Processor design has reached it's physical limits in 2005.
Only with parallel system can progress be achieved.
Creating correct and efficient parallel programs can be troublesome if not done right.
Since 1995 i am working with parallel and distributed systems.
-
Thread programming with shared memory
-
Using multicore processors and networks.
- C/C++: OpenMP, POSIX Threads
- Java 5/Java 7:
java.util.concurrent
-
GPU Computing
-
Graphic cards: 10x faster than CPU, because massively parallel
- NVIDIA CUDA and OpenCL
- Platforms: C/C++, Java/JVM, .NET
-
Software Transactional Memory (STM)
-
Akka, Haskell
-
Actor models
-
Scala, Akka
-
Theoretical background
-
-
PRAM, P-Completeness
-
Parallel, concurrent and distributed algorithms
