Accelerating Band Linear Algebra Operations on GPUs with Application in Model Reduction

被引：0

作者：

Benner, Peter ^{[1
]}

Dufrechou, Ernesto ^{[3
]}

Ezzatti, Pablo ^{[3
]}

Igounet, Pablo ^{[3
]}

Quintana-Orti, Enrique S. ^{[2
]}

Remon, Alfredo ^{[1
]}

机构：

[1] Max Planck Inst Dynam Complex Tech Syst, Magdeburg, Germany

[2] Univ Jaime I, Dep Ingn & Ciencia Computac, Castellon de La Plana, Spain

[3] Univ Republica, Inst Computac, Montevideo 11300, Uruguay

来源：

COMPUTATIONAL SCIENCE AND ITS APPLICATIONS, PART VI - ICCSA 2014 | 2014年 / 8584卷

关键词：

Band linear systems; linear algebra; graphics processors (GPUs); high performance; control theory;

D O I：

暂无

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

In this paper we present new hybrid CPU-GPU routines to accelerate the solution of linear systems, with band coefficient matrix, by off-loading the major part of the computations to the GPU and leveraging highly tuned implementations of the BLAS for the graphics processor. Our experiments with an nVidia S2070 GPU report speed-ups up to 6x for the hybrid band solver based on the LU factorization over analogous CPU-only routines in Intel's MKL. As a practical demonstration of these benefits, we plug the new CPU-GPU codes into a sparse matrix Lyapunov equation solver, showing a 3x acceleration on the solution of a large-scale benchmark arising in model reduction.

引用

页码：386 / 400

页数：15

共 13 条

[1] Anderson E., 1992, LAPACK Users Guide
[2] [Anonymous], 2013, TOP500 LIST
[3] ANTOULAS A. C., 2005, ADV DES CONTROL, DOI 10.1137/1.9780898718713
[4] Barrachina S, 2008, LECT NOTES COMPUT SC, V5168, P739, DOI 10.1007/978-3-540-85451-7_79
[5] Matrix inversion on CPU-GPU platforms with applications in control theory
Benner, Peter
Ezzatti, Pablo
Quintana-Orti, Enrique S.
Remon, Alfredo
[J]. CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2013, 25 (08) : 1170 - 1182
[6] Cuthill E., 1969, P 1969 24 NAT C, P157, DOI [DOI 10.1145/800195.805928, 10.1145/800195.805928]
[7] Du Croz J., 1990, 21 LAPACK U TENN
[8] Farber R, 2011, CUDA APPLICATION DESIGN AND DEVELOPMENT, P1
[9] Kirk D., 2012, PROGRAMMING MASSIVEL, VSecond
[10] A cyclic low-rank smith method for large sparse Lyapunov equations
Penzl, T
[J]. SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2000, 21 (04) : 1401 - 1418

← 1 2 →