Open lemire opened 6 years ago
Do flyweight iterators currently outperform the standard iterators? The benchmarks suggest not.
Do we have a benchmark stressing garbage collection? There are definitively reports that reducing GC pressure helps.
I admit that I don’t know how we would make GC the bottleneck.
Would you be interested in a test tracking the quantity of heap allocated ? It would not check actual GC pressure (as for the same quantity of allocations, it would not track the lifetime of given memory), but it may provide valuable insights.
@blacelle
Anything that might help motivate the flyweight pattern is welcome.
We have fast bulk iterators...
See PR https://github.com/RoaringBitmap/RoaringBitmap/pull/243
We'd like to support them using the flyweight model.