Comparison of Cilk, Kaapi and CUDA for the Jacobi Method
Instituto de Informatica – Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil
11th Workshop on Parallel and Distributed Processing (WSPPD), 2013
While multi-core architectures allow for high performance gains, the use of a specialized framework is often required to maximize the use of the available hardware. This paper compares the performance of three different frameworks, Cilk, Kaapi and CUDA, in implementing the Jacobi method.
August 21, 2013 by hgpu
