Infrastructure
Artemisa (ARTificial Environment for ML and Innovation in Scientific Advanced Computing)
Artemisa is a high performance computing infrastructure dedicated to artificial intelligence and machine learning located on IFIC´s data centre. The and machine learning capacity for handling large amounts of heterogeneous data to learn their underlying characteristic in a computationally efficient way provides new forms to produce empirical models in physics, chemistry, biology and social studies.
Artemisa advanced features and excellent performance make posible the accelerated development of projects involving artificial intelligence areas, machine learning or deep learning can be used for applications such as;
- Image and video object detection
- Signal processing
- Automatic segmentation
- Predictive analysis
- Detection and early warning
- Risk stratification
- Temporary events forecasting
- Large volumes of data management
Their advanced technical characteristics are useful to several IFIC research projects in areas ranging from medical physics to nuclear physics, from neutrino physics to high-energy physics. Nevertheless, Artemisa is an open infrastructure to the entire research community and to any research group affiliated to a Spanish public university or other public research institution which may request their use in the subsequent calls opened specifically for this purpose.
Requests for permission to use are evaluated by a committee that takes account of the scientific interests of the proposal, the research group experience and adequacy of calculations to the Artemisa resources. Artemisa pays special attention to its ML/ IA applications to solve social challenges and to promote their the arrival to the productive sector, specially those related with Comunidad Valenciana RIS3.
Equipment
- 23 workstations with NVIDIA Volta V100 SMX2 GPUs with 32 GB of memory
- NV link as a way of communicating with GPUs, resulting in up to 300 GB/s two-way bandwidth.
- Two Intel Platinum 8180 processors with ECC DDD4 of 768 GB memory
- Solid state disks with 10 Gbps network interfaces.
From the UCIE we are at your fully service through our contact form for any additional information about this infrastructure contact.