Typed DataSet's vs. Custom Collections

I have been reading the blog entries (here, here) lately about Typed DataSets and it makes me wonder if anyone has really sat down and compared the performance of a custom object/collection (using a data reader to populate) to a Typed DataSet?  Performance was critical to us so we decided to go the custom collection route and we were very happy with the results, but we never actually sat down and figured out what the performance difference was.  We assumed that considering all the bells and whistles that are associated with DataSets that custom collections would be a major performance win.  And since we were using CodeSmith to generate those custom collections, it wasn't any more work to use them.  Does anyone know of any benchmarks that I could look at?

4 Comments

Comments have been disabled for this content.