RELIABLE GENERATION OF HIGH- PERFORMANCE MATRIX ALGEBRA
|
|
- Arron Wilkinson
- 5 years ago
- Views:
Transcription
1 RELIABLE GENERATION OF HIGH- PERFORMANCE MATRIX ALGEBRA Jeremy G. Siek, University of Colorado at Boulder Joint work with Liz Jessup, Boyana Norris, Geoffrey Belter, Thomas Nelson Lake Isabelle, Colorado Our SC 12 submission is available on my web page.
2 ABSTRACTION VS. PERFORMANCE A few lines from the PETSc Biconjugate gradient method. 97: betaold = beta; 98: KSP_MatMult(ksp,Amat,Pr,Zr); /* z <- Kp */ 99: VecConjugate(Pl); 100: KSP_MatMultTranspose(ksp,Amat,Pl,Zl); 101: VecConjugate(Pl); 102: VecConjugate(Zl); 103: VecDot(Zr,Pl,&dpi); /* dpi <- z'p */ 104: a = beta/dpi; /* a = beta/p'z */ 105: VecAXPY(X,a,Pr); /* x <- x + ap */
3 LARGER SCOPE = BETTER PERFORMANCE BLAS 3 BLAS 2.5 Optimization Opportunities BLAS 2 BLAS 1 Scope
4 LARGER SCOPE = BETTER PERFORMANCE 1/1 compute to data BLAS 3 BLAS 2.5 Optimization Opportunities BLAS 2 BLAS 1 Scope
5 LARGER SCOPE = BETTER PERFORMANCE 1/1 compute to data BLAS 3 BLAS 2.5 Optimization Opportunities BLAS 1 BLAS 2 But larger scope = more kernels = higher cost Scope
6 COST & PERFORMANCE BLAS, hand-tuned Man hours/kernel BLAS, auto-tuned General Purpose Compilers Domain-specific compilers Reliable Performance
7 BUILD TO ORDER BLAS Kernel Specification A = A + u1 * v1' + u2 * v2' x = beta * (A' * y) + z w = alpha * (A * x) Build to Order BLAS Compiler Optimized Kernel Implementation in C
8 BUILD TO ORDER BLAS Linear Algebra Specific Knowledge Bases General Purpose Infrastructure Kernel Specification Linear Algebra Implementation Dataflow Transformation Engine Dataflow Graph Linear Algebra Optimizations Analytic Performance Predictor Stack of Alternatives Hardware Database C++/ C++/ C Fortran Fortran Empirical Performance Evaluation C
9 KERNEL SPECIFICATION GEMVER in u1 : vector, u2 : vector, v1 : vector, v2 : vector, alpha : scalar, beta : scalar, y : vector, z : vector inout A : dense column matrix out x : vector, w : vector { A = A + u1 * v1' + u2 * v2' x = beta * (A' * y) + z w = alpha * (A * x) }
10 Speed Up Relative to ICC PGI BLAS (MKL) HAND BTO PLUTO AXPYDOT VADD WAXPBY Kernel AXPYDOT VADD WAXPBY ATAX BICGK DGEMV DGEMVT GEMVER GESUMMV Operation z w αv β z T u x w + y + z w αx + βy y A T Ax q Ap s A T r z αax + βy x βa T y + z w αax B A + u 1 v1 T + u 2v2 T x βb T y + z w αbx y αax + βbx Intel Westmere, 24 core, 2.66 GHz ATAX BICGK DGEMV DGEMVT GESUMMV GEMVER
11 RESULTS ON AMD Speedups relative to PGI Kernel BLAS Pluto HAND BTO AXPYDOT VADD WAXPBY ATAX BICGK DGEMV DGEMVT GEMVER GESUMMV AMD Phenom, 6 core, 3.3 GHz Kernel BLAS Pluto HAND BTO AXPYDOT VADD WAXPBY ATAX BICGK DGEMV DGEMVT GEMVER GESUMMV AMD Interlagos, 64 core, 2.2 GHz
12 BUILD TO ORDER BLAS Linear Algebra Specific Knowledge Bases General Purpose Infrastructure Kernel Specification Linear Algebra Implementation Dataflow Transformation Engine Dataflow Graph Linear Algebra Optimizations Analytic Performance Predictor Stack of Alternatives Hardware Database C++/ C++/ Fortran Fortran Empirical Performance Evaluation C++
13 DATAFLOW GRAPH A = A + u1 * v1' + u2 * v2' x = beta * (A' * y) + z w = alpha * (A * x) v2 T x A y alpha z beta u2 + + T x x + x x v1 T x u1 x w
14 TRAVERSAL PATTERNS orientations O ::= C R types τ ::= O τ S. R<S> R<R<S>> C<R<S>>
15 LINEAR ALGEBRA DB Algo Op and Operands Result Type Pipe O τ l + O τ r O τ l + τ r yes S + S S no O τ T O T τ T yes S S S no R τ l R τ r R R τ l τ r yes C τ l C τ r C P τ l C τ r yes R τ l C τ r (τl τ r ) no C τ l R τ r C τ l R τ r yes C τ l R τ r R C τ l τ r yes S O τ O S τ yes Table 1: Sample of the linear algebra knowledge base.
