DOCSLIB.ORG

Evolving Software Repositories

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Evolving Software Repositories 1 Evolving Software Rep ositories http://www.netli b.org/utk/pro ject s/esr/ Jack Dongarra UniversityofTennessee and Oak Ridge National Lab oratory Ron Boisvert National Institute of Standards and Technology Eric Grosse AT&T Bell Lab oratories 2 Pro ject Fo cus Areas NHSE Overview Resource Cataloging and Distribution System RCDS Safe execution environments for mobile co de Application-l evel and content-oriented to ols Rep ository interop erabili ty Distributed, semantic-based searching 3 NHSE National HPCC Software Exchange NASA plus other agencies funded CRPC pro ject Center for ResearchonParallel Computation CRPC { Argonne National Lab oratory { California Institute of Technology { Rice University { Syracuse University { UniversityofTennessee Uniform interface to distributed HPCC software rep ositories Facilitation of cross-agency and interdisciplinary software reuse Material from ASTA, HPCS, and I ITA comp onents of the HPCC program http://www.netlib.org/nhse/ 4 Goals: Capture, preserve and makeavailable all software and software- related artifacts pro duced by the federal HPCC program. Soft- ware related artifacts include algorithms, sp eci cations, designs, do cumentation, rep ort, ... Promote formation, growth, and interop eration of discipline-oriented rep ositories that organize, evaluate, and add value to individual contributions. Employ and develop where necessary state-of-the-art technologies for assisting users in nding, understanding, and using HPCC software and technologies. 5 Bene ts: 1. Faster development of high-quality software so that scientists can sp end less time writing and debugging programs and more time on research problems. 2. Less duplication of software development e ort by sharing of soft- ware mo dules. 3. Less time and e ort sp ent in lo cating relevant software and in- formation through the use of appropriate indexing and search mechanisms and domain-sp eci c exp ert help systems. 4. Reducing information overload through the use of lters and au- tomatic search mechanisms. 6 Intended Audience: HPCC application and computer science community { Source of material for NHSE Users of NASA, NSF, DOE and other sup ercomputer centers { Go o d targets for NHSE { Natural supp ort organization: sup ercomputer center sta Other users of high p erformance computers { Current and p otential industrial users { No natural supp ort organization Applicable to other domains 7 NHSE Comp onents: Discipline Oriented Rep ositories Submission and Review Common Infrastructure { Resource cataloging and distributed system { Rep ository to ols and resource center { Naming and authentication { Publishing to ols HPCC Sp eci c Searching Outreach and technology transition { To the HPCC user community and industry Measurement Hyp ertext Road Map Selective Capitalization of Emerging Technologies Collection Management 8 Physical Physical Repository Physical Repository 2 Repository 1 n catalog catalog info info catalog info file request retrieved file NHSE Search/Browse Interface search results search request Virtual Repository Architecture 9 NHSE Based on Existing Technologies { WWW Browser Mosaic / Netscap e / etc Distributed / Scalable URL: http://www.netlib.org/nhse/ { Netlib Rep ository for math software since 1985 Rep ositories Currently Available { Netlib, Softlib, CITlib { ASSET - Asset Source for SW Engineering Tech. { CARDS - Comprehensive Approachto Reusable Defense SW { ELSA - Electronic Library Services and Appl. { GAMS Virtual Software Rep ository { STARS - SW Technology for Adaptable, Reliable Systems { Many examples related to GC problems Currently Available Information { NHSE currently p oints to 350+ mo dules software catalog, tech rep orts and pap ers parallel pro cessing to ols numerical libraries Grand Challenge prototyp e co des data analysis and visulization b enchmarks 10 Discipline Oriented Rep ositorie s Interop eration The NHSE will catalog and provide access to software and software- related artifacts from all the HPCC software rep ositories. Assets accessible from other existing software rep ositories, such as ASSET, CARDS, DSRS, and ESLA, may also b e of interest to NHSE users. The NHSE will b e participating in a small-scale interop erability exp eriment with the ab ove rep ositories to help de ne require- ments for further interop eration e orts. The NHSE will also b e working with the Reuse Library Interop er- ability Group RIG on establishing standards for unique naming, asset description and classi cation, and asset evaluation. In the future, the NHSE will interop erate with these other rep os- itories so that software from them may b e retrieved directly from the NHSE interface. 