University at Buffalo, Center for Computational Research
XDMoD is an NSF-funded comprehensive resource management tool designed to provide XSEDE HPC center personnel and leadership with the ability to obtain detailed operational metrics of HPC systems, coupled with an extensive analytical capability. This allows users to optimize performance at the system and job level, ensure quality of service and provide accurate data to guide system upgrades and acquisitions. The original project to monitor HPC resource utilization has been expanded to integrate XDMoD with the OnDemand utility, to develop a federated version of Open XDMoD in which the utilization from multiple clusters or resources can be grouped together to present an integrated view and to integrate XDMoD with the newly developed ColdFront allocation utility. During a brief overview of XDMoD, both basic capabilities and new ones will be introduced. We will introduce the user to the core features of XDMoD in a live demonstration.