16 C TYPE/ALGO INFERENCE R<> v2 T x A y alpha z beta u2 + + T x x + x x v1 T x u1 x w C
17 C TYPE/ALGO INFERENCE R<> v2 T x A y alpha z beta u2 + + T x x + x x v1 T add u1 x R<> R<> x w C
18 C TYPE/ALGO INFERENCE R<> R<> v2 u2 T x add A y alpha z beta + + T x x + x x v1 T add u1 x R<> R<> x w R<> C
19 C TYPE/ALGO INFERENCE R<S> R<> R<> v2 u2 T x outer2 add A y alpha z beta v1 T outer2 + + T x x + x x add u1 R<S> x R<> R<> R<> x w C
20 BUILD TO ORDER BLAS Linear Algebra Specific Knowledge Bases General Purpose Infrastructure Kernel Specification Linear Algebra Implementation Dataflow Transformation Engine Dataflow Graph Linear Algebra Optimizations Analytic Performance Predictor Stack of Alternatives Hardware Database C++/ C++/ Fortran Fortran Empirical Performance Evaluation C++
21 DATAFLOW REFINEMENT j = 1..n Column Matrix R<> A Column Matrix R<> Column Matrix R<> A A(:,j) Column Vector Column Matrix R<> + C + C(:,j)<- C B B B(:,j) Column Matrix R<> Column Matrix R<> j = 1..n Column Matrix R<> A A(:,j) i = 1..m A(i,j) Column Vector Scalar S Column Matrix R<> + C(i,j)<- C(:,j)<- C B Column Matrix R<> B(:,j) B(i,j)
22 OPTIMIZATION FUSION (1) A A i = 1..n i = 1..n i = 1..n A(i) A(i) A(i)
23 OPTIMIZATION FUSION (2) i = 1..n i = 1..n A(i)<- i = 1..n A A(i)
24 BUILD TO ORDER BLAS Linear Algebra Specific Knowledge Bases General Purpose Infrastructure Kernel Specification Linear Algebra Implementation Dataflow Transformation Engine Dataflow Graph Linear Algebra Optimizations Analytic Performance Predictor Stack of Alternatives Hardware Database C++/ C++/ Fortran Fortran Empirical Performance Evaluation C++
25 SEARCH SPACE Illegal Legal Complete Search Space BTO Considered Search Space BTO Legal Points BTO Pruned
26 ENUMERATING THE SPACE We try to avoid even considering illegal points Loop fusion is an equivalence relation Can t fuse inner loops if you haven t already fused their outer loops. y βa T Ax. a : {{1}}{{2}}{{3}} b : {{1}{2}}{{3}} c : {{12}}{{3}} d : {{1}{3}}{{2}} e : {{1}{2}{3}} f : {{123}} a : b : for i in 1 to M for j in 1 to N t0[ i ] += A[i, j ] x[j ] for i in 1 to M for j in 1 to N t1[ j ] += A[i, j ] t0[ i ] for j in 1 to N y[ j ] = t1[ j ] beta
27 PARTITIONING for i in 1 to M for j in 1 to N t0[ i ] += A[i, j ] x[j ] for i in 1 to M rix or ve { i { j 1}} }, where A A { p(i) { { }}} { p(i) { i { j 1}}} { p(j) { i { j 1}}} { { { }}}
28 MFGA SEARCH ALGORITHM We start with a greedy search technique that we call max-fuse (MF). Then we mutate to seed a genetic algorithm (GA). add or remove fusions add or remove partitions change direction of partition (horizontal/vertical) increment/decrement number of threads assigned to a partition
29 SEARCH TIME VS. PERFORMANCE For GEMVER on Intel Westmere, 24 core
30 FUTURE WORK/ CONCLUSIONS We obtain reliable, high-performance matrix algebra 1. high-level specification language 2. careful enumeration of optimization choices 3. search algorithm: max-fuse + genetic Future work: More parallelism using MPI, GPUs More matrix formats: banded, triangular, sparse
A Reliable Generation of High-Performance Matrix Algebra
A Reliable Generation of High-Performance Matrix Algebra THOMAS NELSON, University of Colorado, Boulder GEOFFREY BELTER, University of Colorado, Boulder JEREMY G. SIEK, University of Colorado, Boulder
More informationAutomating the Generation of Composed Linear Algebra Kernels
utomating the Generation of Composed Linear lgebra Kernels Geoffrey Belter geoffrey.belter@colorado.edu Dept. of ECEE, University of Colorado E. R. Jessup jessup@cs.colorado.edu Dept. of Computer Science,
More informationDSLs and Search for Linear Algebra Performance Optimization
University of Colorado, Boulder CU Scholar Computer Science Graduate Theses & Dissertations Computer Science Spring 1-1-2015 DSLs and Search for Linear Algebra Performance Optimization Thomas Harrison
More informationHIGH PERFORMANCE NUMERICAL LINEAR ALGEBRA. Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA
1 HIGH PERFORMANCE NUMERICAL LINEAR ALGEBRA Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA 2 BLAS BLAS 1, 2, 3 Performance GEMM Optimized BLAS Parallel
More informationBLAS. Basic Linear Algebra Subprograms
BLAS Basic opera+ons with vectors and matrices dominates scien+fic compu+ng programs To achieve high efficiency and clean computer programs an effort has been made in the last few decades to standardize
More informationRuntime Prediction of Fused Linear Algebra in a Compiler Framework
University of Colorado, Boulder CU Scholar Computer Science Graduate Theses & Dissertations Computer Science Spring 1-1-2011 Runtime Prediction of Fused Linear Algebra in a Compiler Framework Ian Karlin
More informationAutotuning. John Cavazos. University of Delaware UNIVERSITY OF DELAWARE COMPUTER & INFORMATION SCIENCES DEPARTMENT
