John Feo is the director of the Center for Adaptive Supercomputer Software (CASS) at the Pacific Northwest Laboratory. John and his team are using a Cray XMT 1 supercomputer since three years to study a large spectrum of applications like graphs algorithms, Dynamic network analysis (DNA), Emerging Subnetwork Patterns, bioinformatics, semantic search, and semantic database development.
The CASS group is also developing compilers and runtime systems to improve the compilation of code and is working on new programming models for highly multithreated machines like the XMT.
For John one important application of the Cray XMT will be in companies like Facebook, Google, or banks doing graphs algorithms, semantic search and knowledge discovery on huge data sets.
In this interview John explains the particularities of developing applications for this new kind of architectures and how his team is supporting users and scientists in porting their code.