DOCSLIB.ORG

User's Guide on Software at Uow HPC Cluster

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text

Load more

User’s Guide on Software at UoW HPC Cluster Information Technology Services University of Wollongong ( Last updated on October 10, 2011) Contents 1 Requests for Software4 2 Environment Modules4 3 Software List4 4 How to Use7 4.1 Parallel Programming Libraries/Tools...................7 4.1.1 LAM/MPI.............................7 4.1.2 MPICH...............................7 4.1.3 MPICH2..............................7 4.1.4 OpenMPI..............................7 4.2 Compilers..................................8 4.2.1 GNU Compiler Collection (GCC).................8 4.2.2 Intel Fortran Compiler.......................8 4.2.3 Intel C Compiler..........................8 4.2.4 CMake...............................9 4.3 Scripting Languages............................9 4.3.1 Python...............................9 4.3.2 Perl.................................9 4.3.3 Java.................................9 4.4 Editors....................................9 4.4.1 Emacs................................9 4.4.2 Nano................................ 10 4.5 Math Libraries............................... 10 CONTENTS CONTENTS 4.5.1 FFTW................................ 10 4.5.2 FFTW3............................... 10 4.5.3 Intel Math Kernel Library (IMKL)................. 10 4.5.4 Basic Linear Algebra Subroutines (BLAS)............ 11 4.5.5 Linear Algebra PACKage (LAPACK)............... 11 4.5.6 Basic Linear Algebra Communication Subprograms (BLACS).. 12 4.5.7 Scalable LAPACK (ScaLAPACK)................. 12 4.5.8 Automatically Tuned Linear Algebra Software (ATLAS)..... 13 4.5.9 AMD Core Math Library (ACML)................. 13 4.5.10 GotoBLAS2............................ 14 4.6 Debuggers, Profilers and Simulators.................... 14 4.6.1 Valgrind............................... 14 4.6.2 Intel Debugger........................... 14 4.7 Graphics................................... 15 4.7.1 GNUPlot.............................. 15 4.8 Statistics and Mathematics Environments................. 15 4.8.1 R.................................. 15 4.9 Computational Chemistry.......................... 15 4.9.1 Quantum ESPRESSO....................... 15 4.9.2 GROMACS............................. 16 4.9.3 GAUSSIAN............................ 17 4.9.4 VMD................................ 18 4.9.5 NAMD............................... 18 4.9.6 ABINIT............................... 19 4.9.7 CPMD............................... 20 4.9.8 XCrySDen............................. 21 4.9.9 PSI3................................. 21 4.9.10 TINKER.............................. 22 4.9.11 MOPAC 2009............................ 23 4.9.12 DOCK............................... 23 4.9.13 NWChem.............................. 24 4.9.14 GAMESS.............................. 25 4.9.15 OpenBabel............................. 26 4.10 Informatics................................. 26 4.10.1 GhostScript............................. 26 4.10.2 ImageMagick............................ 26 4.11 Engineering................................. 26 CONTENTS CONTENTS 4.11.1 MATLAB.............................. 26 4.11.2 ANSYS............................... 31 4.11.3 ABAQUS.............................. 31 4.11.4 Ansoft Maxwell.......................... 31 4.11.5 LAMMPS.............................. 31 4.11.6 Materials Studio.......................... 32 4.12 Biology................................... 32 4.12.1 MrBayes.............................. 32 4.12.2 ATSAS............................... 33 3 SOFTWARE LIST 1 Requests for Software If you require software that is not installed on the UoW HPC cluster please send an email to the ITS admin team at hpc [email protected]. Please note: As there is currently no central budget for software licenses, any costs incurred for the requested software must be covered by user’s research group or the department of the university that the user is employed by. Please see the Section3 (Software List) for the details of the available software. 2 Environment Modules Package Environment Modules provide a great way to easily customize your shell environment to the requirements of whatever software you wish to use. They are available on the whole system. Always put the following line in your job script to active the environment modules: source /etc/profile.d/00-modules.sh 3 Software List Some software installed on the UoW HPC cluster is licensed and paid for by various departments and research groups on campus. The * under the ‘Prerequisite’ column indicates software that requires some discussion with the HPC admin to be able to access. Read the ‘How to Use’ section of this guide for more details. Please Note: All software installed on the UoW HPC cluster has passed the standard test suite (if existing). However, it is a user’s responsibility to check the correctness