Autotuning John Cavazos University of Delaware What is Autotuning? Searching for the best code parameters, code transformations, system configuration settings, etc. Search can be Quasi-intelligent: genetic
More informationAdministrative Issues. L11: Sparse Linear Algebra on GPUs. Triangular Solve (STRSM) A Few Details 2/25/11. Next assignment, triangular solve
Administrative Issues L11: Sparse Linear Algebra on GPUs Next assignment, triangular solve Due 5PM, Tuesday, March 15 handin cs6963 lab 3 Project proposals Due 5PM, Wednesday, March 7 (hard
More informationModule 16: Data Flow Analysis in Presence of Procedure Calls Lecture 32: Iteration. The Lecture Contains: Iteration Space.
The Lecture Contains: Iteration Space Iteration Vector Normalized Iteration Vector Dependence Distance Direction Vector Loop Carried Dependence Relations Dependence Level Iteration Vector - Triangular
More informationBLAS: Basic Linear Algebra Subroutines I
BLAS: Basic Linear Algebra Subroutines I Most numerical programs do similar operations 90% time is at 10% of the code If these 10% of the code is optimized, programs will be fast Frequently used subroutines
More informationAccelerating GPU Kernels for Dense Linear Algebra
Accelerating GPU Kernels for Dense Linear Algebra Rajib Nath, Stan Tomov, and Jack Dongarra Innovative Computing Lab University of Tennessee, Knoxville July 9, 21 xgemm performance of CUBLAS-2.3 on GTX28
More informationLinear Algebra for Modern Computers. Jack Dongarra
Linear Algebra for Modern Computers Jack Dongarra Tuning for Caches 1. Preserve locality. 2. Reduce cache thrashing. 3. Loop blocking when out of cache. 4. Software pipelining. 2 Indirect Addressing d
More informationBLAS: Basic Linear Algebra Subroutines I
BLAS: Basic Linear Algebra Subroutines I Most numerical programs do similar operations 90% time is at 10% of the code If these 10% of the code is optimized, programs will be fast Frequently used subroutines
More informationScheduling of QR Factorization Algorithms on SMP and Multi-core Architectures
Scheduling of Algorithms on SMP and Multi-core Architectures Gregorio Quintana-Ortí Enrique S. Quintana-Ortí Ernie Chan Robert A. van de Geijn Field G. Van Zee quintana@icc.uji.es Universidad Jaime I de
More informationLinear Algebra libraries in Debian. DebConf 10 New York 05/08/2010 Sylvestre
Linear Algebra libraries in Debian Who I am? Core developer of Scilab (daily job) Debian Developer Involved in Debian mainly in Science and Java aspects sylvestre.ledru@scilab.org / sylvestre@debian.org
More informationGPU ACCELERATION OF WSMP (WATSON SPARSE MATRIX PACKAGE)
GPU ACCELERATION OF WSMP (WATSON SPARSE MATRIX PACKAGE) NATALIA GIMELSHEIN ANSHUL GUPTA STEVE RENNICH SEID KORIC NVIDIA IBM NVIDIA NCSA WATSON SPARSE MATRIX PACKAGE (WSMP) Cholesky, LDL T, LU factorization
More informationParallelisation. Michael O Boyle. March 2014
Parallelisation Michael O Boyle March 2014 1 Lecture Overview Parallelisation for fork/join Mapping parallelism to shared memory multi-processors Loop distribution and fusion Data Partitioning and SPMD
More informationMAGMA a New Generation of Linear Algebra Libraries for GPU and Multicore Architectures
MAGMA a New Generation of Linear Algebra Libraries for GPU and Multicore Architectures Stan Tomov Innovative Computing Laboratory University of Tennessee, Knoxville OLCF Seminar Series, ORNL June 16, 2010
More informationAutomatic Performance Tuning. Jeremy Johnson Dept. of Computer Science Drexel University
Automatic Performance Tuning Jeremy Johnson Dept. of Computer Science Drexel University Outline Scientific Computation Kernels Matrix Multiplication Fast Fourier Transform (FFT) Automated Performance Tuning
More informationBehavioral Data Mining. Lecture 12 Machine Biology
Behavioral Data Mining Lecture 12 Machine Biology Outline CPU geography Mass storage Buses and Networks Main memory Design Principles Intel i7 close-up From Computer Architecture a Quantitative Approach
More informationIntel C++ Compiler User's Guide With Support For The Streaming Simd Extensions 2
Intel C++ Compiler User's Guide With Support For The Streaming Simd Extensions 2 This release of the Intel C++ Compiler 16.0 product is a Pre-Release, and as such is 64 architecture processor supporting
More information9. Linear Algebra Computation
9. Linear Algebra Computation Basic Linear Algebra Subprograms (BLAS) Routines that provide standard, low-level, building blocks for performing basic vector and matrix operations. Originally developed
More informationStudy and implementation of computational methods for Differential Equations in heterogeneous systems. Asimina Vouronikoy - Eleni Zisiou
Study and implementation of computational methods for Differential Equations in heterogeneous systems Asimina Vouronikoy - Eleni Zisiou Outline Introduction Review of related work Cyclic Reduction Algorithm
More informationOptimizing Cache Performance in Matrix Multiplication. UCSB CS240A, 2017 Modified from Demmel/Yelick s slides
