I've been known to dabble in the pranksters' art from time to time:-D

True, Normally I wouldn't do such things in a benchmark; I feel benchmarks like this should only perform the work necessary to achieve the same end result, and not equivalent processing. I massaged the code a little and ran it under DevPartner Profiler to obtain (for my actual scenario) more accurate results. Thanks and Regards

Good Idea and something I thought about myself. Although, the code that I am working with at the moment does *allot* of small rectangle adjustments during the drawing and I am a bit worried about what the performance is going to look like under .NET. I am currently starting to re-think things, and either I am going to change the drawing logic (I doubt) and/or, to fall back and to code up a rectangle variant that allows L,T,R,B adjustments and translate those into X,Y,W,H adjustments *auto-magically*. Regards