I am a newbie to Fortran. Please look at the code below:
c main program
call foo(2)
print*, 2
stop
end
subroutine foo(x)
x = x + 1
return
end
In some implementations of Fortran IV, the above code would print a 3. Why is that? Can you suggest an explanation?
How do you suppose more recent Fortran implementations get around the problem?
Help is very much appreciated. Thank You.
The program breaks the language rules - the dummy argument x in the subroutine is modified via the line x = x + 1, but it is associated with something that is an expression (a simple constant). In general, values that result from expressions cannot be modified.
That specific code is still syntactically valid Fortran 2008. It remains a programming error in Fortran 2008 - as it was in Fortran IV/66. This isn't something that compilers are required to diagnose. Some may, perhaps with additional debugging options, and perhaps not till runtime.
Because the program breaks the language rules anything could happen when you run the program. Exactly what depends on the code generated by the compiler. Compilers may have set aside modifiable storage for the value that results from the expression such that it internally looks like a variable (the program might print three and the program carries on), that modifiable storage might be shared across the program for other instances of the constant 2 (suddenly the value of 2 becomes three everywhere!), the storage for the value of the constant might in non-modifiable memory (the program may crash), the compiler may issue an error message, the program may get upset and sulk in its bedroom, the program might declare war on a neighbouring nation - it is a programming error - what happens is unspecified.
As of Fortran 90, facilities were introduced into the language to allow programmers to write new code that is practical for compilers to check for errors such as these (and in some cases compilers are required to check for errors if they are to be regarded as standard conforming).
For the code as presented, the main program and the subroutine are to be regarded as separately compiled - the main program is unaware of the details of the subroutine and vice versa (it is possible that the subroutine could be compiled long after the main program, on a different machine, with the outputs of the two being linked together at some later stage - without fancy link time behaviour or static analysis it is therefore not possible to resolve errors such as this). Language rules are such that when compiling the main program the compiler must implicitly assume the details of the interface of the subroutine based only on the way the subroutine is referenced - inside the main program the subroutine has an implicit interface.
Fortran 90 introduced the concept of an explicit interface, where the compiler is explicitly told what the interface of the subroutine in various ways, and can then check that any reference to the subroutine is consistent with that interface. If a procedure is a module procedure, internal procedure or intrinsic procedure - that interface is automatically realized, alternatively for external subprograms, procedure pointers, etc, the programmer can explicitly describe the interface using an interface block.
In addition, Fortran 90 introduced the intent attribute - a characteristic of a dummy argument of a procedure that is also then a characteristic of the interface for a procedure. The intent of the argument indicates to the compiler whether the procedure may define the argument (it also may implications for default initialization and component allocation status) and hence whether an expression could be a valid actual argument. x in subroutine foo would typically be declared INTENT(INOUT).
Collectively these new language features provide a robust defence against this sort of programming error when using compilers with a basic level of implementation quality. If you are starting with the language then it is recommended that these new features become part of your standard approach - i.e. use implicit none, all procedures should generally be module procedures or internal procedures, use external procedures only when absolutely required, always specify dummy argument intent, use free form source.
Related
To use a subroutine and allow the compiler to check
argument mismatch, one needs to place the defintion of that subroutine in a module, for which case, fortan creates an explicit interface for the calling unit to check argumetn mismatch etc.
For stand-alone subroutines that are not contained in a module, we say they have only "implicit interfaces" and no "explicit interfaces", so that the compiler can not check argument mismatch.
Why does not the compiler also create "explicit interfaces" for the stand-alone subroutines? What difficulties prevent the compiler from doing this?
When we are considering interfaces, we are not talking about some mythical object which may be implicit or explicit.
Consider the two main programs:
program main1
implicit none (external)
external sub
call sub
end program main1
and
program main2
implicit none (external)
interface
subroutine sub()
end subroutine sub
end interface
call sub
end program main2
These two main programs are (perhaps) concerned about the same external subroutine sub. main1 has an implicit interface for sub and main2 has an explicit interface for sub.
Whichever main program we are wondering about, the subroutine sub always has an explicit interface for itself:
subroutine sub()
implicit none (external)
call sub
end subroutine sub
I've just defined a (pretty terrible) subroutine called sub. I imagine that's going to be a popular name: next time you want to use that name (don't forget: external procedures are global entities) for an external procedure are you going to use the one I've just defined? If you aren't, how are you going to tell the compiler which sub you are talking about?