and validity of the software installed on the cluster prior to publishing any result by using them. Table 1: List of software Name Version Site Prerequisite 1. Parallel Programming Libraries/Tools (Sec. 4.1) LAMMPI 7.1.4 System MPICH 1.2.7 System MPICH2 1.3.1 System OpenMPI 1.3; 1.4.2; 1.4.3 System 2. Compilers (Sec. 4.2) GCC 4.1.2;4.5.2 System Intel Fortran Compiler 10.1.023; 11.1.046 System Intel C/C++ Compiler 10.1.023 System Continued on next page 4 3 SOFTWARE LIST Table 1 – continued from previous page Name Version Site Prerequisite CMake 2.8.3; 2.8.5 System 3. Scripting Languages (Sec. 4.3) Python 2.4.3 System Perl 5.12.2 System Java 1.6.0 17 Head-node 4. Editors (Sec. 4.4) Emacs 23.2.1 System Nano 2.2.6 System 5. Math Libraries (Sec. 4.5) FFTW 2.1.5 System FFTW3 3.2.2 System Intel MKL 10.2 System BLAS — System LAPACK 3.2.2 System BLACS — System SCALAPACK 1.8.0 System ATLAS 3.9.32 System ACML 4.4.0 System GotoBLAS2 1.13 System 6. Debuggers,Profilers and Simulators (Sec. 4.6) Valgrind 3.2.1 Head-node Intel Debugger 11.1.046 Head-node 7. Graphics (Sec. 4.7) GNUPlot 4.0 Head-node 8. Statistics and Mathematics Environments (Sec. 4.8) R 2.12.0 System 9. Computational Chemistry (Sec. 4.9) Quantum Espresso 4.2.1; 4.3 System GROMACS 4.5.1 System GAUSSIAN g03e01; g09a02 System * VMD 1.8.7; 1.9 System * NAMD 2.7b4; 2.7; 2.8 System * ABINIT 6.4.1 System CPMD 3.13.2; 3.15.1 System * Continued on next page 5 3 SOFTWARE LIST Table 1 – continued from previous page Name Version Site Prerequisite XCrySDen 1.5.21 System PSI3 3.4 System TINKER 5.0.19 System MOPAC2009 11.053 System DOCK 6.5 System * NWChem 6.0 System GAMESS 1Oct2010 System * OpenBabel 2.3.0 System 10. Informatics (Sec. 4.10) GhostScript 8.15 System ImageMagick 6.6.4-10 System 11. Engineering (Sec. 4.11) MATLAB R2010b; R2011a System ANSYS 12,13 System * ABAQUS 6.91 System * Ansoft Maxwell 14.0 System * LAMMPS 05Nov10 System Materials Studio 5.5 System * 12. Biology (Sec. 4.12) MrBayes 3.1.2 System ATSAS 2.4.1-4 System 6 4 HOW TO USE 4 How to Use 4.1 Parallel Programming Libraries/Tools 4.1.1 LAM/MPI To use LAM/MPI version 7.1.4 in the current shell, do %module load mpi/lam-7.1.4 NOTE: ‘%’ represents command line prompt all through this paper. 4.1.2 MPICH To use MPICH version 1.2.7 (gcc build) in the current shell, do %module load mpich/1.2.7-gcc 4.1.3 MPICH2 To use MPICH2 version 1.3.1 (gcc build) in the current shell, do %module load mpi/mpich2-1.3.1-gcc To use Intel Compiler build of MPICH2 %module load mpich2/1.3.1-itl 4.1.4 OpenMPI To use OpenMPI version 1.3 in the current shell, do %module load mpi/openmpi-1.3 To use OpenMPI version 1.4.2 in the current shell, do %module load mpi/openmpi-1.4.2 To use OpenMPI version 1.4.3 in the current shell, do %module load mpi/openmpi-1.4.3 OpenMPI 1.4.3 is RECOMMENDED as your default MPI environment. NOTE: MPI environment can be set for future shells by using environment switcher package. Please refer to ‘UOW HPC Cluster User Guide’ for further detail. 7 4.2 Compilers 4 HOW TO USE 4.2 Compilers 4.2.1 GNU Compiler Collection (GCC) GCC 4.1.2 is installed in system path which can be used directly. To use GCC version 4.5.2, do %module load gcc/4.5.2 in the command line or in your job script. 4.2.2 Intel Fortran Compiler The Intel Fortran Compiler is called ifort. To use Intel Fortran Compiler version 11.1.046, do %module load intel-fc/11.1.046 in the command line or add it in your job script. NOTE: This version is incompatible with intel-cc/10.1.023. To use Intel Fortran Compiler version 10.1.023, do %module load intel-fc/10.1.023 This version is compatible with intel-cc/10.2.023. To see all other versions of the compiler available do %module avail intel-fc Note that the debugger idb is now the graphical debugger. To use the non-graphical interface use idbc command. 4.2.3 Intel C Compiler The Intel C/C++ Compiler is called as icc and the C++ compiler is icpc. To use Intel C/C++ Compiler version 10.1.023, do %module load intel-cc/10.1.023 NOTE: This version is incompatible with intel-fc/11.1.046. To see all installed versions do %module avail intel-cc 8 4.3 Scripting Languages 4 HOW TO USE 4.2.4 CMake The cross-platform, open-source build system CMake is a family of tools designed to build, test and package software. CMake is used to control the software compilation process using simple platform and compiler independent configuration files. CMake generates native makefiles and workspaces that can be used in the compiler environment of your choice. To use CMake, type module load cmake/2.8.3 or module load cmake/2.8.5 in the command line. 4.3 Scripting Languages 4.3.1 Python Python is installed system wide (which means user can submit python based applications to the queue system). To use python, do %python 4.3.2 Perl To use Perl, do module load perl/5.12.2 in your job script.
Recommended publications
  • (LINUX) Computer in Three Parts

