RT Journal Article
T1 Study of basic vector operations on Intel Xeon Phi and NVIDIA Tesla using OpenCL
A1 Coronado Barrientos, Edoardo
A1 Indalecio Fernández, Guillermo
A1 García Loureiro, Antonio Jesús
K1 Nvidia Tesla
K1 Intel Xeon Phi
K1 OpenCL
K1 Kernel
K1 Vector operations
K1 AXPY
K1 DOT
K1 SPMV
AB The present work is an analysis of the performance of the basic vector operations AXPY,DOT and SpMV using OpenCL. The code was tested on the NVIDIA Tesla S2050 GPU andIntel Xeon Phi 3120A coprocessor. Due to the nature of the AXPY function, only two versionswere implemented, the routine to be executed by the CPU and the kernel to be executed onthe previously mentioned devices. It was studied how they perform for different vector’s sizes.Their results show the NVIDIA architecture better suited for the smaller vectors sizes and theIntel architecture for the larger vector’s sizes. For the DOT and SpMV functions, there are threeversions implemented. The first is the CPU routine, the second one is an OpenCL kernel thatuses local memory and the third one is an OpenCL kernel that only uses global memory. Thekernels that use local memory are tested by varying the size of the work-group; the kernels thatonly uses global memory are tested by varying the arrays size. In the case of the first ones, theresults show the optimum work-group size and that the NVIDIA architecture benefits from theuse of local memory. For the latter kernels, the results show that larger computational loadsbenefits the Intel architecture
PB Universidad de Granada
SN 2341-3158
YR 2015
FD 2015
LK http://hdl.handle.net/10347/17698
UL http://hdl.handle.net/10347/17698
LA eng
NO Coronado Barrientos, E., Indalecio Fernández, G. and García Loureiro, A. (2015). Study of basic vector operations on Intel Xeon Phi and NVIDIA Tesla using OpenCL, v. 2, n. 1, pp. 26-40
NO This work has been supported by FEDER funds and Xunta de Galicia under contract GRC 2014/008, and by Spanish Government (MCYT) under project TEC2010-17320 and TIN-2013-41129-P
DS Minerva
RD 27 abr 2026