Resolution of which external subroutine sub is going to be used is often a problem for the linker, not the compiler. The compiler may never see a particular external procedure so cannot automatically generate for a program unit an explicit interface for an external procedure.1
That said, there are cases where a compiler will do some work for you:
program main3
implicit none (external)
external sub
call sub
end program main3
subroutine sub(x)
implicit none (external)
real x
end subroutine sub
(as a single file, or separate files, perhaps with specific compiling options) may well prompt your compiler to complain, just as though sub had an explicit interface inside main3.
1 An external procedure can also be defined by means other than Fortran. In such a case, the Fortran compiler doesn't need to even understand it, let alone worry about having to generate an explicit interface for it.
It simply the way the language works and have always worked since the inception in the 1950s and first standardization in 1960s.
It had little other choice back then, even though some other languages like Algol 68, or later Modula, pioneered the use of modules a decade or more later. Pascal, for example, relied on nesting.
Mind you, modules are a relatively new thing. C++ got them only a few years ago. Otherwise C (and C++) used to include files to allow the compiler to see headers from other source files by including everything into the current file. But implicit interfaces were also an option in original C.
In Fortran, the compilation units are individual program units, not source files. One can write or even automatically generate (e.g. the -gen-interfaces compiler option mentioned in the comments) interface blocks and include them using include files, but this approach is rarely recommendable.
Even though explicit interfaces would theoretically theoretically be possible without modules or included headers, such a concept does not really exist much in the programming world for traditional compiled languages.
The concept of explicit interface was introduced in Fortran 1990. It allows some new ways of calling more advanced forms of Fortran subprograms (array results, assumed shape, optional arguments,...). However it also allows checks that were not possible before. Enforcing such checks would break existing code. The success (and its lack) is to a large extent also based on backward compatibility. You cannot brake existing code so much - it would really be a lot.
Compilers will now often warn you that you are doing something not allowed, that the call of an external subprogram is not conforming, if you allow such warnings and if you enable or do not disable such warnings. But it will always remain a warning and it does not bring you the new possibilities. Legacy codes remain working.
These warnings are also possible at the linking stage, especially when link-time optimizations are used.
When dealing with optional arguments in fortran I believe it's typical to branch using the present() intrinsic, i.e.:
subroutine foo(ii,jj)
implicit none
integer, intent(in) :: ii
integer, optional :: jj
if (present(jj)) then
! do something
else
! do something else
end if
end subroutine foo
My assumption (coming from a C++ world) is that present() is hopefully a compile-time construct, and that there won't be any associated runtime performance penalty. I expect that the compiler should be able to (should be required to?) optimise the if statement shown above away, depending on whether foo(ii) or foo(ii,jj) is called.
How is the present() intrinsic handled by compilers in practice? Does the fortran spec guarantee certain behaviour?
PRESENT is a runtime concept - to the extent that the presence of an argument may depend on aspects of the program that cannot be determined until runtime (based on input read from a file or similar).
The implementation of the equivalent of PRESENT in a C++ program would similarly be a runtime concept. Note that C++'s defaulted arguments feature is not quite the same concept as Fortran's optional arguments feature (though it could be part of an equivalent implementation).
Beyond compiler optimization cleverness, if you want compile time resolution of things similar to C++'s defaulted arguments or function overloading, then consider using multiple specific procedures behind a generic name, one variant of the specific with the "optional" argument, the other without.
That depends (what else did you expect? :-)
If you use gfortran with a recent version, and do a lot of work in at least one of the branches, and use LTO or put everything into one file, the compiler will clone the function for you (via constant propagation). Otherwise, probably not. If you want to find out if the procedure foo has been cloned, grep for foo.*constprop in the assembly file.
An alternative to one subroutine with an optional argument is two subroutines, one with the argument, and one without, sharing the same interface. I don't think there is a run-time cost for doing this as opposed to giving the two versions of the subroutine different names. Schematically, it looks like this:
interface foo
module procedure foo_1,foo_2
end interface foo
contains
subroutine foo_1(ii)
! some code
subroutine foo_2(ii,jj)
! some code
Am working with some legacy FORTRAN code. The author defined a function (not a subroutine, but a function -- that's going to be important) called REDUCE_VEC(). It accepts a 1D array and returns a scalar real*8. So, if you want to "reduce" your vector, you make a call to the function
RV = REDUCE_VEC(V1)
and everything is fine. But occasionally, he has lines that look like
CALL REDUCE_VEC(V2)
So, two questions: 1) What the heck would this second form of the call do? (Note that there is no way to return data.) 2) This won't even compile under gfortran, even though it did with PGI, so is this even legal FORTRAN?
Thanks.