    (LINUX) Computer in Three Parts

    Science on a (LINUX) computer in three parts Introductory short couse, Part II Thorsten Becker University of Southern California, Los Angeles September 2007 The first part dealt with ● UNIX: what and why ● File system, Window managers ● Shell environment ● Editing files ● Command line tools ● Scripts and GUIs ● Virtualization Contents part II ● Typesetting ● Programming – common languages – philosophy – compiling, debugging, make, version control – C and F77 interfacing – libraries and packages ● Number crunching ● Visualization tools Programming: Traditional languages in the natural sciences ● Fortran: higher level, good for math – F77: legacy, don't use (but know how to read) – F90/F95: nice vector features, finally implements C capabilities (structures, memory allocation) ● C: low level (e.g. pointers), better structured – very close to UNIX philosophy – structures offer nice way of modular programming, see Wikipedia on C ● I recommend F95, and use C happily myself Programming: Some Languages that haven't completely made it to scientific computing ● C++: object oriented programming model – reusable objects with methods and such – can be partly realized by modular programming in C ● Java: what's good for commercial projects (or smart, or elegant) doesn't have to be good for scientific computing ● Concern about portability as well as general access Programming: Compromises ● Python – Object oriented – Interpreted – Interfaces easily with F90/C – Numerous scientific packages Programming: Other interpreted, high- abstraction languages
  • Some Notes on BLAS, SIMD, MIC and GPGPU (For Octave and Matlab)

    Some Notes on BLAS, SIMD, MIC and GPGPU (For Octave and Matlab)

    Some Notes on BLAS, SIMD, MIC and GPGPU (for Octave and Matlab) Christian Himpe ([email protected]) WWU Münster Institute for Computational and Applied Mathematics Software-Tools für die Numerische Mathematik 15.10.2014 About1 BLAS & LAPACK Affinity SIMD Automatic Offloading 1 Get your Buzzword-Bingo ready! BLAS & LAPACK BLAS (Basic Linear Algebra System) Level 1: vector-vector operations (dot-product, vector norms, generalized vector addition) Level 2: matrix-vector operations (generalized matrix-vector multiplication) Level 3: matrix-matrix operations (generalized matrix-matrix multiplication) LAPACK (Linear Algebra Package) Matrix Factorizations: LU, QR, Cholesky, Schur Least-Squares: LLS, LSE, GLM Eigenproblems: SEP, NEP, SVD Default (un-optimized) netlib reference implementation. Used by Octave, Matlab, SciPy/NumPy (Python), Julia, R. MKL Intel’s implementation of BLAS and LAPACK. MKL (Math Kernel Library) can offload computations to XeonPhi can use OpenMP provides additionally FFT, libm and 1D-interpolation Current Version: 11.0.5 (automatic offloading to Phis) Costs (we have it, and Matlab ships with it, too) Go to: https://software.intel.com/en-us/intel-mkl ACML AMD doesn’t want to be left out. ACML (AMD Core Math Libraries) Can use OpenCL for automatic offloading to GPUs Special version to exploit FMA(4) instructions Choice of compiled binaries (GFortran, Intel Fortran, Open64, PGI) Current Version: 6 (automatic offloading to GPUs) Free (but not open source) Go to: http://developer.amd.com/tools-and-sdks/ cpu-development/amd-core-math-library-acml OpenBLAS Open Source, the third kind. OpenBLAS Fork of GotoBLAS Good performance for dense operations (close to the MKL) Can use OpenMP and is compiled for specific architecture Current Version: 0.2.11 Open source (!) Go to: http://github.com/xianyi/OpenBLAS FlexiBLAS So many BLAS implementations, so little time..
  • Parallel Programming Models for Dense Linear Algebra on Heterogeneous Systems

