Improved user interface speed

Before a plot can be displayed or an animation started, mesh data and results need to be loaded into memory and converted. In addition, further data structure needs to be built and stored to generate cutting planes and iso-surfaces, which are involved in each minimum/maximum animation frame. If cutting planes exist, the display on the cutting planes also needs to be updated for each frame. For large models, this means that an animation build up can take time.

The post-processing speed of the user interface has been increased by 2-3 times, due to the implementation of a special map that retrieves data from an index with the same speed as from an array. Historically, one of the major bottlenecks for data loading has been the building of the two-way maps between mesh nodes and elements and vtk points or cells. These maps are needed for results query and display, but data retrieval is slow and memory inefficient. With the new implementation, post processing loading, animation, and cutting plane speed has be improved significantly.