演講公告

新聞標題：　( 2010-03-30 )

• 演講主題：Introduction of GPU Computing

• 主講人：余啟哲教授 (交通大學應用數學系)

• 演講日期：99年4月6日(星期二) <br> 下午3:10 –4:00

• 演講地點：(光復校區)科學一館223室

• 茶會時間：當天下午1:30科學一館205室

• 摘要內容：
  

  Abstract. The original applications of GPU were primarily concentrated on real-time 3D visualization, e.g. 3D animations and games. However, in the past five years, the general purpose computing with graphics processor units (GPGPU) is becoming increasingly viable and powerful. According to reports, efficiency of GPU implementation averages from 10 to 100 times of CPU, but there are still many limitations in programming. The realization of parallelism in GPU coding is thoroughly distinct from that in CPU structures, some important but unfamiliar attributes (e.g. hardware architecture) which lead the overall performance must be setup and modified, based on individual numerical algorithms. In this talk, we will sketch our strategy for CUDA with finite volume method.
  

go back