Enhancement Of Computational Engineering Within An Undergraduate Mechanical Engineering Curriculum

The NSF supported Course, Curriculum and Laboratory Initiative (CCLI) project described herein addresses concerns regarding undergraduate education at research universities as highlighted in the 1998 Boyer Commission by incorporating advances in information technology into the curriculum. This has been accomplished by developing an “emphasis” to the department’s undergraduate mechanical engineering degree in the area of computational engineering. To complete the emphasis, students need to complete four upper division elective courses related to computational engineering. Three courses concentrate on applied modeling and simulation; the fourth (which was developed under the CCLI award) concentrates on implementing algorithms on parallel computing architectures. To support the emphasis, the authors have designed and assembled a PC Beowulf teaching cluster. The cluster consists of a server node where students can log in and develop their programs, as well 10 dual Opteron compute nodes for running and testing parallel codes. AMD Opteron CPUs were selected for the cluster since they may be used to teach both shared and distributed memory programming techniques, and for their strong price/performance ratio. Using gigabit networking technology, the cluster was built at a cost of approximately $15k. In addition to traditional coursework, a cluster computing workshop was developed and offered for the first time during the summer of 2004. Key issues covered were designing a Beowulf cluster, implementing and programming a cluster, and tuning/profiling of programs.