Performance evaluation of data migration methods between the host and the device in CUDA-based programming
Carregando...
Data
2016-04-01
Orientador
Coorientador
Pós-graduação
Curso de graduação
Título da Revista
ISSN da Revista
Título de Volume
Editor
Tipo
Capítulo de livro
Direito de acesso
Acesso aberto
Resumo
CUDA-based programming model is heterogeneous – composed of two components: host (CPU) and device (GPU). Both components have separated memory spaces and processing units. A great challenge to increase GPU-based application performance is the data migration between these memory spaces. Currently, the CUDA platform supports the following data migration methods: UMA, zero-copy, pageable and pinned memory. In this paper, we compare the zero-copy performance method with the other methods by considering the overall application runtime. Additionally, we investigated the aspects of data migration process to enunciate causes of the performance variations. The obtained results demonstrated in some cases the zero-copy memory can provide an average performance on 19% higher than the pinned memory transfer. In the studied situation, this method was the second most efficient. Finally, we present limitations of zero-copy memory as a resource for improving performance of CUDA applications.
Descrição
Palavras-chave
Idioma
Inglês
Como citar
Advances in Intelligent Systems and Computing, v. 448, p. 689-700.