Registered Member
|
Hi
In my application I use a lot of small Eigen::MatrixXd structures (400,000 or more 50x50 matrices) in a std::vector. My problem is that deallocation of this vector consumes a lot of time: 25% of total running time for example. Since Eigen::MatrixXd does not accept custom allocators, I ruled out any memory pool implementations. Do you have any suggestions? Thanks. |
Registered users: Bing [Bot], Google [Bot], Yahoo [Bot]