earticle

영어

Recently, the concept of HPC as a Service (HPCaaS) was proposed to bring the traditional high performance computing field into the era of cloud computing. One of its goals aims to allow users to get easier access to HPC facilities and applications. This paper deals with related job submission and scheduling issues to achieve such goal. Traditionally, HPC users in supercomputing centers are required to specify the amount of processors to use upon job submission. However, we think this requirement might no longer be necessary for HPCaaS users since most modern parallel jobs are moldable and users usually could not know how to choose an appropriate amount of processors to allow their jobs to finish earlier. With moldable property, parallel programs can choose to exploit different parallelisms for execution just before their start of running. Therefore, we propose two moldable job scheduling approaches which take advantage of information of application speedup model to not only relieve HPC users’ burden of selecting an appropriate number of processors but also achieve even better system performance than existing methods. The experimental results indicate that the proposed approaches can achieve up to 78% and 89% performance improvement in terms of average turnaround time.