Optimizing Cache Performance in Matrix Multiplication UCSB CS240A, 2017 Modified from Demmel/Yelick s slides 1 Case Study with Matrix Multiplication An important kernel in many problems Optimization ideas
More informationDynamic Selection of Auto-tuned Kernels to the Numerical Libraries in the DOE ACTS Collection
Numerical Libraries in the DOE ACTS Collection The DOE ACTS Collection SIAM Parallel Processing for Scientific Computing, Savannah, Georgia Feb 15, 2012 Tony Drummond Computational Research Division Lawrence
More informationLoop Transformations! Part II!
Lecture 9! Loop Transformations! Part II! John Cavazos! Dept of Computer & Information Sciences! University of Delaware! www.cis.udel.edu/~cavazos/cisc879! Loop Unswitching Hoist invariant control-flow
More informationIntel Math Kernel Library (Intel MKL) BLAS. Victor Kostin Intel MKL Dense Solvers team manager
Intel Math Kernel Library (Intel MKL) BLAS Victor Kostin Intel MKL Dense Solvers team manager Intel MKL BLAS/Sparse BLAS Original ( dense ) BLAS available from www.netlib.org Additionally Intel MKL provides
More informationA Parallelizing Compiler for Multicore Systems
A Parallelizing Compiler for Multicore Systems José M. Andión, Manuel Arenaz, Gabriel Rodríguez and Juan Touriño 17th International Workshop on Software and Compilers for Embedded Systems (SCOPES 2014)
More informationScientific Computing. Some slides from James Lambers, Stanford
Scientific Computing Some slides from James Lambers, Stanford Dense Linear Algebra Scaling and sums Transpose Rank-one updates Rotations Matrix vector products Matrix Matrix products BLAS Designing Numerical
More informationA polyhedral loop transformation framework for parallelization and tuning
A polyhedral loop transformation framework for parallelization and tuning Ohio State University Uday Bondhugula, Muthu Baskaran, Albert Hartono, Sriram Krishnamoorthy, P. Sadayappan Argonne National Laboratory
More informationBrief notes on setting up semi-high performance computing environments. July 25, 2014
Brief notes on setting up semi-high performance computing environments July 25, 2014 1 We have two different computing environments for fitting demanding models to large space and/or time data sets. 1
More informationOvercoming the Barriers to Sustained Petaflop Performance. William D. Gropp Mathematics and Computer Science
Overcoming the Barriers to Sustained Petaflop Performance William D. Gropp Mathematics and Computer Science www.mcs.anl.gov/~gropp But First Are we too CPU-centric? What about I/O? What do applications
More information' $ Sparse Compilation & % 1
Sparse Compilation 1 Motivation for Sparse Codes Consider ux, heat, or stresses interactions are between neighbors. Linear equations are sparse. Therefore, matrices are sparse. 2 Three Sparse Matrix Representations
More informationSome notes on efficient computing and high performance computing environments
Some notes on efficient computing and high performance computing environments Abhi Datta 1, Sudipto Banerjee 2 and Andrew O. Finley 3 July 31, 2017 1 Department of Biostatistics, Bloomberg School of Public
More informationResources for parallel computing
Resources for parallel computing BLAS Basic linear algebra subprograms. Originally published in ACM Toms (1979) (Linpack Blas + Lapack). Implement matrix operations upto matrix-matrix multiplication and
More informationSpeedup Altair RADIOSS Solvers Using NVIDIA GPU
Innovation Intelligence Speedup Altair RADIOSS Solvers Using NVIDIA GPU Eric LEQUINIOU, HPC Director Hongwei Zhou, Senior Software Developer May 16, 2012 Innovation Intelligence ALTAIR OVERVIEW Altair
More informationPorting the NAS-NPB Conjugate Gradient Benchmark to CUDA. NVIDIA Corporation
Porting the NAS-NPB Conjugate Gradient Benchmark to CUDA NVIDIA Corporation Outline! Overview of CG benchmark! Overview of CUDA Libraries! CUSPARSE! CUBLAS! Porting Sequence! Algorithm Analysis! Data/Code
More informationLocality-Aware Automatic Parallelization for GPGPU with OpenHMPP Directives
Locality-Aware Automatic Parallelization for GPGPU with OpenHMPP Directives José M. Andión, Manuel Arenaz, François Bodin, Gabriel Rodríguez and Juan Touriño 7th International Symposium on High-Level Parallel
More informationApplications of Berkeley s Dwarfs on Nvidia GPUs
Applications of Berkeley s Dwarfs on Nvidia GPUs Seminar: Topics in High-Performance and Scientific Computing Team N2: Yang Zhang, Haiqing Wang 05.02.2015 Overview CUDA The Dwarfs Dynamic Programming Sparse
More informationIssues In Implementing The Primal-Dual Method for SDP. Brian Borchers Department of Mathematics New Mexico Tech Socorro, NM
Issues In Implementing The Primal-Dual Method for SDP Brian Borchers Department of Mathematics New Mexico Tech Socorro, NM 87801 borchers@nmt.edu Outline 1. Cache and shared memory parallel computing concepts.