11 Netlib - Network Access to Mathematical Software and Data Began in 1985 { JD and Eric Grosse, AT&T Bell Labs Motivated by the need for cost-e ective, timely distribution of high-quality mathematical software to the community. Designed to send, by return electronic mail, requested items. Automatic mechanism for the disseminate of public domain soft- ware. { Still in use and growing { Mirrored at a numb er of sites netlib2.cs.utk.edu netlib1.epm.ornl.gov research.att.com netlib.no unix.hensa.ac.uk ftp.zip-b erlin.de nchc.edu.tw Mo derated collection of high-quality math software Distributed maintenance Mo del for domain-sp eci c rep ositories 12 Netlib { Network access to mathematical software and data Jack Dongarra Univ. Tenn. and ORNL Eric Grosse AT&T Bel l Labs, Murray Hil l NJ 13 netlib Started in 1985. Motivated by the research community Uses email for the distribution. Has grown in p opularity and scop e. Funded by the NSF and Bell Labs. 14 The development of NETLIB was motivated by the need for cost- e ective, timely distribution of high-quality mathematical soft- ware to the research community at large. The system was designed to send, by return electronic mail, re- quested routines together with subsidiary routines and any re- lated do cuments or test programs supplied by the authors. Automatic mechanism for the disseminate of public domain soft- ware. 15 Try: mail [email protected] mail [email protected] send index send rs from eispack who is Golub Collection includes: Linpack Eispack Fishpack Odepack ACM TOMS Benchmark Bihar Blas BMP Conformal f2c FMM Fnlib Fftpack Hompack Lanczos LP/data Minpack Napack NL2SOL Odepack Paranoia Pltmg Polyhedra Port Pppack Quadpack SIAM memship Sparspak Typ esetting Vanhu el Voronoi ... 16 Netlib provides the following features: There are no administrativechannels to go through. Since no human pro cesses the request, it is p ossible to get software at any time, even in the middle of the night. The most up-to-date version is always available. Individual routines or pieces of a package can b e obtained instead of a whole collection. Over around 15000 requests a day. Software collection ab out 1 Gbytes 21K le in 330 libraries/directories 17 Interdisciplinary resource Software Parallel pro cessing collection Data To ols Rep orts Do cumentation Benchmarks Journal information 18 Synchronization and Netlib Sites: Still in use and growing Mirrored at a numb er of sites { netlib2.cs.utk.edu { netlib1.epm.ornl.gov { research.att.com { netlib.no { unix.hensa.ac.uk { ftp.zip-b erlin.de 19 O sho ots Other sites running the netlib pro cessor, but to supp ort other databases: statlib@temp er.stat.cmu.edu statistics [email protected] T X E [email protected] Reduce symbolic algebra [email protected] Maple symbolic algebra [email protected] benchmarks Well overahundred copies of netlib itself have b een shipp ed. 20 NETLIB do es not o er Technical assistance in determining and correcting problems with library software. Pro cedures for testing or validating co des. A uniform style for programming and do cumentation. Uniform error handling within the library. 21 Requests Made to the Netlib Repositories at the Univ. of Tennessee & ORNL 8,946,816 total requests to these repositories as of Feb 15, 1996 4000 3000 2000 Requests x 1,000 1000 785,025 0 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Year HTTP Gopher FTP XNetlib EMail Data as of - Feb 15, 1996 at 02:07:27 22 Breakdown of requests to each Netlib library (an alphabetical listing is also available.) Data as of 02/15/96 at 02:09:20 Library Name Number of accesses lapack 475,979 pvm3 379,849 linpack 256,403 slatec 248,292 blas 178,728 clapack 129,256 linalg 127,022 eispack 126,116 slatec/src 118,366 toms 117,152 f2c 98,025 c++ 96,774 benchmark 85,552 master 69,997 f2c/src 67,415 minpack 60,632 fn 59,781 fftpack 58,805 na-digest 50,970 port 49,496 slatec/lin 46,800 hence 45,229 confdb 43,118 slatec/chk 37,640 c++/answerbook 37,524 napack 36,719 23 Discipline Oriented Rep ositorie s Di erent discipline s will maintain their own software rep ositories Users should not need to access each of these rep ositories sepa- rately NHSE will provide a uniform interface to a virtual HPCC software rep ository which will b e built on top of the distributed set of discipline-oriented rep ositories. The interface will assist the user in lo cating relevant resources and in retrieving these resources. A combined browse/searchinterface will allow the user to explore the various HPCC areas and b ecome familiar with the available resources. A longer term goal of the NHSE is to provide users with domain- sp eci c exp ert help in lo cating and understanding relevant re- sources. 24 Discipline Oriented Rep ositorie s HPCC Cataloging To enable searching, cataloging information must b e made avail- able for NHSE assets. Eachphysical rep ository will b e resp onsible for maintaining a network-accessible le containing such cataloging information. These les will b e retrieved and indexed by an NHSE indexer on a regular basis, and the resulting searchable index will b e replicated for reliability.
