Scientific Computing Service (SCI-uc)
Servicio de Computación Avanzada para la Investigación · University of Cantabria
The Scientific Computing Service (SCI-uc) is a specialized unit of the University of Cantabria that provides advanced computing infrastructure and technical support for university research groups, research organizations, and companies involved in research and innovation. Its mission is to facilitate access to and effective use of HPC infrastructure, supporting scientific and technological work.
SCI-uc supports two infrastructure models: Hosting, where infrastructure is operated through a common SCI-uc service framework, and Housing, where research groups operate and manage their own clusters. Access for external research users and industry collaborations is subject to project application and technical validation. See the SCI-uc website for current service information.
Get Started
| Task | Documentation |
|---|---|
| Request an account and connect to the clusters | Request access · Connection guide |
| Find partitions and run compute jobs | Partitions · Slurm user guide · MPI jobs |
| Find and manage software | Software Management |
| Choose storage or transfer data | Storage · Data transfer service |
| Monitor resources or get help | Resource monitoring · Support channels |
About the Service
SCI-uc provides the following services:
- Equipment housing and infrastructure hosting: secure rack space, power, cooling, network connectivity, and monitoring at the 3Mares data center for research computing equipment, including externally acquired systems.
- HPC system installation and configuration: design, deployment, and optimization of environments tailored to research projects and integrated with existing infrastructure and workflows.
- Maintenance and technical support: continuous monitoring, maintenance, and periodic upgrades to support reliability, performance, and service continuity.
- Specialist consulting: tailored technical solutions and advice on efficient resource use and HPC best practices.
- Procurement advice: guidance on hardware and software acquisition, considering project performance, scalability, and budget requirements.
The 3Mares facilities are intended to provide a professionally managed, highly available environment for hosted equipment.
This documentation focuses on practical access and use. For cluster equipment, resource specifications, software, storage policies, and service procedures, use the guides linked above; these pages are maintained in their respective sections so operational details can be updated independently of this introduction.
For help with access, resources, or technical issues, contact soporte.sci@unican.es or consult the support channels.
Research Groups and Collaborating Organizations
The following list shows the research groups and collaborating organizations, together with their respective departments, that are integrated into the SCI-uc service as of January 2026:
- Department of Computer Science and Electronics
- Department of Earth Sciences and Condensed Matter Physics (CITIMAC)
- Department of Water Science and Environmental Engineering
- Department of Applied Mathematics and Computational Sciences
- Department of Applied Physics
- Department of Economics
- Department of Geography, Urban Planning and Spatial Planning
- Department of Mathematics, Statistics and Computing
- Department of Ground and Materials Science and Engineering (DCITYM)
- Institute of Physics of Cantabria (IFCA, SCIC)
- Other entities
Detailed Infrastructure and Software Reference
(**) The infrastructure and software information below is current as of January 2026.
3Mares Data Center
80 m2, divided into two work areas:
- Assembly area for equipment assembly and repair.
- Operations area ("cold" room) with a CUBO AP system by Schneider Electric.
- 23 IT racks.
- Four cooling racks (InROWs) and a free-cooling system, with up to 120 kW of cooling power (70 kW free cooling).
- Three power racks (UPS and PDUs), with up to 320 kW of electrical power.
Stack hardware
Advanced computing servers
Advanced computing clusters for research groups (Housing) and the center/service (Hosting):
- More than 90 compute working nodes, with more than 4,600 cores and 15 TB RAM.
- Two GPU working nodes, with two NVIDIA Quadro RTX 4000 GPUs.
Networking
- 1/10 Gbps Ethernet network.
- 200 Gbps HPC InfiniBand network for computation and data storage.
Storage
- Lustre filesystem: more than 1.2 PB of high-performance remote storage.
- NFS filesystem: more than 80 TB of remote home storage.
- Local scratch (SSD/SATA): more than 100 TB.
- NAS servers: two units with 46 TB and 64 TB for backups.
Stack Base software
- All working nodes run Rocky Linux 8.10.
Stack HPC software
Objectives
- Hide hardware complexity from end users.
- Facilitate usability of the overall system.
- Simplify management.
Main HPC management platforms and software:
- Slurm v24.11.3.
- Spack v0.23.0 and Module Environment (Lua) v8.7.65:
gcc-13.3.0/gcc-8.5.0-rt6fdintel-oneapi-compilers-2024.2.0/gcc-8.5.0-st4ziintel-oneapi-mkl-2024.2.0/gcc-8.5.0-q44bfintel-oneapi-mpi-2021.14.0/gcc-8.5.0-iywn6openblas-0.3.28/gcc-8.5.0-qmqz7openmpi-5.0.5/gcc-8.5.0-cuim6openmpi-5.0.5/gcc-8.5.0-slurm-ogwvqhdf5-1.14.5/gcc-8.5.0-q24s5
System management properties
- Custom-developed for learning and optimization.
- Based exclusively on open-source software.
- Uses server virtualization.
- Continuously evolving and improving.
- 99% of the software is free.
Main infrastructure management platforms:
- XCP-ng and Xen Orchestra (XOA).
- Container infrastructure (Docker).
- Rocky Linux 8.10 and 9.6.
- Slurm job manager.
- Ganglia monitor and Nagios Core.