MOTIVATION 

• Increasing presence of heterogeneous technologies on pre-exascale supercomputers 

• Need to port key HPC and emerging applications to these systems on time for exascale 

OBJECTIVES 

• Extend the programmability of large-scale heterogeneous systems with GPUs, FPGAs, HBM and NVM 

• Introduce new concepts and functionalities, and implement them in two widely-used HPC programming systems for large-scale supercomputers: MPI and GASPI 

• Maximize the productivity of application development on heterogeneous supercomputers by: – providing auto-tuned collective communication
  – a framework for automatic code generation for FPGAs
  – a memory abstraction device comprised of APIs
  – a runtime for automatic data placement on diverse memories and a DSL for large-scale deep-learning frameworks.