More informationOptimizing the operations with sparse matrices on Intel architecture
Optimizing the operations with sparse matrices on Intel architecture Gladkikh V. S. victor.s.gladkikh@intel.com Intel Xeon, Intel Itanium are trademarks of Intel Corporation in the U.S. and other countries.
More informationMatrix Multiplication
Matrix Multiplication CPS343 Parallel and High Performance Computing Spring 2013 CPS343 (Parallel and HPC) Matrix Multiplication Spring 2013 1 / 32 Outline 1 Matrix operations Importance Dense and sparse
More informationBLAS. Christoph Ortner Stef Salvini
BLAS Christoph Ortner Stef Salvini The BLASics Basic Linear Algebra Subroutines Building blocks for more complex computations Very widely used Level means number of operations Level 1: vector-vector operations
More informationMore Data Locality for Static Control Programs on NUMA Architectures
More Data Locality for Static Control Programs on NUMA Architectures Adilla Susungi 1, Albert Cohen 2, Claude Tadonki 1 1 MINES ParisTech, PSL Research University 2 Inria and DI, Ecole Normale Supérieure
More informationTransforming Imperfectly Nested Loops
Transforming Imperfectly Nested Loops 1 Classes of loop transformations: Iteration re-numbering: (eg) loop interchange Example DO 10 J = 1,100 DO 10 I = 1,100 DO 10 I = 1,100 vs DO 10 J = 1,100 Y(I) =
More information*Yuta SAWA and Reiji SUDA The University of Tokyo
Auto Tuning Method for Deciding Block Size Parameters in Dynamically Load-Balanced BLAS *Yuta SAWA and Reiji SUDA The University of Tokyo iwapt 29 October 1-2 *Now in Central Research Laboratory, Hitachi,
More informationJose Aliaga (Universitat Jaume I, Castellon, Spain), Ruyman Reyes, Mehdi Goli (Codeplay Software) 2017 Codeplay Software Ltd.
SYCL-BLAS: LeveragingSYCL-BLAS Expression Trees for Linear Algebra Jose Aliaga (Universitat Jaume I, Castellon, Spain), Ruyman Reyes, Mehdi Goli (Codeplay Software) 1 About me... Phd in Compilers and Parallel
More informationMatrix Multiplication
Matrix Multiplication CPS343 Parallel and High Performance Computing Spring 2018 CPS343 (Parallel and HPC) Matrix Multiplication Spring 2018 1 / 32 Outline 1 Matrix operations Importance Dense and sparse
More informationAdvanced School in High Performance and GRID Computing November Mathematical Libraries. Part I
1967-10 Advanced School in High Performance and GRID Computing 3-14 November 2008 Mathematical Libraries. Part I KOHLMEYER Axel University of Pennsylvania Department of Chemistry 231 South 34th Street
More informationIntel Math Kernel Library
Intel Math Kernel Library Release 7.0 March 2005 Intel MKL Purpose Performance, performance, performance! Intel s scientific and engineering floating point math library Initially only basic linear algebra
More informationAchieve Better Performance with PEAK on XSEDE Resources
Achieve Better Performance with PEAK on XSEDE Resources Haihang You, Bilel Hadri, Shirley Moore XSEDE 12 July 18 th 2012 Motivations FACTS ALTD ( Automatic Tracking Library Database ) ref Fahey, Jones,
More informationOn Level Scheduling for Incomplete LU Factorization Preconditioners on Accelerators
On Level Scheduling for Incomplete LU Factorization Preconditioners on Accelerators Karl Rupp, Barry Smith rupp@mcs.anl.gov Mathematics and Computer Science Division Argonne National Laboratory FEMTEC
More informationCray Scientific Libraries. Overview
Cray Scientific Libraries Overview What are libraries for? Building blocks for writing scientific applications Historically allowed the first forms of code re-use Later became ways of running optimized
More informationSoftware Distributed Shared Memory with High Bandwidth Network: Production and Evaluation
,,.,, InfiniBand PCI Express,,,. Software Distributed Shared Memory with High Bandwidth Network: Production and Evaluation Akira Nishida, The recent development of commodity hardware technologies makes
More informationLattice Simulations using OpenACC compilers. Pushan Majumdar (Indian Association for the Cultivation of Science, Kolkata)
