fortran return a different value respect to the one expected

[afylot1]

I checked with a debugger that a value of a variable(the name of the function) is computed, but then a different value is returned. I'd think this is related to type translation, but I don't see how. I have the following function. double precision function func_energy(temp) use func_params implicit none double precision energy_f_temp, temp func_energy = energy_f_temp(temp, density) - p_invert return end function func_energy func_params has been declared as the following module func_params double precision, save :: p_invert double precision, save :: density end module func_params I checked that the initialization of func_params is correct. When I am inside func_energy I found that energy_f_temp(temp, density) - p_invert is equal to -42476827268725768, but then when I check at the next line the value of func_energy I get 588030290385264, and this is the value that is returned by the function. I don't understand what is going on. It looks to me that everything is double precision. What could be going wrong here? I'm using ifort with compiler option: -g -O0 -check bounds -warn all -traceback -align all -align rec8byte These functions are part of a fortran library that I am calling from a C main program. I have read somewhere that when the calling convention are not right you could be corrupting the stack. Does it apply here? The point is that the error appeared suddently. thanks

[Steven_L_Intel1]

Calling convention mismatch does not apply here. How is energy_f_temp declared? If it is not also double precision, you could have this problem.

Without a complete source we can build and run, it's very difficult to diagnose the problem. I'd suggest putting energy_f_temp and any other functions you call in a module and USEing that module, instead of relying on implicit interfaces., to make sure that everything is ok. -warn all will give you generated interface checking, but if you compiled the called routine later it would not pick that up.