Harnessing CUDA Dynamic Parallelism for the Solution of Sparse Linear Systems (CROSBI ID 635190)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Aliaga, José ; Davidović, Davor ; Pérez, Joaquín ; Quintana-Ortí, Enrique S.
engleski
Harnessing CUDA Dynamic Parallelism for the Solution of Sparse Linear Systems
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.
Graphics processing units (GPUs); CUDA dynamic parallelism; sparse linear systems; iterative solvers; high performance; energy efficiency
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
217-226.
2016.
objavljeno
Podaci o matičnoj publikaciji
Advances in Parallel Computing Volume 27: Parallel Computing: On the Road to Exascale
Joubert, G.R., Leather, H., Parsons, M., Peters, F., Sawyer, M.
IOS Press
978-1-61499-620-0
Podaci o skupu
Nepoznat skup
predavanje
29.02.1904-29.02.2096