Harnessing CUDA Dynamic Parallelism for the Solution of Sparse Linear Systems

被引：1

作者：

Aliaga, Jose ^{[1
]}

Davidovic, Davor ^{[2
]}

Perez, Joaquin ^{[1
]}

Quintana-Orti, Enrique S. ^{[1
]}

机构：

[1] Univ Jaume 1, Dept Ingn Ciencia Comp, Castellon de La Plana, Spain

[2] Inst Ruder Baskovic, Ctr Informat & Racunarstvo CIR, Zagreb, Croatia

来源：

PARALLEL COMPUTING: ON THE ROAD TO EXASCALE | 2016年 / 27卷

关键词：

Graphics processing units (GPUs); CUDA dynamic parallelism; sparse linear systems; iterative solvers; high performance; energy efficiency;

D O I：

10.3233/978-1-61499-621-7-217

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

We leverage CUDA dynamic parallelism to reduce execution time while significantly reducing energy consumption of the Conjugate Gradient (CG) method for the iterative solution of sparse linear systems on graphics processing units (GPUs). Our new implementation of this solver is launched from the CPU in the form of a single "parent" CUDA kernel, which invokes other "child" CUDA kernels. The CPU can then continue with other work while the execution of the solver proceeds asynchronously on the GPU, or block until the execution is completed. Our experiments on a server equipped with an Intel Core i7-3770K CPU and an NVIDIA "Kepler" K20c GPU illustrate the benefits of the new CG solver.

引用

页码：217 / 226

页数：10

共 50 条

[41] Remarks concerning linear systems with parallelism
Ostrom, TG
MOSTLY FINITE GEOMETRIES: IN CELEBRATION OF T G OSTROM'S 80TH BIRTHDAY, 1997, 190 : 1 - 7
[42] PRECONDITIONING LINEAR-SYSTEMS AND PARALLELISM
CODENOTTI, B
LEONCINI, M
COMPUTERS AND ARTIFICIAL INTELLIGENCE, 1990, 9 (05): : 471 - 491
[43] Systematic Fusion of CUDA Kernels for Iterative Sparse Linear System Solvers
Aliaga, Jose I.
Perez, Joaquin
Quintana-Orti, Enrique S.
EURO-PAR 2015: PARALLEL PROCESSING, 2015, 9233 : 675 - 686
[44] PARALLELISM IN THE SOLUTION OF LARGE FULL LINEAR-SYSTEMS USING A MATRIX PARTITION APPROACH
HINTZ, RM
AIAA JOURNAL, 1989, 27 (06) : 825 - 826
[45] A numerical evaluation of sparse direct solvers for the solution of large sparse symmetric linear systems of equations
Gould, Nicholas I. M.
Scott, Jennifer A.
Hu, Yifan
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2007, 33 (02):
[46] Attract-Repulse Fireworks Algorithm and its CUDA Implementation Using Dynamic Parallelism
Ding, Ke
Tan, Ying
INTERNATIONAL JOURNAL OF SWARM INTELLIGENCE RESEARCH, 2015, 6 (02) : 1 - 31
[47] Accelerating BIRCH for Clustering Large Scale Streaming Data Using CUDA Dynamic Parallelism
Dong, Jianqiang
Wang, Fei
Yuan, Bo
INTELLIGENT DATA ENGINEERING AND AUTOMATED LEARNING - IDEAL 2013, 2013, 8206 : 409 - 416
[48] AN ITERATIVE SOLUTION METHOD FOR SOLVING SPARSE NONSYMMETRIC LINEAR-SYSTEMS
MAKINSON, GJ
SHAH, AA
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 1986, 15 (03) : 339 - 352
[49] Numerical performance of preconditioning techniques for the solution of complex sparse linear systems
Mazzia, A
Pini, G
COMMUNICATIONS IN NUMERICAL METHODS IN ENGINEERING, 2003, 19 (01): : 37 - 48
[50] Application of the Cramer rule in the solution of sparse systems of linear algebraic equations
Mittal, RC
Al-Kurdi, A
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 2001, 136 (1-2) : 1 - 15

← 1 2 3 4 5 →