How did you measure startup time?
David Air


We measured dyld performance by building and running a debug build of our app with the DYLD_PRINT_STATISTICS environment variable defined after a reboot of the device (to clear out the dyld cache). This allowed us to measure the relative performance of all steps of our app’s launch. However, we noted dramatic improvements in launch times outside of just this scenario—release and enterprise builds both launched noticeably faster after this change as well. While it’s quite likely that the difference in launch times will be smaller outside of debug builds, the most important aspect to us was the relative (2x) difference, which should occur regardless of build type. We weren’t able to measure any other scenarios though, as we couldn’t find a way to do so. If you know of a way to measure dyld performance outside of the context of a debug build, please share it!

Like what you read? Give Eric Horacek a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.