With Shaheen II compute capacity and capability, nicely layered data storage infrastructure and cross-campus network connectivity, it is expected that KAUST scientists will be able to generate a large amount of scientific data. Data Analytics using state of the art solutions will become a pressing need for the user to analyze all scientific output of simulations processed on the Shaheen II system. A Cray Urika-GD data analytics appliance is installed at KAUST for the challenges of the discovery process, transforming these massive amounts of seemingly unrelated data into relevant insights and new scientific breakthroughs. The Urika-GD appliance consists of:
- Graph analytics platform, providing graph-optimized hardware with 2TB of global shared-memory, 64 Threadstorm4 processors with 128 hardware threads per processor and more than 75 TB of Lustre parallel filesystem for scalable I/O,
- Graph analytics database, providing an RDF triplestore and SPARQL query engine, and
- Graph analytics application services, providing management, security and data pipeline functions.