Remora: a resource monitoring tool for everyone

Knowing about the requirements of HPC applications is a common question that users of high performance systems ask often. However, answering this question requires the collaboration of administrators and sometimes the answer does not contain the amount of detail that users demand. This work introduces a new user space resource monitoring tool, REMORA. REMORA stands for REsource MOnitoring for Remote Applications, and provides a simple interface to gather important system utilization data while running on HPC systems. REMORA is designed to provide a brief text report and post-processing tools to analyze the very detailed records taken during an application run. Users can configure the tool to achieve the amount of detail that they want and perform the analysis of the results at any point in time. REMORA helps users achieving a better understanding of their applications by providing a high level profile of their executions and users can take advantage of that information to improve their codes.