The HPC platform offers generous resources for free to the researchers. This baseline service should be sufficient for most research projects. However, some projects might require more resources or fully dedicated resources for a limited time or permanently. Allocating large amount of resources for a few users will reduce the pool of freely accessible resources for all the other users. The win-win offer is then to propose extra paid resources for the few users and the generated incomes are re-invested in the HPC infrastructure to further increase globally the available resources.
Some research projects require more storage space than the default offered for free. We can quickly offer a few extra terabytes on request. For much larger volumes (> 10 TB), please contact us well in advance to determine the feasability and the roadmap to obtain the extra storage space.
It has been the default offer for several years: renting (a) full compute node(s). You gain access to (a) fully dedicated server(s) available 24/7 with a dedicated queue to run jobs on the rented node(s). We'll continue this offer but, with the development of alternative offers, this solution will be limited to very specific cases. A rented node not being used 100% all the time is a wasted resource.
If you know in advance that extra computing resources will be needed, you can ask for a reservation of such resources. A time range and the required number of cores is all what will be needed. Once the time has arrived, you'll get for you, or your team, the requested resources fully available to run your jobs.
If you plan to ask a grant for extra compute nodes or obtained a budget to purchase compute nodes, get in touch with us to evaluate the options for hosting the compute nodes in our HPC platform. Be aware that we impose some constrains (rack space and power is limited). It is therefore mandatory to contact us as early as possible if you plan to benefit from our hosting service.