Lattice Simulations using OpenACC compilers Pushan Majumdar (Indian Association for the Cultivation of Science, Kolkata) OpenACC is a programming standard for parallel computing developed by Cray, CAPS,
More informationPorting Scientific Research Codes to GPUs with CUDA Fortran: Incompressible Fluid Dynamics using the Immersed Boundary Method
Porting Scientific Research Codes to GPUs with CUDA Fortran: Incompressible Fluid Dynamics using the Immersed Boundary Method Josh Romero, Massimiliano Fatica - NVIDIA Vamsi Spandan, Roberto Verzicco -
More informationCray Scientific Libraries: Overview and Performance. Cray XE6 Performance Workshop University of Reading Nov 2012
Cray Scientific Libraries: Overview and Performance Cray XE6 Performance Workshop University of Reading 20-22 Nov 2012 Contents LibSci overview and usage BFRAME / CrayBLAS LAPACK ScaLAPACK FFTW / CRAFFT
More informationOpenACC Fundamentals. Steve Abbott November 13, 2016
OpenACC Fundamentals Steve Abbott , November 13, 2016 Who Am I? 2005 B.S. Physics Beloit College 2007 M.S. Physics University of Florida 2015 Ph.D. Physics University of New Hampshire
More informationMatrix-free multi-gpu Implementation of Elliptic Solvers for strongly anisotropic PDEs
Iterative Solvers Numerical Results Conclusion and outlook 1/18 Matrix-free multi-gpu Implementation of Elliptic Solvers for strongly anisotropic PDEs Eike Hermann Müller, Robert Scheichl, Eero Vainikko
More informationBig Data Analytics Performance for Large Out-Of- Core Matrix Solvers on Advanced Hybrid Architectures
Procedia Computer Science Volume 51, 2015, Pages 2774 2778 ICCS 2015 International Conference On Computational Science Big Data Analytics Performance for Large Out-Of- Core Matrix Solvers on Advanced Hybrid
More informationParallel Programming in C with MPI and OpenMP
Parallel Programming in C with MPI and OpenMP Michael J. Quinn Chapter 17 Shared-memory Programming 1 Outline n OpenMP n Shared-memory model n Parallel for loops n Declaring private variables n Critical
More informationToward a supernodal sparse direct solver over DAG runtimes
Toward a supernodal sparse direct solver over DAG runtimes HOSCAR 2013, Bordeaux X. Lacoste Xavier LACOSTE HiePACS team Inria Bordeaux Sud-Ouest November 27, 2012 Guideline Context and goals About PaStiX
More informationAdvanced Numerical Techniques for Cluster Computing
Advanced Numerical Techniques for Cluster Computing Presented by Piotr Luszczek http://icl.cs.utk.edu/iter-ref/ Presentation Outline Motivation hardware Dense matrix calculations Sparse direct solvers
More informationqr_mumps a multithreaded, multifrontal QR solver The MUMPS team,
qr_mumps a multithreaded, multifrontal QR solver The MUMPS team, MUMPS Users Group Meeting, May 29-30, 2013 The qr_mumps software the qr_mumps software qr_mumps version 1.0 a multithreaded, multifrontal
More informationAMath 483/583 Lecture 22. Notes: Another Send/Receive example. Notes: Notes: Another Send/Receive example. Outline:
AMath 483/583 Lecture 22 Outline: MPI Master Worker paradigm Linear algebra LAPACK and the BLAS References: $UWHPSC/codes/mpi class notes: MPI section class notes: Linear algebra Another Send/Receive example
More informationCMSC 714 Lecture 6 MPI vs. OpenMP and OpenACC. Guest Lecturer: Sukhyun Song (original slides by Alan Sussman)
CMSC 714 Lecture 6 MPI vs. OpenMP and OpenACC Guest Lecturer: Sukhyun Song (original slides by Alan Sussman) Parallel Programming with Message Passing and Directives 2 MPI + OpenMP Some applications can
More informationOpenMP. A parallel language standard that support both data and functional Parallelism on a shared memory system
OpenMP A parallel language standard that support both data and functional Parallelism on a shared memory system Use by system programmers more than application programmers Considered a low level primitives
More informationDense matrix algebra and libraries (and dealing with Fortran)
