Регистрация слушателей Открытого курса лекций Георги Гайдаджиева на английском языке
Программа курса:
1. Accelerators, Stream Processing and CUDA, Kernel-Based Data Parallel Execution Model.
2. Memory Model, Locality and Bandwidths, Atomic operations, Privatization.
3. Barrier Synchronization, Algorithmic and Execution overheads, Double-buffering and work efficiency analysis.
4. Overlapping execution with data movements and in-depth Numeric considerations.