Yeah, it takes forever to run on my machine (two or three hours for very large programs), and once it skipped the main section of my program, and I never did get it to start profiling it. It was a click event of a button called an external piece of code, and it showed up in the coverage file, but when I ran the profiler, it didn't track it.

So, any way, I mainly stick with the debug window. Most of my programs are of the "conversion" sort, wherein I loop an obscene number of times. If I cut down the size of the dataset, then the profiler works nicely. But cutting down the data set isn't always an option.

>Yes. I agree. Also for checking whether some functions are getting called an obscene number of times. Have you used the coverage profiler yet in 6.0?
>
>>I'll often run my code once all the bugs are out, and look for places where it seems like the little pointer pauses. That tells me that I might have an optimization problem there.