Last week I ported Iain Lobb’s Bunnymark performance test to Corona, as I was curious to see how it would perform. I posted the results on Twitter, but thought that I’d follow up with a post showing the Bunnymark code, and also talk about how the results have changed after testing again using the graphics optimizations recently announced by Ansca.
First off, I’ve posted the basic Bunnymark code to GitHub. Included in it is a simple FPS counter that I put together. If you look at the code for the counter, you’ll see that you can specify a ‘target’ frame rate, and the counter will dispatch an event when the frame rate drops below that target. This makes it a bit easier to stop any particular code from running if the FPS drops too low. Also, to account for any temporary dips in the frame rate (e.g. by creating a couple hundred display objects in a single frame) it tracks the frame rate as an average of the rate over the last 60 frames.