Configurable Enter Play Mode

I tried this out on an empty project using nothing but the entities package. It lowered the enter play time from 4s to 0.3s on my machine when not having made any code changes. Which is amazing!

However! It does not seem to have any impact of the complete iteration time of making a code change, switching to the editor and entering play mode. Around 5 seconds of editor freeze regardless what option are used for me (In realistic projects this is of course way longer). Considering how frequent we make code changes and switch to the editor, the configurable play mode is barely scratching the surface in reducing the complete iteration time.

I would very much like to see the diagram covering complete iteration time. Obviously recompilation is a varying factor but I can assure you it is not the bottleneck.

Try a URP project, remove all example assets, install entities package (possibly also physics). Take a dummy script and change a number 5 to a number 6 and measure the complete time till you can enter play.

I take the opportunity in making a shameless plug for my proposal of a ligthweight runner rewritten for programmers. Where neither domain reloading nor freezing the editor would even be necessary because the runner/editor should not run the game in the same process.

1 Like