## Jenkins Jenkins VisReleL1l2o1 DOXiS4_CSB-Integration_Matrix-Comb j-l1l2o1 #593  Zurück zum Projekt   Matrix Reloaded Compare environment Vorheriger Build Nachfolgender Build  Gestartet durch vorgelagertes Projekt DOXiS4_CSB-Integration_Matrix-Comb, Build 593 ursprünglich ausgelöst durch: • Gestartet durch vorgelagertes Projekt DOXiS4 CSB-Start

## +21 ## GraphBLAS: Building a C++ Matrix API for Graph Algorithms BENJAMIN BROCK, SCOTT MCMILLAN 20 21 October 24-29 ## About Us ![Image](/uploads/documents/e/7/5/a/e75ae90af61e15b82b344610c6597bfa/p2_1 Engineering Institute at permission@sei.cmu.edu. ## DM21-0916 ## This Talk Background: How and why to use matrix algebra for graphs? What are the important data structures and concepts? Prior work in the GraphBLAS interoperate with standard C++, graph library proposal? ## This Talk Background: How and why to use matrix algebra for graphs? What are the important data structures and concepts? Prior work in the GraphBLAS

support for linear algebra (pining for Fortran...) • Your options essentially were: • Write your own Matrix class and operations • Convince your boss to buy a commercial library ## A Little History: Open-Source release 1.29.0 ## • BLAS functionality ➤ Matrix and vector representations ➤ Matrix addition and subtraction ➤ Matrix multiplication • Matrix × Matrix • Matrix × Vector • Vector dot products • Scalar Scalar multiplication ➤ Norms ➤ Still on your own for linear solvers and matrix decompositions ## • Good (but not outstanding) performance • Last major improvement was in 2008 (See uBLAS FAQ's) ##

Copying data is not always bad ..... 379 34.18 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 380 34.19 Avoid string interpolation for I/O ..... 380 34.20 Optimize network Linear Algebra 1029 75.1 Special matrices 1031 75.2 Matrix factorizations 1034 75.3 Standard functions 1034 75.4 Low-level matrix operations 1116 75.5 BLAS functions 1121 75.6 LAPACK Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1216 89.2 Sparse Vector Storage ..... 1217 89.3 Sparse Vector and Matrix Constructors ..... 1217 89.4 Sparse matrix operations ..... 1218 89.5

Copying data is not always bad ..... 379 34.18 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 380 34.19 Avoid string interpolation for I/O ..... 380 34.20 Optimize network Linear Algebra 1029 75.1 Special matrices 1031 75.2 Matrix factorizations 1034 75.3 Standard functions 1034 75.4 Low-level matrix operations 1116 75.5 BLAS functions 1121 75.6 LAPACK Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1216 89.2 Sparse Vector Storage ..... 1217 89.3 Sparse Vector and Matrix Constructors ..... 1217 89.4 Sparse matrix operations ..... 1218 89.5

Copying data is not always bad ..... 379 34.18 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 380 34.19 Avoid string interpolation for I/O ..... 380 34.20 Optimize network Linear Algebra 1029 75.1 Special matrices 1031 75.2 Matrix factorizations 1034 75.3 Standard functions 1034 75.4 Low-level matrix operations 1116 75.5 BLAS functions 1121 75.6 LAPACK Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1216 89.2 Sparse Vector Storage ..... 1217 89.3 Sparse Vector and Matrix Constructors ..... 1217 89.4 Sparse matrix operations ..... 1218 89.5

Copying data is not always bad ..... 379 34.18 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 380 34.19 Avoid string interpolation for I/O ..... 380 34.20 Optimize network Linear Algebra 1029 75.1 Special matrices 1031 75.2 Matrix factorizations 1034 75.3 Standard functions 1034 75.4 Low-level matrix operations 1116 75.5 BLAS functions 1121 75.6 LAPACK Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1216 89.2 Sparse Vector Storage ..... 1217 89.3 Sparse Vector and Matrix Constructors ..... 1217 89.4 Sparse matrix operations ..... 1218 89.5

Copying data is not always bad ..... 388 34.17 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 389 34.18 Avoid string interpolation for I/O ..... 389 34.19 Optimize network 1019 Matrix factorizations ..... 1020 The uniform scaling operator ..... 1020 73.2 Matrix factorizations ..... 1022 73.3 Standard functions ..... 1022 73.4 Low-level matrix operations Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1205 87.2 Sparse Vector Storage ..... 1206 87.3 Sparse Vector and Matrix Constructors ..... 1206 87.4 Sparse matrix operations ..... 1207 87.5

16 Copying data is not always bad 371 34.17 Consider StaticArrays.jl for small fixed-size vector/matrix operations 372 34.18 Avoid string interpolation for I/O 372 34.19 Optimize network I/O during 75 Linear Algebra 983 75.1 Special matrices 985 75.2 Matrix factorizations 988 75.3 Standard functions 988 75.4 Low-level matrix operations 1068 75.5 BLAS functions 1073 75.6 LAPACK functions Compressed Sparse Column (CSC) Sparse Matrix Storage 1166 89.2 Sparse Vector Storage 1167 89.3 Sparse Vector and Matrix Constructors 1167 89.4 Sparse matrix operations 1168 89.5 Correspondence

Copying data is not always bad ..... 388 34.17 Consider StaticArrays.jl for small fixed-size vector/matrix operations ..... 389 34.18 Avoid string interpolation for I/O ..... 389 34.19 Optimize network 1031 Matrix factorizations ..... 1032 The uniform scaling operator ..... 1032 75.2 Matrix factorizations ..... 1034 75.3 Standard functions ..... 1034 75.4 Low-level matrix operations Compressed Sparse Column (CSC) Sparse Matrix Storage ..... 1219 89.2 Sparse Vector Storage ..... 1220 89.3 Sparse Vector and Matrix Constructors ..... 1220 89.4 Sparse matrix operations ..... 1221 89.5