Looking at the code, it does not appear that particle systems loaded from snippets are cached. I.e. if multiple instances of the same system are used, the snippet will be loaded and parsed each time. Can you please confirm my understanding? Also would you recommend limiting the creation of new instances of particle systems using a pooling strategy? Or is it okay from a perf and resource perspective to simply clone the already loaded instance everytime we need a new instance (attached to a new emitter)?
Cloning might be faster than reloading but will go with the same perf implications (a new system)
That said the overhead is probably amazingly thin knowing the snippet request should be cached by the browser.
Snippet request is cached by browser, but the parsing of the json and the spinning up of the particle system would be happening each time. Also I’m guessing that since PS has a dispose it potentially consumes enough resources that we should get rid of them as soon as we are done using them?
yup or you could recycle them if some setups are similar