D2.6. Deployment of the final, refined version of the Model Execution Environment

Deliverable 2.6 describes the final deployment of the Model Execution Environment (MEE) tool integrated with the Prometheus supercomputer and used in the PRIMAGE project to simplify model creation and execution on the distributed High Performance Computing (HPC) resources. This document constitutes the sixth deliverable of WP2. It is a continuation of the D2.1 [D2.1] and D2.4 [D2.4] deliverables where we mainly focus on HPC. It covers the final version of the HPC infrastructure and accompanying tools (MEE, Rimrock and PLG-Data) used to run advanced multiscale simulations on distributed resources. This document comprises a description of the architecture and technical aspects connected with service usage, improvement and deployment. We also show WP5 models integrated with MEE, which, later on, are used by the orchestrator to run complex multiscale simulation on the Prometheus HPC infrastructure. Detailed information connected with granting access to the system and its usage can be found in previous WP2 deliverables (D2.1 and D2.4). The main target of this document is the global technical team of the PRIMAGE project that is using, or will be using, HPC resources to create and run complex simulations which can take benefits from huge distributed resources delivered by the Prometheus HPC cluster. It also serves the dissemination and exploitation role for PRIMAGE partners, enabling them to understand the architecture of the service and integration between the cluster, the PLGrid infrastructure, and the European Open Science Cloud (EOSC), especially with the Check-in [Checkin] service. This document is organized in seven sections, starting with this introduction. Section 2 summarizes the Model Execution Environment (MEE) architecture, focusing on the main building blocks and integration with accompanying services. Section 3 presents extended MEE features implemented since the release of D2.4. Section 4 describes the repositories and service URLs that have been used in the framework of the project. Section 5 shows example WP5 models and their configuration required to run such models from MEE on the distributed Prometheus HPC resources. Finally, Section 6 presents conclusions and Section 7 provides references