Title :
High performance linear equation solver using NVIDIA GPUs
Author :
Leow Yoon Kah ; Akoglu, Ali ; Guven, Ibrahim ; Madenci, Erdogan
Author_Institution :
Univ. of Arizona, Tucson, AZ, USA
Abstract :
The solution of a linear system of equations constitutes an important part in the field of linear algebra that is widely used in industries like aerospace, aeronautics, solid mechanics, fluid dynamics, oil research and numerous others. A direct method for solving these equations is Gaussian Elimination, which consists of forward elimination and back substitution. We have tailored this method to take advantage of the massive parallelism offered by NVIDIA GPU architectures. Thorough evaluations have been performed for variants of our implementation that exploit different memory features on an NVIDIA Tesla C1060 GPU. Compared to a serial implementation on an Intel Core I7, the execution time for forward elimination on the GPU is reduced by a factor of 183X when using both global and shared memory systems, and by a factor of 185X when using only global memory.
Keywords :
computer graphic equipment; coprocessors; finite element analysis; linear algebra; parallel architectures; shared memory systems; storage management; Gaussian elimination; NVIDIA GPU architecture; NVIDIA Tesla C1060 GPU; aeronautics; aerospace; back substitution; fluid dynamics; forward elimination; global memory system; high performance linear equation solver; linear algebra; massive parallelism; memory feature; oil research; shared memory system; solid mechanics; Equations; Graphics processing unit; Instruction sets; Kernel; Mathematical model; Matrices; Random access memory;
Conference_Titel :
Adaptive Hardware and Systems (AHS), 2011 NASA/ESA Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4577-0598-4
Electronic_ISBN :
978-1-4577-0597-7
DOI :
10.1109/AHS.2011.5963960
