Cambridge’s Centre for Theoretical Cosmology (CTC) has facilitated the development of a high-performance computing (HPC) ecosystem for 20 years to advance development of the theories of the universe that are both mathematically consistent and observationally testable.
In so doing, CTC has engaged in long-standing industrial collaborations with both Intel and Hewlett Packard Enterprise (HPE)/SGI, which have brought significant benefits to these commercial partners through co-design and improvement of the HPC hardware and software.
For instance, the collaboration of CTC with HPE has resulted in new capabilities and improvements for the SGI UV2 and UV3 supercomputers. These were incorporated in the HPE Superdome Flex server, HPE’s flagship in-memory system deployed by commercial customers with large data flows worldwide.
CTC has also collaborated with Intel on advancing hardware acceleration on HPC systems, as well as creating in-situ visualisation capabilities. Collaborative innovations have been released in open-source libraries and applications used worldwide, such as incorporated in 150–200 commercial graphics tools and commonly used on Top 500 Supercomputers.