DS2. Performance and Energy Aware Workload Partitioning on Heterogeneous Platforms
Student: Li Tang (University of Notre Dame)
Advisor: Xiaobo Sharon Hu (University of Notre Dame)
Abstract: Heterogeneous platforms which employ a mix of CPUs and accelerators such as GPUs have been widely used in HPC. Such heterogeneous platforms have the potential to offer higher performance at lower energy cost than homogeneous platforms. However, it is rather challenging to actually achieve the high performance and energy efficiency promised by heterogeneous platforms. To address this issue, this work proposes a framework to assist application developers to partition workload on heterogeneous platforms for achieving high performance or energy efficiency before actual implementation. The framework includes both analytical performance/energy models and two sets of workload partitioning guidelines. Based on the design goal, application developers can obtain a workload partitioning guideline for a given platform and then follow it to partition workload. Then the performance/energy models can be used to estimate the performance or energy of the obtained workload partitions and help select the appropriate workload partition.
Doctoral Showcase Index