This will compile with many processors if the interface is implicit, because the compiler cannot check that, it just calls some symbol. Consider the following:
function f(a)
dimension a(*)
f = 0
do i=1,10
f = f + a(i)
end do
end function
program p
call f([1.,2.,3.,4.,5.,6.,7.,8.,9.,10.])
end program
Compiles and does not even crash immediately with ifort, whereas sunf90 and gfortran will compile that only if it is in separate source files and then the result also does not crash on my machine. If the return value is placed in register, it may cause no harm to the rest of the program, but a stack corruption is quite likely otherwise.
It is not legal Fortran. As presented it is more than likely a programming error (it is possible for the same name in different scopes to refer to different things, but that's not what is implied by the question). If the Fortran processor happens to support an extension to the language that allows this, then what happens is up to the Fortran processor. Otherwise, "anything" could happen, where "anything" could include (but is not limited to) "nothing", or "very, very bad things".
from the source code published in the programmer's manual of a commercial program, I have isolated a code snippet which puzzles me quite a lot.
The function below is expected to be called multiple times by a kernel and is supposed to implement the temporal behaviour of a component in a system consisting of many interconnected components (I have removed the Input/Output parameters from the function prototype because they are irrelevant to the point I intend to rise).
To distinguish between different instances of the same block type the kernel pass an instance number in the INFO(1) element.
As far as I have understood, the designer of this program took a great deal of effort trying to save the time spent in copying the values of the parameters from the PAR vector to local variables with meaningful names at each call (as if they were not aware of the optimizations a compiler can do). It seems to me that they wanted to assign them to the local variables only in the first call, or when the caller switch to a different instance of the same type.
However I can't understand how this could work if the local variables are not declared static with the "save" keyword. Does FORTRAN store local variables statically i.e. not on a stack? (I am sorry if the question sounds stupid, I am used to the C/C++ languages)
Thank you.
SUBROUTINE TYPE151(PAR, INFO, *)
IMPLICIT NONE
INTEGER*4 INFO(15), IUNIT
DOUBLE PRECISION PAR, QMAX
PARAMETER (NP=1)
DIMENSION PAR(NP)
! First call
IF (INFO(7).EQ.-1) THEN
IUNIT = INFO(1)
QMAX = PAR(1)
RETURN 1
ENDIF
! later calls
IF(INFO(1).NE.IUNIT) THEN
IUNIT = INFO(1)
QMAX = PAR(1)
ENDIF
! Making use of QMAX in some ways...
RETURN 1
END SUBROUTINE TYPE151
Storage methods are not part of the language standard. Old FORTRAN compilers (FORTRAN 77 and earlier) frequently stored all variables statically. The language requires that you use "SAVE" for variables for which the values should be retained across calls to the procedure. But many programmers ignored this requirement and relied on the behavior that all variables retained their values because of the typical design of compilers in the FORTRAN 77 era.
Modern Fortran compilers typically use memory differently and local variables of procedures do not always retain their values if SAVE is omitted. This frequently causes bugs when old programs are compiled with current compilers. Compilers typically provide an option to restore the old behavior. Otherwise it could be a great deal of work to identify all variables in a large legacy program that needed to have the SAVE attribute added to their declaration.
In my Fortran code I made the following call to the dnrm2 routine:
d = dnrm2(n, ax, 1)
Just a simple call that would return me a double precision result.
The question is, should I declare the function at the start of my script? I found that if I don't declare it, when I compile the code in 32 bit Windows, then the result is correct.
But if I compile the code in 64 bit Windows, then the result isn't be correct.
Why is this so? Must an external routine always be declared in Fortran?
If you don't correctly describe your subprograms (subroutines and functions) to a calling program, the compiler may not correctly call them. Fortran compiles each unit separately, so the compiler doesn't "know", by default, about the characteristics of other subprograms. There are several ways that you can describe/declare a subprogram in Fortran 90/95/2003.
The easiest and best method is to place your subprograms into a module and then "use" that module in the calling program. This automatically makes the interface known to the compiler and will enable the compiler to check the consistency of actual arguments (in the call) and dummy arguments in the subprogram. It will also make known the return type of a function. The various subprograms in a module have their interfaces known to each other.
You can also write an "interface" containing a subprogram declaration that matches the declarations of the actual subprogram. (This method can be very similar to the style of including header files in C.) This method is more work and prone to error because you have to manually maintain consistency between the actual subprogram and interface whenever changes are made. The interface method is useful when you don't have the code to the subprogram or the subprogram is written in a language other than Fortran.
Or you can simply declare a function name to specify the type-return of the function, but this won't give you any checking of the arguments. In my opinion this method is weaker since having the compiler check argument consistency eliminates a major class of programming mistakes.
I don't do Fortran, but in C, the size of a pointer and the size of a long int varies between 32 and 64 bit OS'es, but the size of an int does not. Perhaps the program is using ints to do pointer arithmetic?