Project overview
IBM Research - Zurich has developed over the years a deep and broad competence in the field of parallel and distributed computing. Our internal work on scientific and technical computing is used as a test bed and driver for innovative parallel alghorithms and implementations on diverse computer hardware ranging from general-use high-end supercomputers (p-Series clusters) and Linux clusters (x-Series) to novel massively computer architectures such as BlueGene ( L and P) and heterogeneous architectures as the Cell processor.
Our key competencies are:
- distributed memory parallelization using MPI,
- shared memory parallelization using OpenMP and direct threading,
- mixed MPI/SMP schemes
- generation parallel programing
Projects
- CPMD- selected tab,
- Performance evaluation
- GROMOS
- Joint projects
CPMD code parallelization and tuning
The CPMD code has been implemented and tuned for the entire generation of IBM supercomputers. This made the code a reference in the world of high-perfomance simulations. Our recent dual-level (distributed memory/shared memory) implementation of CPMD is able to sustain 1 Teraflop on 32 clustered (via colony switches) p690 systems (1024 processors), with 45% parallel efficiency (1 to 1024 processors). These are still the best results for similarly complex codes.
Currently, a system of about 1000 atoms of silicon (corresponding to a memory
size of 250 GB) can be handled with 0.5 TF/s.