Load more

Recommended publications
  • Fortran Resources 1
    Fortran Resources 1 Ian D Chivers Jane Sleightholme May 7, 2021 1The original basis for this document was Mike Metcalf’s Fortran Information File. The next input came from people on comp-fortran-90. Details of how to subscribe or browse this list can be found in this document. If you have any corrections, additions, suggestions etc to make please contact us and we will endeavor to include your comments in later versions. Thanks to all the people who have contributed. Revision history The most recent version can be found at https://www.fortranplus.co.uk/fortran-information/ and the files section of the comp-fortran-90 list. https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=comp-fortran-90 • May 2021. Major update to the Intel entry. Also changes to the editors and IDE section, the graphics section, and the parallel programming section. • October 2020. Added an entry for Nvidia to the compiler section. Nvidia has integrated the PGI compiler suite into their NVIDIA HPC SDK product. Nvidia are also contributing to the LLVM Flang project. Updated the ’Additional Compiler Information’ entry in the compiler section. The Polyhedron benchmarks discuss automatic parallelisation. The fortranplus entry covers the diagnostic capability of the Cray, gfortran, Intel, Nag, Oracle and Nvidia compilers. Updated one entry and removed three others from the software tools section. Added ’Fortran Discourse’ to the e-lists section. We have also made changes to the Latex style sheet. • September 2020. Added a computer arithmetic and IEEE formats section. • June 2020. Updated the compiler entry with details of standard conformance.
    [Show full text]
  • Life As a Developer of Numerical Software
    A Brief History of Numerical Libraries Sven Hammarling NAG Ltd, Oxford & University of Manchester First – Something about Jack Jack’s thesis (August 1980) 30 years ago! TOMS Algorithm 589 Small Selection of Jack’s Projects • Netlib and other software repositories • NA Digest and na-net • PVM and MPI • TOP 500 and computer benchmarking • NetSolve and other distributed computing projects • Numerical linear algebra Onto the Rest of the Talk! Rough Outline • History and influences • Fortran • Floating Point Arithmetic • Libraries and packages • Proceedings and Books • Summary Ada Lovelace (Countess Lovelace) Born Augusta Ada Byron 1815 – 1852 The language Ada was named after her “Is thy face like thy mother’s, my fair child! Ada! sole daughter of my house and of my heart? When last I saw thy young blue eyes they smiled, And then we parted,-not as now we part, but with a hope” Childe Harold’s Pilgramage, Lord Byron Program for the Bernoulli Numbers Manchester Baby, 21 June 1948 (Replica) 19 Kilburn/Tootill Program to compute the highest proper factor 218 218 took 52 minutes 1.5 million instructions 3.5 million store accesses First published numerical library, 1951 First use of the word subroutine? Quality Numerical Software • Should be: – Numerically stable, with measures of quality of solution – Reliable and robust – Accompanied by test software – Useful and user friendly with example programs – Fully documented – Portable – Efficient “I have little doubt that about 80 per cent. of all the results printed from the computer are in error to a much greater extent than the user would believe, ...'' Leslie Fox, IMA Bulletin, 1971 “Giving business people spreadsheets is like giving children circular saws.