    Parallel Programming Models for Dense Linear Algebra on Heterogeneous Systems

    DOI: 10.14529/jsfi150405 Parallel Programming Models for Dense Linear Algebra on Heterogeneous Systems M. Abalenkovs1, A. Abdelfattah2, J. Dongarra 1,2,3, M. Gates2, A. Haidar2, J. Kurzak2, P. Luszczek2, S. Tomov2, I. Yamazaki2, A. YarKhan2 c The Authors 2015. This paper is published with open access at SuperFri.org We present a review of the current best practices in parallel programming models for dense linear algebra (DLA) on heterogeneous architectures. We consider multicore CPUs, stand alone manycore coprocessors, GPUs, and combinations of these. Of interest is the evolution of the programming models for DLA libraries – in particular, the evolution from the popular LAPACK and ScaLAPACK libraries to their modernized counterparts PLASMA (for multicore CPUs) and MAGMA (for heterogeneous architectures), as well as other programming models and libraries. Besides providing insights into the programming techniques of the libraries considered, we outline our view of the current strengths and weaknesses of their programming models – especially in regards to hardware trends and ease of programming high-performance numerical software that current applications need – in order to motivate work and future directions for the next generation of parallel programming models for high-performance linear algebra libraries on heterogeneous systems. Keywords: Programming models, runtime, HPC, GPU, multicore, dense linear algebra. Introduction Parallelism in today’s computer architectures is pervasive – not only in systems from large supercomputers to laptops, but also in small portable devices like smart phones and watches. Along with parallelism, the level of heterogeneity in modern computing systems is also gradually increasing. Multicore CPUs are often combined with discrete high-performance accelerators like graphics processing units (GPUs) and coprocessors like Intel’s Xeon Phi, but are also included as integrated parts with system-on-chip (SoC) platforms, as in the AMD Fusion family of ap- plication processing units (APUs) or the NVIDIA Tegra mobile family of devices.
  • AMD Core Math Library (ACML)

    AMD Core Math Library (ACML)

    AMD Core Math Library (ACML) Version 4.2.0 Copyright c 2003-2008 Advanced Micro Devices, Inc., Numerical Algorithms Group Ltd. AMD, the AMD Arrow logo, AMD Opteron, AMD Athlon and combinations thereof are trademarks of Advanced Micro Devices, Inc. i Short Contents 1 Introduction................................... 1 2 General Information ............................. 2 3 BLAS: Basic Linear Algebra Subprograms ............. 19 4 LAPACK: Package of Linear Algebra Subroutines ........ 20 5 Fast Fourier Transforms (FFTs) .................... 24 6 Random Number Generators....................... 75 7 ACML MV: Fast Math and Fast Vector Math Library .... 163 8 References .................................. 229 Subject Index ................................... 230 Routine Index ................................... 233 ii Table of Contents 1 Introduction ............................... 1 2 General Information ....................... 2 2.1 Determining the best ACML version for your system ........... 2 2.2 Accessing the Library (Linux) ................................ 4 2.2.1 Accessing the Library under Linux using GNU gfortran/gcc ......................................................... 4 2.2.2 Accessing the Library under Linux using PGI compilers pgf77/pgf90/pgcc ......................................... 5 2.2.3 Accessing the Library under Linux using PathScale compilers pathf90/pathcc ........................................... 6 2.2.4 Accessing the Library under Linux using the NAGWare f95 compiler.................................................
  • 35.232-2016.43 Lamees Elhiny.Pdf

