Artelys Crystal HPC: Our integrated, sovereign High-Performance Computing (HPC) managed platform for industries

Artelys Crystal HPC is a fully managed HPC-as-a-Service platform designed for all types of complex modeling and intensive simulation.

Thanks to its advanced workload orchestrator, both interactive sessions (such as Desktop and Jupyter Lab) and heavy production tasks are managed within a unified scheduling system. This ensures the frictionless, simultaneous execution of jobs, spanning from exploratory research to full-scale production. 

• Collaborative environment: Teams work together on a unified platform without interfering with each other’s processes.  
• Seamless orchestration: Launch your simulations in parallel without worrying about resource contention or workflows collision. 

Performance is a core pillar of Artelys Crystal HPC, providing a stable and highly scalable environment for the most demanding computational tasks.

The platform is engineered to handle massive data volumes and complex simulations with exceptional numerical stability and high throughput. 

• Cutting-edge hardware: In partnership with OVHcloud, we leverage high-end Bare Metal servers (featuring up to 3 TB of RAM per node). This architecture eliminates the performance overhead typical of virtualization, ensuring your code runs at its full potential. 
• A fully managed experience: As a managed service, Artelys Crystal HPC frees you from the complexities of hardware lifecycle management. You no longer need to worry about server maintenance, obsolescence, or procurement. 
• Transparent evolution: Upgrading machine configurations or migrating to the latest generation of processors is seamless and transparent. Your environment stays at the forefront of technology without any manual intervention or hardware reinvestment on your part. 
• Built-in scalability: The platform is designed to scale alongside your projects, capable of executing industrial-grade workloads comparable to the most advanced research infrastructures.

Artelys Crystal HPC is designed to provide a secure environment for sensitive workloads through localized data management and robust access protocols. 

• Data and workflow containment: To ensure maximum protection, all datasets and computational workflows are strictly confined to the specific cluster where the platform is deployed. Information never leaves the secure environment during processing. 
• Secure access management: The platform leverages Single Sign-On (SSO) for user authentication, ensuring that access to the infrastructure is centralized, encrypted, and strictly controlled. 
• SecNumCloud: For the handling of highly sensitive or critical data, Artelys Crystal HPC can be transparently deployed on SecNumCloud-certified clusters. This allows users to benefit from the highest level of cybersecurity and sovereign compliance (ANSSI) without changing their operational experience. 

Artelys Crystal HPC provides comprehensive visibility into computing workloads through an integrated monitoring stack, ensuring users maintain full control over their simulations and resource consumption.

• Real-time performance tracking: The platform utilizes Prometheus for the automated collection of system metrics, including CPU and RAM utilization, disk I/O, and network health. This data is visualized through Grafana dashboards, allowing users to monitor job execution and resource efficiency in real time.
• Secure monitoring infrastructure: Access to the monitoring suite is strictly protected. Both Prometheus and Grafana operate behind a reverse proxy, with Grafana requiring OAuth-based authentication to ensure that performance data is only accessible to authorized personnel.
• Data and workflow sovereignty: Users retain direct access and total oversight of their data, logs, and simulation results. The platform provides an intuitive interface to manage the entire lifecycle of a study without abstracting away the underlying control of the workflows. 

Artelys Crystal HPC reduces your Total Cost of Ownership (TCO) by focusing on two key financial services: competitive infrastructure rates and the smart optimization of expensive software assets.

• Competitive cloud pricing: Leveraging our long-standing collaboration with OVHcloud, Artelys Crystal HPC offers access to high‑performance computing resources under attractive commercial conditions.
• Resource and license pooling: The platform enables significant savings by centralizing and mutualizing software licenses across multiple teams.  
  • • License optimization: Multiple users can run production workloads on the same high-performance nodes, maximizing license utilization.
    • Guaranteed isolation: The integration of the SLURM workload manager ensures that this pooling does not compromise performance. It manages fair resource allocation and strict task isolation, allowing teams to share infrastructure without any interference or degradation in execution speed.