Dense matrix algebra and libraries (and dealing with Fortran) CPS343 Parallel and High Performance Computing Spring 2018 CPS343 (Parallel and HPC) Dense matrix algebra and libraries (and dealing with Fortran)
More informationGPU Implementation of Elliptic Solvers in NWP. Numerical Weather- and Climate- Prediction
1/8 GPU Implementation of Elliptic Solvers in Numerical Weather- and Climate- Prediction Eike Hermann Müller, Robert Scheichl Department of Mathematical Sciences EHM, Xu Guo, Sinan Shi and RS: http://arxiv.org/abs/1302.7193
More informationLecture 9 Basic Parallelization
Lecture 9 Basic Parallelization I. Introduction II. Data Dependence Analysis III. Loop Nests + Locality IV. Interprocedural Parallelization Chapter 11.1-11.1.4 CS243: Parallelization 1 Machine Learning
More informationLecture 9 Basic Parallelization
Lecture 9 Basic Parallelization I. Introduction II. Data Dependence Analysis III. Loop Nests + Locality IV. Interprocedural Parallelization Chapter 11.1-11.1.4 CS243: Parallelization 1 Machine Learning
More informationA Few Numerical Libraries for HPC
A Few Numerical Libraries for HPC CPS343 Parallel and High Performance Computing Spring 2016 CPS343 (Parallel and HPC) A Few Numerical Libraries for HPC Spring 2016 1 / 37 Outline 1 HPC == numerical linear
More informationAccelerating Data Warehousing Applications Using General Purpose GPUs
Accelerating Data Warehousing Applications Using General Purpose s Sponsors: Na%onal Science Founda%on, LogicBlox Inc., IBM, and NVIDIA The General Purpose is a many core co-processor 10s to 100s of cores
More informationProgress on GPU Parallelization of the NIM Prototype Numerical Weather Prediction Dynamical Core
Progress on GPU Parallelization of the NIM Prototype Numerical Weather Prediction Dynamical Core Tom Henderson NOAA/OAR/ESRL/GSD/ACE Thomas.B.Henderson@noaa.gov Mark Govett, Jacques Middlecoff Paul Madden,
More informationParallel Numerics. 1 Data Dependency Graphs & DAGs. Exercise 3: Vector/Vector Operations & P2P Communication II
Technische Universität München WiSe 2014/15 Institut für Informatik Prof. Dr. Thomas Huckle Dipl.-Inf. Christoph Riesinger Sebastian Rettenberger, M.Sc. Parallel Numerics Exercise 3: Vector/Vector Operations
More informationSequoia. Mike Houston Stanford University July 9 th, CScADS Workshop
Sequoia Mike Houston Stanford University July 9 th, 2007 - CScADS Workshop Today s outline Sequoia programming model Sequoia targets Tuning in Sequoia http://sequoia.stanford.edu - Supercomputing 2006
More informationOn the limits of (and opportunities for?) GPU acceleration
On the limits of (and opportunities for?) GPU acceleration Aparna Chandramowlishwaran, Jee Choi, Kenneth Czechowski, Murat (Efe) Guney, Logan Moon, Aashay Shringarpure, Richard (Rich) Vuduc HotPar 10,
More informationAutomatic Tuning of Sparse Matrix Kernels
Automatic Tuning of Sparse Matrix Kernels Kathy Yelick U.C. Berkeley and Lawrence Berkeley National Laboratory Richard Vuduc, Lawrence Livermore National Laboratory James Demmel, U.C. Berkeley Berkeley
More informationCombined Iterative and Model-driven Optimization in an Automatic Parallelization Framework
Combined Iterative and Model-driven Optimization in an Automatic Parallelization Framework Louis-Noël Pouchet The Ohio State University pouchet@cse.ohio-state.edu Uday Bondhugula IBM T.J. Watson Research
More informationOil and Water Can Mix: An Integration of Polyhedral and AST-based Transformations
Oil and Water Can Mix: An Integration of Polyhedral and AST-based Transformations Jun Shirako Rice University Louis-Noël Pouchet University of California Los Angeles Vivek Sarkar Rice University Abstract
More informationSolving Dense Linear Systems on Platforms with Multiple Hardware Accelerators
Solving Dense Linear Systems on Platforms with Multiple Hardware Accelerators Francisco D. Igual Enrique S. Quintana-Ortí Gregorio Quintana-Ortí Universidad Jaime I de Castellón (Spain) Robert A. van de
More informationCS4961 Parallel Programming. Lecture 9: Task Parallelism in OpenMP 9/22/09. Administrative. Mary Hall September 22, 2009.