    35.232-2016.43 Lamees Elhiny.Pdf

    LOAD PARTITIONING FOR MATRIX-MATRIX MULTIPLICATION ON A CLUSTER OF CPU-GPU NODES USING THE DIVISIBLE LOAD PARADIGM by Lamees Elhiny A Thesis Presented to the Faculty of the American University of Sharjah College of Engineering in Partial Fulfillment of the Requirements for the Degree of Master of Science in Computer Engineering Sharjah, United Arab Emirates November 2016 c 2016. Lamees Elhiny. All rights reserved. Approval Signatures We, the undersigned, approve the Master’s Thesis of Lamees Elhiny Thesis Title: Load Partitioning for Matrix-Matrix Multiplication on a Cluster of CPU- GPU Nodes Using the Divisible Load Paradigm Signature Date of Signature (dd/mm/yyyy) ___________________________ _______________ Dr. Gerassimos Barlas Professor, Department of Computer Science and Engineering Thesis Advisor ___________________________ _______________ Dr. Khaled El-Fakih Associate Professor, Department of Computer Science and Engineering Thesis Committee Member ___________________________ _______________ Dr. Naoufel Werghi Associate Professor, Electrical and Computer Engineering Department Khalifa University of Science, Technology & Research (KUSTAR) Thesis Committee Member ___________________________ _______________ Dr. Fadi Aloul Head, Department of Computer Science and Engineering ___________________________ _______________ Dr. Mohamed El-Tarhuni Associate Dean, College of Engineering ___________________________ _______________ Dr. Richard Schoephoerster Dean, College of Engineering ___________________________ _______________ Dr. Khaled Assaleh Interim Vice Provost for Research and Graduate Studies Acknowledgements I would like to express my sincere gratitude to my thesis advisor, Dr. Gerassi- mos Barlas, for his guidance at every step throughout the thesis. I am thankful to Dr. Assim Sagahyroon, and to the Department of Computer Science & Engineering as well as the American University of Sharjah for offering me the Graduate Teaching Assistantship, which allowed me to pursue my graduate studies.
  • Parallel Siesta.Graffle

    Parallel Siesta.Graffle

    Siesta in parallel Toby White Dept. Earth Sciences, University of Cambridge ❖ How to build Siesta in parallel ❖ How to run Siesta in parallel ❖ How Siesta works in parallel ❖ How to use parallelism efficiently ❖ How to build Siesta in parallel Siesta SCALAPACK BLACS LAPACK MPI BLAS LAPACK Vector/matrix BLAS manipulation Linear Algebra PACKage http://netlib.org/lapack/ Basic Linear Algebra Subroutines http://netlib.org/blas/ ATLAS BLAS: http://atlas.sf.net Free, open source (needs separate LAPACK) GOTO BLAS: http://www.tacc.utexas.edu/resources/software/#blas Free, registration required, source available (needs separate LAPACK) Intel MKL (Math Kernel Library): Intel compiler only Not cheap (but often installed on supercomputers) ACML (AMD Core Math Library) http://developer.amd.com/acml.jsp Free, registration required. Versions for most compilers. Sun Performance Library http://developers.sun.com/sunstudio/perflib_index.html Free, registration required. Only for Sun compilers (Linux/Solaris) IBM ESSL (Engineering & Science Subroutine Library) http://www-03.ibm.com/systems/p/software/essl.html Free, registration required. Only for IBM compilers (Linux/AIX) Parallel MPI communication Message Passing Infrastructure http://www.mcs.anl.gov/mpi/ You probably don't care - your supercomputer will have (at least one) MPI version installed. Just in case, if you are building your own cluster: MPICH2: http://www-unix.mcs.anl.gov/mpi/mpich2/ Open-MPI: http://www.open-mpi.org And a lot of experimental super-fast versions ... SCALAPACK BLACS Parallel linear algebra Basic Linear Algebra Communication Subprograms http://netlib.org/blacs/ SCAlable LAPACK http://netlib.org/scalapack/ Intel MKL (Math Kernel Library): Intel compiler only AMCL (AMD Core Math Library) http://developer.amd.com/acml.jsp S3L (Sun Scalable Scientific Program Library) http://www.sun.com/software/products/clustertools/ IBM PESSL (Parallel Engineering & Science Subroutine Library) http://www-03.ibm.com/systems/p/software/essl.html Previously mentioned libraries are not the only ones - just most common.
  • AMD Core Math Library (ACML) ©

