Introduction to GPU Computation
Matthew Smith李汶樺National Cheng-Kung University 下午 1:00–5:00 鴻經館 201-4 室Description of this intensive short course
Modern engineering disciplines require computations and insight which cannot be obtained through analytical solutions alone. To cope with the demand for solutions to such real life problems, engineers and scientists have dedicated large amounts of time and effort to employ modern computational technology to aid them in their engineering analysis. Such efforts have led to the creation of simulation tools such as Fluent, ANSYS to name a few.
While these tools have been beneficial, the limit to which we can use such tools to solve real life problems is limited by the speed at which these computations can be performed. With increasing demands placed upon engineers for increased fidelity in their computations, there is a demand for larger scale simulation results. Hence, modern engineers require parallel computing to obtain suitable “modern” results.
The goal of this intense workshop is to apply a problem-centred approach to parallel computing using GPU devices. Students will be presented with everyday (common) problems in mathematics and science and will be required to use the GPU to solve these problems. Hence, the approach used during this workshop is a form of Problem Based Learning. As such, during activities, students will be grouped into pairs and will work together to solve these potentially challenging problems.