You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As #26 shows there is no clear correlation between number of inserted elements and performance of std::vector::push_back.
This is not the case for deferred_vector. More elements inserted - more time it will take to insert another one. For example, when there are 25 (cell A2) elements inserted, it is only 36 (cell C2) times slower than insertion into vector<scoped_ptr>, but if there are 4975 (cell A200) elements, it is 1498 (cell C200) times slower.
Inserting 1 million elements into deferred_vector will probably take N days.
The text was updated successfully, but these errors were encountered:
Right, this is a proof of concept at smaller scale and not a production-quality scalable library. I invite people who write scalable GC arenas to use this interface as a starting point to experiment with a C++ opt-in GC heap.
As #26 shows there is no clear correlation between number of inserted elements and performance of std::vector::push_back.
This is not the case for deferred_vector. More elements inserted - more time it will take to insert another one. For example, when there are 25 (cell A2) elements inserted, it is only 36 (cell C2) times slower than insertion into vector<scoped_ptr>, but if there are 4975 (cell A200) elements, it is 1498 (cell C200) times slower.
Inserting 1 million elements into deferred_vector will probably take N days.
The text was updated successfully, but these errors were encountered: