Does anyone know what happens to shared memory via fork when you start invading swap? My assumption is that it will swap out OS pages which will cause a CoW fault and any shared memory will copied for each process.
@Fryguy I am thinking our testing of Aaron's compact-gc branch was flawed. I believe we were compacting before each fork. To truly see if it helps, we'd need to eager load more core code and compact only once before the first fork, not each fork