I'm running some C code on the device and trying to figure out what I can do to make it faster. Unfortunately, I'm having trouble figuring out how to get the information I need from the tools available.
In this case, 99.7% of the time spent is in "hidden (pdpathfinder.c:568". Is this part of the API? Is there a way to tell where I'm calling this code from? Something like stack sampling would be really nice or, failing that, some of the source for the internal libraries would really help with debugging / profiling.
The Simulator has a Sampler window, which can be set to measure Device instead of itself. The reported functions have disclosure triangles you can click to drill down, with code line numbers. Is that what you need?
In the Mac Sim (only) if you control click the Sample button you can select the game's elf file. That will allow game symbols to show up in the sample. I am currently working on making this more obvious and also more useful by showing game and OS symbols together, among other things. I'm also adding game elf support to the Win/Linux sim. This will hopefully land in 1.13.1.
Thanks for this info! I tried this but I'm not sure it worked. I'm on an M1 Mac though so that may be why. I'll try on an intel device.
It would be great to have more sampling opportunities. I noticed that there's no JTAG or SWD header on the board. Are you guys willing to share the pinout of the test pads? I would definitely enjoy tracing my code to figure out how to make things go faster.
Sorry, we have RDP enabled on the device so that people can't just connect and dump the firmware. We do need to try and keep some stuff protected for Catalog, and we're dreading the day China makes a Playdate clone that runs our code. Maybe it's overly paranoid, but we've invested a lot of time, energy, and money into Playdate and it's unsettling to think that our work could get sniped like that.