    [Show full text]
  • Numerical and Parallel Libraries
    Numerical and Parallel Libraries Uwe Küster University of Stuttgart High-Performance Computing-Center Stuttgart (HLRS) www.hlrs.de Uwe Küster Slide 1 Höchstleistungsrechenzentrum Stuttgart Numerical Libraries Public Domain commercial vendor specific Uwe Küster Slide 2 Höchstleistungsrechenzentrum Stuttgart 33. — Numerical and Parallel Libraries — 33. 33-1 Overview • numerical libraries for linear systems – dense – sparse •FFT • support for parallelization Uwe Küster Slide 3 Höchstleistungsrechenzentrum Stuttgart Public Domain Lapack-3 linear equations, eigenproblems BLAS fast linear kernels Linpack linear equations Eispack eigenproblems Slatec old library, large functionality Quadpack numerical quadrature Itpack sparse problems pim linear systems PETSc linear systems Netlib Server best server http://www.netlib.org/utk/papers/iterative-survey/packages.html Uwe Küster Slide 4 Höchstleistungsrechenzentrum Stuttgart 33. — Numerical and Parallel Libraries — 33. 33-2 netlib server for all public domain numerical programs and libraries http://www.netlib.org Uwe Küster Slide 5 Höchstleistungsrechenzentrum Stuttgart Contents of netlib access aicm alliant amos ampl anl-reports apollo atlas benchmark bib bibnet bihar blacs blas blast bmp c c++ cephes chammp cheney-kincaid clapack commercial confdb conformal contin control crc cumulvs ddsv dierckx diffpack domino eispack elefunt env f2c fdlibm fftpack fishpack fitpack floppy fmm fn fortran fortran-m fp gcv gmat gnu go graphics harwell hence hompack hpf hypercube ieeecss ijsa image intercom itpack
    [Show full text]
  • Over-Scalapack.Pdf
    1 2 The Situation: Parallel scienti c applications are typically Overview of ScaLAPACK written from scratch, or manually adapted from sequen- tial programs Jack Dongarra using simple SPMD mo del in Fortran or C Computer Science Department UniversityofTennessee with explicit message passing primitives and Mathematical Sciences Section Which means Oak Ridge National Lab oratory similar comp onents are co ded over and over again, co des are dicult to develop and maintain http://w ww .ne tl ib. org /ut k/p e opl e/J ackDo ngarra.html debugging particularly unpleasant... dicult to reuse comp onents not really designed for long-term software solution 3 4 What should math software lo ok like? The NetSolve Client Many more p ossibilities than shared memory Virtual Software Library Multiplicityofinterfaces ! Ease of use Possible approaches { Minimum change to status quo Programming interfaces Message passing and ob ject oriented { Reusable templates { requires sophisticated user Cinterface FORTRAN interface { Problem solving environments Integrated systems a la Matlab, Mathematica use with heterogeneous platform Interactiveinterfaces { Computational server, like netlib/xnetlib f2c Non-graphic interfaces Some users want high p erf., access to details {MATLAB interface Others sacri ce sp eed to hide details { UNIX-Shell interface Not enough p enetration of libraries into hardest appli- Graphic interfaces cations on vector machines { TK/TCL interface Need to lo ok at applications and rethink mathematical software { HotJava Browser
    [Show full text]
  • Prospectus for the Next LAPACK and Scalapack Libraries
    Prospectus for the Next LAPACK and ScaLAPACK Libraries James Demmel1, Jack Dongarra23, Beresford Parlett1, William Kahan1, Ming Gu1, David Bindel1, Yozo Hida1, Xiaoye Li1, Osni Marques1, E. Jason Riedy1, Christof Voemel1, Julien Langou2, Piotr Luszczek2, Jakub Kurzak2, Alfredo Buttari2, Julie Langou2, Stanimire Tomov2 1 University of California, Berkeley CA 94720, USA, 2 University of Tennessee, Knoxville TN 37996, USA, 3 Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA, 1 Introduction Dense linear algebra (DLA) forms the core of many scienti¯c computing appli- cations. Consequently, there is continuous interest and demand for the devel- opment of increasingly better algorithms in the ¯eld. Here 'better' has a broad meaning, and includes improved reliability, accuracy, robustness, ease of use, and most importantly new or improved algorithms that would more e±ciently use the available computational resources to speed up the computation. The rapidly evolving high end computing systems and the close dependence of DLA algo- rithms on the computational environment is what makes the ¯eld particularly dynamic. A typical example of the importance and impact of this dependence is the development of LAPACK [1] (and later ScaLAPACK [2]) as a successor to the well known and formerly widely used LINPACK [3] and EISPACK [3] libraries. Both LINPACK and EISPACK were based, and their e±ciency depended, on optimized Level 1 BLAS [4]. Hardware development trends though, and in par- ticular an increasing Processor-to-Memory speed gap of approximately 50% per year, started to increasingly show the ine±ciency of Level 1 BLAS vs Level 2 and 3 BLAS, which prompted e®orts to reorganize DLA algorithms to use block matrix operations in their innermost loops.
