Runs with many iterations and invocations lead to memory issues #132
Description
Similar to what we had with #100, there's still an issue when I am running -it 300 -in 3000. It accumulates quite a bit of memory, and I think, the measurements are not reported to ReBenchDB after each iteration either.
Perhaps there's a way to flush the data to ReBenchDB based on some sensible heuristic, i.e., time passed and memory used.