Personal tools
Log in

High performance computing

COURSE: High performance computing

Code: ФЕИТ06012

ECTS points: 6 ECTS

Number of classes per week: 3+0+0+3

Lecturer: prof. Marija Kacarska

Subject of the course content: Architectures for high-performance computing. Compilers for High Performance Systems. Removal of program loops. Parallelization. High performance systems. Mass memories. Coupling networks and clusters. Switching networks and clusters. Grid structures. Grid computing. Pipelining. Performance estimation and optimization. Applications for grid environment. High-performance microprocessors. Design and evaluation of modern parallel processors. Principles of parallelism. Instruction level parallelism. Multiprocessor systems. Multicore processors. Memory hierarchy design. Scalable parallel computing. Highly parallel systems. Parallel programming models. Communication primitives, techniques for programming and compiling. Fundamental concepts of parallel algorithms. 2D algorithms. Hypercube algorithms. Hypercube architectures. Мessage passing еnvironment (Message Passing Interface-MPI), parallel virtual machine (Parallel Virtual Machine-PVM). Managing the space for data storage. Deadlock. Techniques for synchronization and load balancing.

Literature:

  1. K. Dowd, C. R. Severance, M. Loukides (Editor), "High Performance Computing (Risc Architectures, Optimization & Benchmarks)", O’Reilly, 1998
  2. R. Buyya (Editor), "High Performance Cluster Computing: Architectures and Systems, Vol. 1", Prentice Hall, 1999
  3. F. Berman, G. Fox, T. Hey, (Eds), John, "Grid Computing; Making the Global Infrastructure a Reality", John Wiley, 2003