    [Show full text]
  • Jack Dongarra: Supercomputing Expert and Mathematical Software Specialist
    Biographies Jack Dongarra: Supercomputing Expert and Mathematical Software Specialist Thomas Haigh University of Wisconsin Editor: Thomas Haigh Jack J. Dongarra was born in Chicago in 1950 to a he applied for an internship at nearby Argonne National family of Sicilian immigrants. He remembers himself as Laboratory as part of a program that rewarded a small an undistinguished student during his time at a local group of undergraduates with college credit. Dongarra Catholic elementary school, burdened by undiagnosed credits his success here, against strong competition from dyslexia.Onlylater,inhighschool,didhebeginto students attending far more prestigious schools, to Leff’s connect material from science classes with his love of friendship with Jim Pool who was then associate director taking machines apart and tinkering with them. Inspired of the Applied Mathematics Division at Argonne.2 by his science teacher, he attended Chicago State University and majored in mathematics, thinking that EISPACK this would combine well with education courses to equip Dongarra was supervised by Brian Smith, a young him for a high school teaching career. The first person in researcher whose primary concern at the time was the his family to go to college, he lived at home and worked lab’s EISPACK project. Although budget cuts forced Pool to in a pizza restaurant to cover the cost of his education.1 make substantial layoffs during his time as acting In 1972, during his senior year, a series of chance director of the Applied Mathematics Division in 1970– events reshaped Dongarra’s planned career. On the 1971, he had made a special effort to find funds to suggestion of Harvey Leff, one of his physics professors, protect the project and hire Smith.
    [Show full text]
  • Randnla: Randomized Numerical Linear Algebra
    review articles DOI:10.1145/2842602 generation mechanisms—as an algo- Randomization offers new benefits rithmic or computational resource for the develop ment of improved algo- for large-scale linear algebra computations. rithms for fundamental matrix prob- lems such as matrix multiplication, BY PETROS DRINEAS AND MICHAEL W. MAHONEY least-squares (LS) approximation, low- rank matrix approxi mation, and Lapla- cian-based linear equ ation solvers. Randomized Numerical Linear Algebra (RandNLA) is an interdisci- RandNLA: plinary research area that exploits randomization as a computational resource to develop improved algo- rithms for large-scale linear algebra Randomized problems.32 From a foundational per- spective, RandNLA has its roots in theoretical computer science (TCS), with deep connections to mathemat- Numerical ics (convex analysis, probability theory, metric embedding theory) and applied mathematics (scientific computing, signal processing, numerical linear Linear algebra). From an applied perspec- tive, RandNLA is a vital new tool for machine learning, statistics, and data analysis. Well-engineered implemen- Algebra tations have already outperformed highly optimized software libraries for ubiquitous problems such as least- squares,4,35 with good scalability in par- allel and distributed environments. 52 Moreover, RandNLA promises a sound algorithmic and statistical foundation for modern large-scale data analysis. MATRICES ARE UBIQUITOUS in computer science, statistics, and applied mathematics. An m × n key insights matrix can encode information about m objects ˽ Randomization isn’t just used to model noise in data; it can be a powerful (each described by n features), or the behavior of a computational resource to develop discretized differential operator on a finite element algorithms with improved running times and stability properties as well as mesh; an n × n positive-definite matrix can encode algorithms that are more interpretable in the correlations between all pairs of n objects, or the downstream data science applications.