    AMD Core Math Library (ACML) ©

    ¨ AMD Core Math Library (ACML) © Version 5.3.1 Copyright c 2003-2013 Advanced Micro Devices, Inc., Numerical Algorithms Group Ltd. All rights reserved. The contents of this document are provided in connection with Advanced Micro Devices, Inc. (\AMD") products. AMD makes no representations or warranties with respect to the accuracy or completeness of the contents of this publication and reserves the right to make changes to specifications and product descriptions at any time without notice. The information contained herein may be of a preliminary or advance nature and is subject to change without notice. No license, whether express, implied, arising by estoppel, or oth- erwise, to any intellectual property rights are granted by this publication. Except as set forth in AMD's Standard Terms and Conditions of Sale, AMD assumes no liability whatso- ever, and disclaims any express or implied warranty, relating to its products including, but not limited to, the implied warranty of merchantability, fitness for a particular purpose, or infringement of any intellectual property right. AMD's products are not designed, intended, authorized or warranted for use as components in systems intended for surgical implant into the body, or in other applications intended to support or sustain life, or in any other application in which the failure of AMD's product could create a situation where personal injury, death, or severe property or environmental damage may occur. AMD reserves the right to discontinue or make changes to its products at any time without notice. Trademarks AMD, the AMD Arrow logo, and combinations thereof, AMD Athlon, and AMD Opteron are trademarks of Advanced Micro Devices, Inc.
  • AMD Optimizing CPU Libraries User Guide Version 2.1

    AMD Optimizing CPU Libraries User Guide Version 2.1

    [AMD Public Use] AMD Optimizing CPU Libraries User Guide Version 2.1 [AMD Public Use] AOCL User Guide 2.1 Table of Contents 1. Introduction .......................................................................................................................................... 4 2. Supported Operating Systems and Compliers ...................................................................................... 5 3. BLIS library for AMD .............................................................................................................................. 6 3.1. Installation ........................................................................................................................................ 6 3.1.1. Build BLIS from source .................................................................................................................. 6 3.1.1.1. Single-thread BLIS ..................................................................................................................... 6 3.1.1.2. Multi-threaded BLIS .................................................................................................................. 6 3.1.2. Using pre-built binaries ................................................................................................................. 7 3.2. Usage ................................................................................................................................................. 7 3.2.1. BLIS Usage in FORTRAN ................................................................................................................
  • Best Practice Guide - Generic X86 Vegard Eide, NTNU Nikos Anastopoulos, GRNET Henrik Nagel, NTNU 02-05-2013

    Best Practice Guide - Generic X86 Vegard Eide, NTNU Nikos Anastopoulos, GRNET Henrik Nagel, NTNU 02-05-2013

    Best Practice Guide - Generic x86 Vegard Eide, NTNU Nikos Anastopoulos, GRNET Henrik Nagel, NTNU 02-05-2013 1 Best Practice Guide - Generic x86 Table of Contents 1. Introduction .............................................................................................................................. 3 2. x86 - Basic Properties ................................................................................................................ 3 2.1. Basic Properties .............................................................................................................. 3 2.2. Simultaneous Multithreading ............................................................................................. 4 3. Programming Environment ......................................................................................................... 5 3.1. Modules ........................................................................................................................ 5 3.2. Compiling ..................................................................................................................... 6 3.2.1. Compilers ........................................................................................................... 6 3.2.2. General Compiler Flags ......................................................................................... 6 3.2.2.1. GCC ........................................................................................................ 6 3.2.2.2. Intel ........................................................................................................
  • Accelerating Dense Linear Algebra for Gpus, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach

    Accelerating Dense Linear Algebra for Gpus, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach

