PDC 2009 was an exciting event, with announcements about Azure, Silverlight 4, and Office 2010 popping up one after another. For me, there was another reason why this year’s PDC was exciting – it was my first chance to present in a major conference.
Concurrency
Most Common Performance Issues in Parallel Programs
Performance of parallel programs is an interesting – but also tricky – issue. I put together an article for our team blog that talks about the most common reasons why a parallel program may not scale as desired: Developers ask why one program shows a parallel speedup but another one does not, or how to … Read more
Big Oh in the parallel world
Big-Oh notation is a simple and powerful way to express how running time of a particular algorithm depends on the size of the input. When you say that a particular algorithm runs in O(N2) time, you mean that the number of steps the algorithm takes is proportional to the input size squared. Or, in mathematical terms, there is some fixed constant C, such that to process input of size N, the algorithm needs at most C x N2 steps. One interesting question is how to define a “step”. The beauty of the Big-Oh notation is that any somewhat reasonable definition of a step will do. The step could be a clock cycle, a hardware instruction, or an expression in source code. If an algorithm takes O(N) steps according to one definition of a “step”, it takes O(N) according to all reasonable definitions.
While this is great, you already probably heard it a thousand times. But, what about the complexity of parallel programs? A major departure from the sequential case is that the number of steps an algorithm takes and the actual real-world time may not be proportional. In the parallel case, we may have potentially many cores executing the computation steps!
Overview of concurrency in .NET Framework 3.5
There is a lot of information on the concurrent primitives and concepts exposed by the .NET Framework 3.5 available on MSDN, blogs, and other websites. The goal of this post is to distill the information into an easy-to-digest high-level summary: what are the different pieces, where they differ and how they relate. If you want to know the difference between a Thread and a BackgroundWorker, or what is the point of interlocked operations, you are reading the right article.