    [Show full text]
  • C++ API for BLAS and LAPACK
    2 C++ API for BLAS and LAPACK Mark Gates ICL1 Piotr Luszczek ICL Ahmad Abdelfattah ICL Jakub Kurzak ICL Jack Dongarra ICL Konstantin Arturov Intel2 Cris Cecka NVIDIA3 Chip Freitag AMD4 1Innovative Computing Laboratory 2Intel Corporation 3NVIDIA Corporation 4Advanced Micro Devices, Inc. February 21, 2018 This research was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of two U.S. Department of Energy organizations (Office of Science and the National Nuclear Security Administration) responsible for the planning and preparation of a capable exascale ecosystem, including software, applications, hardware, advanced system engineering and early testbed platforms, in support of the nation's exascale computing imperative. Revision Notes 06-2017 first publication 02-2018 • copy editing, • new cover. 03-2018 • adding a section about GPU support, • adding Ahmad Abdelfattah as an author. @techreport{gates2017cpp, author={Gates, Mark and Luszczek, Piotr and Abdelfattah, Ahmad and Kurzak, Jakub and Dongarra, Jack and Arturov, Konstantin and Cecka, Cris and Freitag, Chip}, title={{SLATE} Working Note 2: C++ {API} for {BLAS} and {LAPACK}}, institution={Innovative Computing Laboratory, University of Tennessee}, year={2017}, month={June}, number={ICL-UT-17-03}, note={revision 03-2018} } i Contents 1 Introduction and Motivation1 2 Standards and Trends2 2.1 Programming Language Fortran............................ 2 2.1.1 FORTRAN 77.................................. 2 2.1.2 BLAST XBLAS................................. 4 2.1.3 Fortran 90.................................... 4 2.2 Programming Language C................................ 5 2.2.1 Netlib CBLAS.................................. 5 2.2.2 Intel MKL CBLAS................................ 6 2.2.3 Netlib lapack cwrapper............................. 6 2.2.4 LAPACKE.................................... 7 2.2.5 Next-Generation BLAS: \BLAS G2".....................
    [Show full text]
  • 2 Accessing Matlab at ER4
    Department of Electrical Engineering EE281 Intro duction to MATLAB on the Region IV Computing Facilities 1 What is Matlab? Matlab is a high-p erformance interactive software package for scienti c and enginnering numeric computation. Matlab integrates numerical analysis, matrix computation, signal pro cessing, and graphics in an easy-to-use envi- ronment without traditional programming. The name Matlab stands for matrix laboratory. Matlab was originally written to provide easy access to the matrix software develop ed by the LIN- PACK and EISPACK pro jects. Matlab is an interactive system whose basic data element is a matrix that do es not require dimensioning. Furthermore, problem solutions are expressed in Matlab almost exactly as they are written mathematically. Matlab has evolved over the years with inputs from many users. Matlab Toolboxes are sp ecialized collections of Matlab les designed for solving particular classes of functions. Currently available to olb oxes include: signal pro cessing control system optimization neural networks system identi cation robust control analysis splines symb olic mathematics image pro cessing statistics 2 Accessing Matlab at ER4 Sit down at a workstation. Log on. login: smithj <press return> password: 1234js <press return> Use the mouse button to view the Ro ot Menu. Drag the cursor to \Design Tools" then to \Matlab." Release the mouse button. You are now running the Matlab interactive software package. The prompt is a double \greater than" sign. You may use emacs line editor commands and the arrow keys while typing in Matlab. 1 3 An Intro ductory Demonstration Execute the following command to view aquick intro duction to Matlab.
