.NET Matters: False Sharing in Parallel Programming

There’s no programming model that magically eliminates all concurrency-related issues, so understanding how memory and caches work is still important to write efficient parallel programs.

In the October 2008 issue of MSDN Magazine, Stephen Toub, Igor Ostrovsky, and Huseyin Yildiz explain how architectural limitations of memory cache can lead to incorrectly sharing data between threads and demonstrate how you can identify and correct such “false sharing” in your parallel applications.

For more .NET programming guidance from Steve, check out previous .NET Matters installments in our MSDN Magazine archives.