3 ms·
In generational GCs there are two or more allocation regions. New objects are put in the "younger" generation, which is garbage collected separated from the oth
by cratermoon 3y ago
In generational GCs there are two or more allocation regions. New objects are put in the "younger" generation, which is garbage collected separated from the other generations. This sort of resolves the issue of tracking individual object lifetimes by having all the short-lived objects subject to rapid GC. This means most of the effort tracking lifetimes is reserved for the fewer long-lived objects.
- lll-o-lll 3y agoIt’s “medium term” objects that cause all the trouble. Async operations. That’s why .net invested heavily into zero-alloc tasks, as the gc would kill scaling in async heavy code.