Re: [deal.II] Re: Speeding up characteristic interpolation

2017-09-05 Thread Martin Kronbichler
Dear Praveen, Regarding the performance of your code, there are several fundamental problems: - You run the constructor of Quadrature (allocates memory) and FEValues (allocates lots of memory, evaluates a number of things you don't need) that are both not made for use within the scenario of

Re: [deal.II] Re: Speeding up characteristic interpolation

2017-09-05 Thread Bruno Turcksin
Praveen, 2017-09-05 9:18 GMT-04:00 Praveen C : > I have set the correct cell by call to set_active_cell. Does the value > function again check if the point is in this cell ? Probably not. > I implemented my own function to evaluate the solution as follows, but this > is

Re: [deal.II] Re: Speeding up characteristic interpolation

2017-09-05 Thread Praveen C
Hello Bruno > Are you sure that FEFieldFunction.value() is the slowest part? In 3D, > point_inside could also be slow (you may need to do a few Newton Iterations). > It would be nice if you ran your code using a profiler to see what is the > bottleneck (building the quadrature in value(p1),