Precise Simulation wrote
This is normal and expected as the Matlab runtime is single threaded.
I see. And the solver is not executing in some cancel-able event ....
I know nothing about windows signals, but thinking in c/unix terms ... could the SIGTERM handler be replaced while the solver is running and use something like setjmp()/longjump() to save the state before calling the solver and to return to that state from the SIGTERM handler (signaled by "Cancel") after restoring the signal state?
Sorry... I shouldn't be guessing how things are implemented ... just wondering...
I suppose as a work-around, I can save to a temp.fea file before each solve attempt so that if I had to kill the process I could return to where I was after restarting MATLAB/FEATool...
Perhaps FEATool could do that as well; if so, then I could post that as an enhancement request.
Anyway, thanks for the answer...
BTW, I have observed that a grid generation cancellation has worked for me.
Kind regards,
Randal