Uma arquitetura de distribuição dinâmica de tarefas entre CPU e GPU em jogos digitais

• Ano de defesa: 2007
• Autor(a) principal: Zamith, Marcelo Panaro de Moraes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Programa de Pós-Graduação em Computação; Computação
• Palavras-chave em Português: Ciência da computação; Jogos em 3D; Computação gráfica; Game engine; Computar graphics; 3D games; 3D engines; General poupose GPUs; Real time rendering; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO::COMPUTABILIDADE E MODELOS DE COMPUTACAO
• Link de acesso: https://app.uff.br/riuff/handle/1/17886

Graphic Unit Processors (GPU) are evolving constantly. Recently, they are converging to technology that allow others tasks, different from typically computer graphic one, to be executed. This has created a new computer graphic research field: GPGPU (General Purpose Computation in GPU) due to its increasing capacity in general purpose processing. Digital games are requiring much than available hardware, since more complex tasks and stages are being executed, adopting more extensively AI and physics modules to define the behavior of objects in scene. Physics simulations have characteristics that need a high mathematic computation power, where as the equations that define the body behavior are widely applied, arising the question of how to use GPU to solve physics problems. This work, in spite of ensuring continuity to UFF research field on the digital games development, starts a new research line in visual computation and interface field: the use of GPU to solve general problems, allowing their usage for applications of other research line. This work aims to show a new architecture pattern of digital games with dynamic task allocation that consists in allocating a task in one of the processors (CPU or GPU), making available a set of toolkit allowing an casually implementation of any kind of application. This work also permits that programmers build applications using GPUs but without experience at shader programming