    University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange Masters Theses Graduate School 8-2010 Accelerating Dense Linear Algebra for GPUs, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach Rajib Kumar Nath [email protected] Follow this and additional works at: https://trace.tennessee.edu/utk_gradthes Recommended Citation Nath, Rajib Kumar, "Accelerating Dense Linear Algebra for GPUs, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach. " Master's Thesis, University of Tennessee, 2010. https://trace.tennessee.edu/utk_gradthes/734 This Thesis is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Masters Theses by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council: I am submitting herewith a thesis written by Rajib Kumar Nath entitled "Accelerating Dense Linear Algebra for GPUs, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach." I have examined the final electronic copy of this thesis for form and content and recommend that it be accepted in partial fulfillment of the equirr ements for the degree of Master of Science, with a major in Computer Science. Jack Dongarra, Major Professor We have read this thesis and recommend its acceptance: Stanimire Z. Tomov, Lynne E. Parker Accepted for the Council: Carolyn R. Hodges Vice Provost and Dean of the Graduate School (Original signatures are on file with official studentecor r ds.) To the Graduate Council: I am submitting herewith a thesis written by Rajib Kumar Nath entitled \Accel- erating Dense Linear Algebra for GPUs, Multicores and Hybrid Architectures: an Autotuned and Algorithmic Approach." I have examined the final paper copy of this thesis for form and content and recommend that it be accepted in partial fulfillment of the requirements for the degree of Master of Science, with a major in Computer Science.
  • AMD Core Math Library (ACML) ©

    AMD Core Math Library (ACML) ©

    ¨ AMD Core Math Library (ACML) © Version 6.1.0 Copyright c 2003-2014 Advanced Micro Devices, Inc., Numerical Algorithms Group Ltd. All rights reserved. The contents of this document are provided in connection with Advanced Micro Devices, Inc. (\AMD") products. AMD makes no representations or warranties with respect to the accuracy or completeness of the contents of this publication and reserves the right to make changes to specifications and product descriptions at any time without notice. The information contained herein may be of a preliminary or advance nature and is subject to change without notice. No license, whether express, implied, arising by estoppel, or oth- erwise, to any intellectual property rights are granted by this publication. Except as set forth in AMD's Standard Terms and Conditions of Sale, AMD assumes no liability whatso- ever, and disclaims any express or implied warranty, relating to its products including, but not limited to, the implied warranty of merchantability, fitness for a particular purpose, or infringement of any intellectual property right. AMD's products are not designed, intended, authorized or warranted for use as components in systems intended for surgical implant into the body, or in other applications intended to support or sustain life, or in any other application in which the failure of AMD's product could create a situation where personal injury, death, or severe property or environmental damage may occur. AMD reserves the right to discontinue or make changes to its products at any time without notice. Trademarks AMD, the AMD Arrow logo, and combinations thereof, AMD Athlon, and AMD Opteron are trademarks of Advanced Micro Devices, Inc.
  • Paket Som Matlab, Maple, Mathematica. • Generella

    Paket Som Matlab, Maple, Mathematica. • Generella

    Numerisk programvara NAG - The Numerical Algorithms Group Ltd Olika typer av programvara: http://www.nag.com/ • Paket som Matlab, Maple, Mathematica. NAG offers libraries in Fortran 77, Fortran 90, C, as well as • Generella rutinbibliotek, NAG, IMSL. F˚ar tillhandah˚alla an SMP Library and a Parallel Library for shared memory and huvudprogram etc. och l¨anka till en biblioteksfil. distributed memory parallel computing respectively. We have most recently added Data Mining Components to our product • Bibliotek f¨or speciella problemomr˚aden. range. Lapack (RISC-optimerat) rutiner f¨or Ax = b, minx ||Ax − b||2 och Ax = λx. Inneh˚aller s˚adar 1100 rutiner. (Linpack ¨aldre bibliotek f¨or CISC.) ¨ BLAS, t.ex. σ = xT y, y = Ax och C = AB. List of Contents, Release 4 Finns i NAG, IMSL men ¨aven i: Chapter 1: Utilities • Datorspecifika bibliotek, t.ex. Chapter 3: Special Functions – AMD: ACML (AMD Core Math Library) Chapter 4: Matrix and Vector Operations – Intel: MKL (Intel Math Kernel library). Chapter 5: Linear Equations – Sun: Sunperf (Sun Performance Library). – SGI: complib.sgimath Module 5.1: Solves a general real or complex system of linear (Scientific and Mathematical Library). equations with one or many right-hand sides – IBM: ESSL (Engineering and Scientific Subroutine Library). Module 5.2: Solves a real or complex, symmetric or Hermitian system of linear equations with one or many right-hand sides • Rutiner f¨or speciella till¨ampningar. Finns samlingar p˚awww. Netlib (d¨ar man ¨aven finner Lapack), Module 5.3: Solves a real or complex triangular system of linear Rutinerna ¨ar normalt skrivna i Fortran (finns Fortran66, 77, 90, equations 95).