debugger. Pre-configured BLAS, LAPACK90, ATLAS and PLplot (2D & 3D) graphics libraries are also configured with this Fortran 90/95 compatible compiler.

  • Lahey/Fujitsu Linux64 Fortran Pro (LF64 Pro) compiler version 8.1 is installed & configured on Xeon based Linux server. Wisk, Winteracter graphical packages, BLAS & LAPACK v4.0 libraries are also configured with this Fortran 90/95 compatible compiler.

  • IMSL Fortran numerical libraries for Absoft version 5.1 is also configured on Xeon based Linux server. The IMSL Fortran Library is a complete collection of mathematical and statistical algorithms which integrate the IMSL F90 Library, Math & Stat and Fortran 77 libraries into a single, cohesive package.

  • Math Kernel Library (MKL) has been setup and installed for scientific users. Math Kernel Library 9.0 is configured on 64-bit computing environment of Aryabhatta cluster.

  • IMSL Fortran 90 MP library consists of numerical algorithms using Fortran 90 language and constructs, including Fortran 90 array data types.

  • Nag Fortran Library is installed on Intel based servers on Linux platform.

  • NAG C Library is a comprehensive collection of functions for the solution of numerical and statistical problems.

  • CXML (Compaq Extended Math Library) LIB is a collection of mathematical routines optimized for Alpha based systems. These subroutines perform numerically intensive operations that occur frequently in engineering and scientific computing, such as linear algebra and signal processing. CXML can help reduce the cost of computation, enhance portability, and improve productivity.

  • CPML (Compaq Portable Math Library) is a library of math functions provided for the Linux Operating System running on Alpha platforms. The entry points and implementations in CPML are identical to those in Compaq's Tru64 UNIX. The CPML contains an extensive collection of real and complex math functions required by C99 and Fortran90.

  • CERNLIB - CERN Program Library has been configured on 32-bit & 64-bit Linux based computing servers and made available for computing uses.

  • Mathematica is a general computing environment, organizing many algorithmic, visualization, and user interface capabilities within a document-like user interface paradigm. It was originally conceived by Stephen Wolfram, developed by a team of mathematicians and programmers. It is a fully integrated environment for technical and scientific computing.

  • SYSTAT Software: Windows based software package SYSTAT (version 12) with five network floating licenses has been installed for users. SYSTAT is a powerful and versatile statistical software package. It provides a menu-driven interface to conduct simple analyses and produce 2D and 3D graphics for engineering and scientific data.

  • MATFOR Graphical Library: This graphical library contains high-level graphical procedures that support x-y plot, vector plot, contour, mesh plot, surface plot, isosurface plot, streamline, slicing plot, molecules plot etc. MATFOR Graphics Viewer displays high quality graphs without the need of Windows programming and graphics initialization.

  • Grace is a WYSIWYG 2D plotting tool for X Windows System and Motif.

  
  
• Auto Parallelization Software for multi processor and multi core Linux servers
  

  Auto parallelization software 'Vast-F/ Parallel' has been installed and configured on Intel Xeon and IBM P5+ based Linux servers. This software tool automatically converts serial code into code designed for multi-core or multi-processor systems using Absoft Fortran compiler and carries out additional optimization. Code re-structuring is done automatically and the original source code files are retained in unmodified form. VAST-F/ Parallel also provides OpenMP support. VAST-F/ Parallel 7.0 x64 Linux is used on Intel Xeon Linux multiprocessor server and VAST-F/ Parallel LinuxPPC is used on IBM P5+ based Linux server for auto parallelization.

• User Hall Facility
  

  User hall at IT building is used by RRCAT employees for conducting training programmes and for executing various Windows based applications. The setup has been revamped with Windows 2003 based server and 22 PCs as clients. LDAP based centralized authentication facilitates single sign-on, even for user hall facilities.