Consider allocating more aggressively during different phases of the VM

[dnfield]

For example, during startup, allocate more aggressively until some point where the program "settles", e.g. in a Flutter application has started drawing frames.

@rmacnak-google @a-siva

[a-siva]

For this I believe we could consider the RAIL ( Response, Animation, Idle, Loading) model that v8 GC adopts.
The Engine would have to provide notification on the different phases (like it currently does for I (idle)) and the VM could adapt some of the GC heuristics accordingly.

[iskakaushik]

cc: @iskakaushik

[fzyzcjy]

👀 Is there any updates

[a-siva]

@dnfield @iskakaushik Is there any more work to be done on this ?

[dnfield]

What's the status of it for startup? I'm only up to date on the work that was done for throughput during animation.