Download Presentation.Text.Marked.Image.Marked-Computer Science-Project Presentation and more Slides Computer Science in PDF only on Docsity! Performance Evaluation of Parallel Numerical Libraries docsity.com Outline Introduction To Whom it may concern Parallel Libraries Underlying Architecture Cluster SGI Performance Metrics Project Schedule docsity.com To Whom it may Concern Top500.org provides statistics on parallel computing docsity.com Parallel Libraries Parallel libraries selected for this project PBLAS LAPACK SCALAPACK PETSc Distributed Computing Toolbox (Matlab) docsity.com Netlib.org Source of FREE mathematical libraries Collection of mathematical software, papers, and databases Primary Institutions: – University of Tennessee, Knoxville – Oak Ridge National Laboratory – University of California: Berkeley docsity.com PBLAS Parallel Basic Linear Algebra Subprograms Parallel version of BLAS Performs linear algebra operations on distributed-memory concurrent computers A major component of the ScaLAPACK library docsity.com PBLAS Three Levels Level 1 Vector-Vector Operations (swap, copy, addition, dot product) Level 2 Matrix-Vector Operations (multiply, rank-updates, outer-product) Level 3 Matrix-Matrix Operations (multiply, transpose, rank-updates) docsity.com Who Uses PBLAS?? Mainly used in Scientific and Engineering computations docsity.com
Libraries Hierarchy
ee PVM/MPL/..
docsity.com
PETSc Portable Extensible Toolkit for Scientific Computation Very popular parallel library docsity.com Who Uses PETSc?? PETSc has been successfully employed in applications in: Nano-simulations Biology/Medical Imaging and Surgery Fusion Environmental/Subsurface Flow Computational Fluid Dynamics Wave propagation and the Helmholz equation Many and many more docsity.com Cluster Cluster computing distributes the computational load to collections of similar machines Each Computer (Node) is a complete Computer Connected to each other via a high speed bus Job broken up into blocks of data and executed independently Cheap and easy to produce docsity.com SGI Super Computer SGI at PIEAS 64 processors Connected by a local high-speed computer bus 128 GB memory docsity.com Performance Evaluation Metrics To analyze the performance of a system, we need some performance evaluation metrics Speedup Efficiency Scalability docsity.com Performance Evaluation Metrics • Scalability – Ability of a system to handle growing amount of work – In parallel computing scalability measures the capability to effectively utilize an increasing number of processors – In most of the cases, scalability decreases as more number of processors are involved in computation docsity.com Challenges in performance evaluation Many tools for performance data gathering and analysis PAPI, TAU, SvPablo, Kojak, … Various interfaces, levels of automation, and approaches to information presentation User’s point of view - What do the different tools do? Which is most appropriate for a given application? - (How) can multiple tools be used in concert? - I have tons of performance data, now what? - What automatic tuning tools are available, what exactly do they do? - How hard is it to install/learn/use tool ? - Is instrumented code portable? What’s the overhead of instrumentation? How does code evolution affect the performance analysis process? docsity.com Project Work Flow Step 1 • Selection of Parallel Libraries • SCALAPACK, PETSc, PBLAS, ... Step 2 • Selection of Performance Analysis tool • PAPI, TAU, SvPablo, Kojak, … Step 3 • Performance evaluation of algorithms developed at PIEAS • Cross checking of Performance Analyzing Tool & Algorithms Step 4 • Performance Evaluation of Parallel Libraries docsity.com Project Schedule continued Semester 8 Performance Evaluation of Parallel Numerical Libraries Thesis Writing docsity.com References URL:http://top500.org Alexey Borisenko, “Performance Evaluation in Parallel Systems”, School of information Technology, University of Ottawa, Canada, 2010 URL:http://www.codeproject.com/KB/IP/ClusterC omputing.aspx URL: http://www.mcs.anl.gov/petsc/petsc- 2/documentation URL:http://www.netlib.org/scalapack/scalapack_ home.html docsity.com