Parallel Programming Lecture 9: Task Parallelism in OpenMP Administrative Programming assignment 1 is posted (after class) Due, Tuesday, September 22 before class - Use the handin program on the CADE machines
More informationMPI Programming. Henrik R. Nagel Scientific Computing IT Division
1 MPI Programming Henrik R. Nagel Scientific Computing IT Division 2 Outline Introduction Finite Difference Method Finite Element Method LU Factorization SOR Method Monte Carlo Method Molecular Dynamics
More informationQR Decomposition on GPUs
QR Decomposition QR Algorithms Block Householder QR Andrew Kerr* 1 Dan Campbell 1 Mark Richards 2 1 Georgia Tech Research Institute 2 School of Electrical and Computer Engineering Georgia Institute of
More informationIntroduction to parallel Computing
Introduction to parallel Computing VI-SEEM Training Paschalis Paschalis Korosoglou Korosoglou (pkoro@.gr) (pkoro@.gr) Outline Serial vs Parallel programming Hardware trends Why HPC matters HPC Concepts
More informationINTRODUCTION TO OPENACC. Analyzing and Parallelizing with OpenACC, Feb 22, 2017
INTRODUCTION TO OPENACC Analyzing and Parallelizing with OpenACC, Feb 22, 2017 Objective: Enable you to to accelerate your applications with OpenACC. 2 Today s Objectives Understand what OpenACC is and
More informationEfficient multigrid solvers for strongly anisotropic PDEs in atmospheric modelling
Iterative Solvers Numerical Results Conclusion and outlook 1/22 Efficient multigrid solvers for strongly anisotropic PDEs in atmospheric modelling Part II: GPU Implementation and Scaling on Titan Eike
More informationSolving Dense Linear Systems on Graphics Processors
Solving Dense Linear Systems on Graphics Processors Sergio Barrachina Maribel Castillo Francisco Igual Rafael Mayo Enrique S. Quintana-Ortí High Performance Computing & Architectures Group Universidad
More informationAdrian Tate XK6 / openacc workshop Manno, Mar
Adrian Tate XK6 / openacc workshop Manno, Mar6-7 2012 1 Overview & Philosophy Two modes of usage Contents Present contents Upcoming releases Optimization of libsci_acc Autotuning Adaptation Asynchronous
More informationEssential constraints: Data Dependences. S1: a = b + c S2: d = a * 2 S3: a = c + 2 S4: e = d + c + 2
Essential constraints: Data Dependences S1: a = b + c S2: d = a * 2 S3: a = c + 2 S4: e = d + c + 2 Essential constraints: Data Dependences S1: a = b + c S2: d = a * 2 S3: a = c + 2 S4: e = d + c + 2 S2
More informationOptimizations of BLIS Library for AMD ZEN Core
Optimizations of BLIS Library for AMD ZEN Core 1 Introduction BLIS [1] is a portable software framework for instantiating high-performance BLAS-like dense linear algebra libraries [2] The framework was
More informationG P G P U : H I G H - P E R F O R M A N C E C O M P U T I N G
Joined Advanced Student School (JASS) 2009 March 29 - April 7, 2009 St. Petersburg, Russia G P G P U : H I G H - P E R F O R M A N C E C O M P U T I N G Dmitry Puzyrev St. Petersburg State University Faculty
More informationParallel Programming in C with MPI and OpenMP
Parallel Programming in C with MPI and OpenMP Michael J. Quinn Chapter 17 Shared-memory Programming Outline OpenMP Shared-memory model Parallel for loops Declaring private variables Critical sections Reductions
More informationPerformance Models for Evaluation and Automatic Tuning of Symmetric Sparse Matrix-Vector Multiply
Performance Models for Evaluation and Automatic Tuning of Symmetric Sparse Matrix-Vector Multiply University of California, Berkeley Berkeley Benchmarking and Optimization Group (BeBOP) http://bebop.cs.berkeley.edu
More informationThe Cray Programming Environment. An Introduction
The Cray Programming Environment An Introduction Vision Cray systems are designed to be High Productivity as well as High Performance Computers The Cray Programming Environment (PE) provides a simple consistent
More informationImproving Linear Algebra Computation on NUMA platforms through auto-tuned tuned nested parallelism
Improving Linear Algebra Computation on NUMA platforms through auto-tuned tuned nested parallelism Javier Cuenca, Luis P. García, Domingo Giménez Parallel Computing Group University of Murcia, SPAIN parallelum
More informationExploiting the Performance of 32 bit Floating Point Arithmetic in Obtaining 64 bit Accuracy
Exploiting the Performance of 32 bit Floating Point Arithmetic in Obtaining 64 bit Accuracy (Revisiting Iterative Refinement for Linear Systems) Julie Langou Piotr Luszczek Alfredo Buttari Julien Langou
More information10th August Part One: Introduction to Parallel Computing
Part One: Introduction to Parallel Computing 10th August 2007 Part 1 - Contents Reasons for parallel computing Goals and limitations Criteria for High Performance Computing Overview of parallel computer
More informationPerformance Analysis of BLAS Libraries in SuperLU_DIST for SuperLU_MCDT (Multi Core Distributed) Development
Available online at www.prace-ri.eu Partnership for Advanced Computing in Europe Performance Analysis of BLAS Libraries in SuperLU_DIST for SuperLU_MCDT (Multi Core Distributed) Development M. Serdar Celebi
More informationParallelism in Spiral
Parallelism in Spiral Franz Franchetti and the Spiral team (only part shown) Electrical and Computer Engineering Carnegie Mellon University Joint work with Yevgen Voronenko Markus Püschel This work was
More information