High-Performance Computing Directions
Efficiency of message passing and data parallel programming
- Helped by standards efforts such as PVM, MPI, Open-MP and (HPF)
Distributed memory with a move toward shared memory
- SMPs and Distributed Shared Memory
- Distributed or shared address space w/deep memory hierarchy
Clustering of shared memory machines for scalability
- Emergence of PC commodity systems
- Pentium based, NT or Linux driven
- At UTK cluster of 18 (dual&single) Pentium based > 10 Gflop/s peak
Complementing “Supercomputing” with Metacomputing in the form of the Computational Grid