Jupyter-JSC is an interactive web-based development environment to provide the rich ecosystem for high performance computing and data analytic that runs on JSC compute resources. It is designed to support a wide range of workflows in data science, scientific computing and machine learning. It offers a simple, streamlined experience of creating and sharing documents that contain live code, mathematics, explanatory text and their rich media output. Users can configure their individual workflows and add new extensions to expand functionality with the flexible and modular JupyterLab interface. It can serve Jupyter notebooks for group of users as a JupyterHub, by giving users access to a pre-configured computing environment managed efficiently by system administrators."
LAMEC (Load AI Modules, Environments, and Containers) API is developed within Unique AI Framework to facilitate user access to HPC systems. The LAMEC job script generator is available as command line tool and webpage application. After receiving user input of desired HPC system and software, it automatically selects and outputs the right module setup based on re-usable scripts repository, in addition to correct workload manager and network settings. By constantly reaching out to HPC community, LAMEC is adopted by more systems over time, and until now by LUMI, VEGA, JUWELS, DEEP, JURECA and RUDENS systems. LAMEC API supports mainstream machine learning and simulation frameworks, including Pytorch-DDP, Horovod, DeepSpeed, HeAT and Basilisk.