For search results we create graphic layers dynamically. This allows the user to have various search results open and control them (visibility, filtering, transparency) individually. The results range from 1-1000. We also allow the user to filter the graphics being shown in the graphics layers by attributes, which would add an but more complexity to the issue. I believe this would mean I would have to double the amount of graphics layers (1 for the main layer and 1 for the text) and then manager the filtering/ rendering of both simultaneously.
Short answer is no I have not implemented it yet.
I am planning on giving it a whirl this week but I was looking to see what type of luck people have had with labeling graphics and if they had any tips/ tricks.