    [Show full text]
  • CS 267 Dense Linear Algebra: History and Structure, Parallel Matrix
    Quick review of earlier lecture •" What do you call •" A program written in PyGAS, a Global Address CS 267 Space language based on Python… Dense Linear Algebra: •" That uses a Monte Carlo simulation algorithm to approximate π … History and Structure, •" That has a race condition, so that it gives you a Parallel Matrix Multiplication" different funny answer every time you run it? James Demmel! Monte - π - thon ! www.cs.berkeley.edu/~demmel/cs267_Spr16! ! 02/25/2016! CS267 Lecture 12! 1! 02/25/2016! CS267 Lecture 12! 2! Outline Outline •" History and motivation •" History and motivation •" What is dense linear algebra? •" What is dense linear algebra? •" Why minimize communication? •" Why minimize communication? •" Lower bound on communication •" Lower bound on communication •" Parallel Matrix-matrix multiplication •" Parallel Matrix-matrix multiplication •" Attaining the lower bound •" Attaining the lower bound •" Other Parallel Algorithms (next lecture) •" Other Parallel Algorithms (next lecture) 02/25/2016! CS267 Lecture 12! 3! 02/25/2016! CS267 Lecture 12! 4! CS267 Lecture 2 1 Motifs What is dense linear algebra? •" Not just matmul! The Motifs (formerly “Dwarfs”) from •" Linear Systems: Ax=b “The Berkeley View” (Asanovic et al.) •" Least Squares: choose x to minimize ||Ax-b||2 Motifs form key computational patterns •" Overdetermined or underdetermined; Unconstrained, constrained, or weighted •" Eigenvalues and vectors of Symmetric Matrices •" Standard (Ax = λx), Generalized (Ax=λBx) •" Eigenvalues and vectors of Unsymmetric matrices
    [Show full text]
  • Discontinuous Plane Rotations and the Symmetric Eigenvalue Problem
    December 4, 2000 Discontinuous Plane Rotations and the Symmetric Eigenvalue Problem Edward Anderson Lockheed Martin Services Inc. [email protected] Elementary plane rotations are one of the building blocks of numerical linear algebra and are employed in reducing matrices to condensed form for eigenvalue computations and during the QR algorithm. Unfortunately, their implementation in standard packages such as EISPACK, the BLAS and LAPACK lack the continuity of their mathematical formulation, which makes results from software that use them sensitive to perturbations. Test cases illustrating this problem will be presented, and reparations to the standard software proposed. 1.0 Introduction Unitary transformations are frequently used in numerical linear algebra software to reduce dense matrices to bidiagonal, tridiagonal, or Hessenberg form as a preliminary step towards finding the eigenvalues or singular values. Elementary plane rotation matrices (also called Givens rotations) are used to selectively reduce values to zero, such as when reducing a band matrix to tridiagonal form, or during the QR algorithm when attempting to deflate a tridiagonal matrix. The mathematical properties of rota- tions were studied by Wilkinson [9] and are generally not questioned. However, Wilkinson based his analysis on an equation for computing a rotation matrix that is different from those in common use in the BLAS [7] and LAPACK [2]. While the classical formula is continuous in the angle of rotation, the BLAS and LAPACK algorithms have lines of discontinuity that can lead to abrupt changes in the direction of computed eigenvectors. This makes comparing results difficult and proving backward stability impossible. It is easy to restore continuity in the generation of Givens rotations even while scaling to avoid overflow and underflow, and we show how in this report.
    [Show full text]
  • Comparison of Numerical Methods and Open-Source Libraries for Eigenvalue Analysis of Large-Scale Power Systems
    applied sciences Article Comparison of Numerical Methods and Open-Source Libraries for Eigenvalue Analysis of Large-Scale Power Systems Georgios Tzounas , Ioannis Dassios * , Muyang Liu and Federico Milano School of Electrical and Electronic Engineering, University College Dublin, Belfield, Dublin 4, Ireland; [email protected] (G.T.); [email protected] (M.L.); [email protected] (F.M.) * Correspondence: [email protected] Received: 30 September 2020; Accepted: 24 October 2020; Published: 28 October 2020 Abstract: This paper discusses the numerical solution of the generalized non-Hermitian eigenvalue problem. It provides a comprehensive comparison of existing algorithms, as well as of available free and open-source software tools, which are suitable for the solution of the eigenvalue problems that arise in the stability analysis of electric power systems. The paper focuses, in particular, on methods and software libraries that are able to handle the large-scale, non-symmetric matrices that arise in power system eigenvalue problems. These kinds of eigenvalue problems are particularly difficult for most numerical methods to handle. Thus, a review and fair comparison of existing algorithms and software tools is a valuable contribution for researchers and practitioners that are interested in power system dynamic analysis. The scalability and performance of the algorithms and libraries are duly discussed through case studies based on real-world electrical power networks. These are a model of the All-Island Irish Transmission System with 8640 variables; and, a model of the European Network of Transmission System Operators for Electricity, with 146,164 variables. Keywords: eigenvalue analysis; large non-Hermitian matrices; numerical methods; open-source libraries 1.
    [Show full text]