a dverk performance issue
Posted: February 27 2017
Hi everybody.
I have been working with a model where the equations of motion for the background can't be solved numerically in CAMB, but they can be solved using mathematica or python. Briefly the situation is the following: I have a system of three ODE y'(1), y'(2), y'(3) whose initial conditions are set by theoretical considerations. Both y(2) and y(3) undergo an initial sudden drop (of eight and 23 orders of magnitude, respectively) and then they evolve smoothly. Whereas this situation does not pose any problem in mathematica or python, I have found that dverk is unable to handle this drop and for that reason the program crashes.
So my question is if there are other ODE solvers that can be incorporated to CAMB, besides dverk. If so, any hint or a link where somebody has done something like that would be appreciated.
Regards
I have been working with a model where the equations of motion for the background can't be solved numerically in CAMB, but they can be solved using mathematica or python. Briefly the situation is the following: I have a system of three ODE y'(1), y'(2), y'(3) whose initial conditions are set by theoretical considerations. Both y(2) and y(3) undergo an initial sudden drop (of eight and 23 orders of magnitude, respectively) and then they evolve smoothly. Whereas this situation does not pose any problem in mathematica or python, I have found that dverk is unable to handle this drop and for that reason the program crashes.
So my question is if there are other ODE solvers that can be incorporated to CAMB, besides dverk. If so, any hint or a link where somebody has done something like that would be appreciated.
Regards