I'm wondering why if I run that code in OpenModelica 1.5.0 and then in  Dymola 7.4 (with visual studio 2008)
I have different results

model tentative
Real a(start=0);
Real T=0.1;
Real a_steady(start=0);

algorithm
    a_steady:=a_steady+1;
equation
  der(a)=(a_steady-a)/T;
end tentative;

with the same simulation parameter

with openmodelica a_steady fast grows for each time step of the simulation, instead with dymola a_steady remains constantly to 1, and  a has a few growing.
instead in dymola the system behaves like a_steady=1 for all the simulation.

can anyone help me to understand, please?
what is it supposed to be the normal behavior? the one of dymola or the one of openmodelica?
It can be the compiler, isn't it?
c'est werid for me.

thanks for your time

have good day

here are the pictures
http://www.mypicx.com/09032010/modelica/
I've some problems to attach them I don't know the reason