Access to High-Performance Computing (HPC) infrastructure environment for Data- or AI- based solutions that require dedicated GPUs. The user is provided with access to a Jupyter notebook on a Docker that has access to the GPUs. Alternatively, if the user has other needs that can be deployed in a docker, it can be adapted accordingly. The service includes technical support for the configuration and deployment of the experimentation infrastructure, as well as issue reporting.
SERVICE DESCRIPTION
Access to High-Performance Computing (HPC) infrastructure environment for Data- or AI- based solutions that require dedicated GPUs. The service includes technical support for the configuration and deployment of the experimentation infrastructure, as well as issue reporting.
Features:
- The user will be provided with access to a Jupyter notebook on a Docker that has access to the available GPUs.
- Alternatively, if the user has other needs that can be deployed in a Docker, it can be adapted accordingly.
- The main advantage is that the user has control of deploying everything that is required, using a docker already configured with everything that is necessary to carry out an experiment.
SPECIAL ACCESS CONDITIONS
The user required environment must be a Jupyter notebook or something that can be deployed in a Docker.
PREREQUISITES
The user must be registered
CASE EXAMPLES
A user has a model that has to be trained using GPUs. He is provided with a Jupyter notebook to prepare the training and launch it against the GPUs.
SERVICE CAN BE COMBINED WITH
It can be combined with services or datasets that requires advanced data processing using GPUs. The tools that can be used to interact with the GPUs are either Jupyter (provided) or anything that can be put in a docker.