Access to EIDF

EIDF hosts a number of distinct compute and data management services with different levels of availability and some variation in access patterns. The descriptions in the service catalogue will guide you. 

General access to EIDF services depends on your academic or commercial status and your relationship to the Data-Driven Innovation initiative (DDI). 

Computing services 

EIDF charging 

We define the EIDF Standard Offering as access to:   

• All EIDF's cloud services (virtual machines and notebooks) and storage up to several hundred terabytes.  

• The unique, evolving AI environment of the EIDF. This currently includes the following: 

  • The Ultra2 shared memory, high-performance computing service. 
  • The EIDF GPU Service.  
  • The two Cerebras CS-2 AI supercomputers, selected to rapidly accelerate AI research on EIDF.  

Access is granted on a project basis over a fixed period of time (e.g.  for the duration of a grant). You can, of course, request an extension and we will be happy to work with you on this.  

We define DDI Researchers as researchers from the University of Edinburgh and Heriot-Watt University. 

Academic access 

In 2023 the University of Edinburgh recognised EIDF as a Major Research Facility for the purposes of charging in grant applications.  

For the 12-month period starting 1 August 2023 the charge for Academic research projects  making use of the EIDF Standard Offering will be £1,500 per researcher FTE per annum. This applies to all academics and researchers working in UK academic institutions, working with a partner based within the City Deal Region. These rates will increase with inflation on an annual basis.

Please apply on the EIDF Portal and seek a costing prior to application to the funder.

EIDF Portal. 

Collaborative industry access  

Collaborative industry research projects, where a commercial partner is working with a DDI Researcher, can benefit from access to the EIDF Standard Offering.  

For the 12-month period starting 1 August 2023, the charge for projects involving commercial partners will be £7,500 per DDI Researcher FTE per annum. This applies to all academics and researchers working in UK academic institutions, working with a partner based within the City Deal Region. These rates will increase with inflation on an annual basis. For consultancy and contract research projects we expect this fee to be paid by the University of Edinburgh researchers rather than the companies themselves. 

In addition, companies from the City Region/Scotland can benefit from access to the EIDF Standard Offering free of charge for research and development activity.  There will be a fee for support and development services associated with any proposed projects. Please contact us for more detail.  

Companies from outside the City Region/Scotland can gain access to the EIDF for research and development activity, please contact us for more detail.  

Please apply on the EIDF Portal and seek a costing prior to concluding negotiations with the funder.

EIDF Portal. 

Data management and hosting services 

One of the goals of the DDI programme is the retention and reuse of datasets created by innovation projects. EIDF offers free, long-term hosting of datasets for use in future research and innovation projects. Please contact us for more detail, or see the data management section of the service catalogue. 

Data management services catalogue. 

General commercial access 

EIDF has been designed to support a wide range of industry applications and EPCC can provide access to cloud, HPC and storage resources and to consultancy effort to exploit EIDF effectively. 

For commercial access to EIDF and EPCC expertise outwith a collaborative project please get in touch. 

Resource units 

Where we measure things for cost purposes we use the following units: 

• Storage on EIDF is currently granted in units of GiB, where 1 GiB is 1,048,576 kB. 

• Computing time on the Ultra2 large-memory service is measured by the core-hour. 1 Ultra2 core-hour is the equivalent of 1 physical core used for 1 hour. 

• Computing time on the GPU cluster, Cerebras accelerator and Graphcore systems are measured differently. Contact us for details.