EAR 5.1 is a system software for energy management, accounting and optimization for super computers. Main EAR services are: 1. Application energy optimization. An **easy-to-use** and **lightweight** optimization service to automatically select the optimal CPU, memory and GPU frequency according to the application and the node characteristics. This service is offered by EAR core components: The EAR library and EAR Node Manager. EARL is a runtime library automatically loaded with the applications and it offers application metrics monitoring and it can select the frequencies based on the application behaviour on the fly. The Library is loaded automatically through the EAR Loader (EARLO) and it can be easly integrated with different system batch schedulers (e.g., SLURM). The EARL provides deep application accounting(both power/energy and performance) and energy optimization in a completelly transparent and dynamic way. 2. Job and Node monitoring: A complete **energy and performance accounting and monitoring system** . Node and application monitoring are also provided by the EAR core components (EAR library and EAR node manager). These two components are the data provides and information is reported to the DB using the EAR DB manager (EARDBD). The EARDBD is a distributed service offering buffering and aggregation of data, minimizing the number of connections with the DB server. EAR includes several report plugins for both relational (MariaDB and PostgreSQL) and non-relational Databases such as EXAMON. EAR commands for data reporting are only based on relational DBs. 3. Cluster power management (powercap). A **cluster energy manager** to monitor and control the energy consumed in the system through the EAR Global Manager (EARGMD) and EARD. EAR support a powerful and flexible configuration where different architectures with different power limits can be configured in the same cluster. CPU and CPU+GPU nodes are supported. Visit the [architecture page](https://gitlab.bsc.es/ear_team/ear/-/wikis/Architecture) for a detailed description of each of these components. The [user guide](User-guide) contains information about how to use EAR as an end user in a production environment. The [admin guide](https://gitlab.bsc.es/ear_team/ear/-/wikis/Admin-guide) has all the information related to the installation and setting up, as well as all core components details. Moreover, you can find a list of [tutorials](https://gitlab.bsc.es/ear_team/ear/-/wikis/Tutorials) about how to use EAR. ## License EAR is an open source software licensed under EPL-2.0 license. Full text can be found in [COPYING.EPL-2.0](https://gitlab.bsc.es/ear_team/ear/-/blob/master/COPYING.EPL-2.0) file distributed with the source code. Contact: [ear-support@bsc.es](mailto:ear-support@bsc.es). ## Publications [J. Corbalan, L. Alonso, J. Aneas and L. Brochard, "Energy Optimization and Analysis with EAR," 2020 IEEE International Conference on Cluster Computing (CLUSTER), 2020, pp. 464-472, doi: 10.1109/CLUSTER49012.2020.00067.](https://ieeexplore.ieee.org/document/9229570) [J. Corbalan, O. Vidal, L. Alonso and J. Aneas, "Explicit uncore frequency scaling for energy optimisation policies with EAR in Intel architectures," 2021 IEEE International Conference on Cluster Computing (CLUSTER), 2021, pp. 572-581, doi: 10.1109/Cluster48925.2021.00089.](https://ieeexplore.ieee.org/document/9555970) [J. Corbalan, L. Alonso, C. Navarrete and C. Guillen, "Soft Cluster Powercap at SuperMUC-NG with EAR," 2022 IEEE 13th International Green and Sustainable Computing Conference (IGSC), Pittsburgh, PA, USA, 2022, pp. 1-8, doi: 10.1109/IGSC55832.2022.9969360](https://ieeexplore.ieee.org/document/9969360) ## Current version This documentation corresponds to EAR version 5.1.