美文网首页
2018-04-16

2018-04-16

作者: boliecon | 来源:发表于2018-04-17 04:38 被阅读0次

    Fortran Best Practices

    ======================

    .. highlight:: fortran

    This page collects a modern canonical way of doing things in Fortran. It is meant to be short, and it is assumed that you already know how to program in other languages (like Python, C/C++, ...) and also know Fortran syntax a bit. Some things in Fortran are obsolete, so this guide only shows the "one correct/canonical modern way" how to do things.

    Summary of the language: http://www.cs.umbc.edu/~squire/fortranclass/summary.shtml

    Language features are explained at: http://en.wikipedia.org/wiki/Fortran_language_features

    The best resource is a recent Fortran standard, for example the Fortran 2003 standard: http://www.j3-fortran.org/doc/year/04/04-007.pdf

    Fortran Style Guide

    -------------------

    Naming Convention

    ~~~~~~~~~~~~~~~~~

    Ultimately this is a matter of preference. Here is a style guide that we like

    and that seems to be prevalent in most scientific codes (as well as the Fortran

    standard library) and you are welcome to follow it.

    1. Use lowercase for all Fortran constructs (``do``, ``subroutine``, ``module``,

      ...).

    2. Follow short mathematical notation for mathematical variables/functions

      (``Ylm``, ``Gamma``, ``gamma``, ``Enl``, ``Rnl``, ...).

    3. For other names use all lowercase: try to keep names to one or two

      syllables; if more are required, use underscores to clarify (``sortpair``,

      ``whitechar``, ``meshexp``, ``numstrings``, ``linspace``, ``meshgrid``,

      ``argsort``, ``spline``, ``spline_interp``, ``spline_interpolate``,

      ``stoperr``, ``stop_error``, ``meshexp_der``).

    For example "spline interpolation" can be shortened to

    ``spline_interpolation``, ``spline_interpolate``, ``spline_interp``,

    ``spline``, but not to ``splineint`` ("int" could mean integration, integer,

    etc. --- too much ambiguity, even in the clear context of a computational

    code). This is in contrast to ``get_argument()`` where ``getarg()`` is

    perfectly clean and clear.

    The above are general guidelines.  In general, choosing the right name

    certainly depends on the word being truncated as to whether the first syllable

    is sufficient. Usually it is but clearly not always. Thus some thought should

    go into step "try to keep names to 2 syllables or less" since it can really

    affect the indicativeness and simplicity. Simple consistent naming rules are a

    real help in this regard -- for both collaboration and for one's own sanity

    when going back to some old code you haven't seen in while.

    Indentation

    ~~~~~~~~~~~

    Use 4 spaces indentation (this is again a matter of preference as

    some people prefer 2 or 3 spaces).

    Comparison to Other Languages

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    On the other hand, in most of the rest of the programming world, where the main

    focus is, in one form or another, on defining and using large sets of complex

    objects, with tons of properties and behaviors, known only in the code in which

    they are defined (as opposed to defined by the same  notation throughout the

    literature), it makes more sense to use longer, more descriptive names. The

    naming conventions one sees used in more general-purpose languages such as C++

    and Python, therefore, are perfectly consistent with their more general-purpose

    missions. But Fortran has a different mission (numerical scientific computing).

    .. _floating_point_numbers:

    Floating Point Numbers

    ----------------------

    Somewhere create and export a parameter ``dp``::

        integer, parameter:: dp=kind(0.d0)                  ! double precision

    and declare floats as::

        real(dp) :: a, b, c

    Always write all floating point constants with the ``_dp`` suffix::

        1.0_dp, 3.5_dp, 1.34e8_dp

    and never any other way (see also the gotcha

    :ref:`floating_point_numbers_gotcha`).

    Omitting the dot in the literal constant is also incorrect.

    To print floating point double precision

    numbers without losing precision, use the ``(es23.16)`` format (see

    http://stackoverflow.com/questions/6118231/why-do-i-need-17-significant-digits-and-not-16-to-represent-a-double/).

    It is safe to assign integers to floating point numbers without losing

    accuracy::

        real(dp) :: a

        a = 3

    In order to impose floating point division (as opposed to integer division

    ``1/2`` equal to ``0``), one can convert the integer to a floating point number

    by::

        real(dp) :: a

        a = real(1, dp) / 2      ! 'a' is equal to 0.5_dp

    Integer Division

    ----------------

    Just like in Python 2.x or C, when doing things like ``(N-1)/N`` where ``N`` is an integer and you want a floating point division, force the compiler to use floats at the right hand side, for example by::

        (N - 1.0_dp)/N

    As long as one of the ``/`` operands is a float, a floating point division will be used.

    .. _modules:

    Modules and Programs

    --------------------

    Only use modules and programs. Always setup a module in the following way::

        module ode1d

        use types, only: dp, pi

        use utils, only: stop_error

        implicit none

        private

        public integrate, normalize, parsefunction, get_val, rk4step, eulerstep, &

                rk4step2, get_midpoints, rk4_integrate, rk4_integrate_inward, &

                rk4_integrate_inward2, rk4_integrate3, rk4_integrate4, &

                rk4_integrate_inward4

        contains

        subroutine get_val(...)

        ...

        end subroutine

        ...

        end module

    The ``implicit none`` statement works for the whole module (so you don't need to worry about it). By keeping the ``private`` empty, all your subroutines/data types will be private to the module by default. Then you export things by putting it into the ``public`` clause.

    Setup programs in the following way::

        program uranium

        use fmesh, only: mesh_exp

        use utils, only: stop_error, dp

        use dft, only: atom

        implicit none

        integer, parameter :: Z = 92

        real(dp), parameter :: r_min = 8e-9_dp, r_max = 50.0_dp, a = 1e7_dp

        ...

        print *, "I am running"

        end program

    Notice the "explicit imports" (using Python terminology) in the ``use`` statements. You can also use "implicit imports" like::

        use fmesh

    But just like in Python, this should be avoided ("explicit is better than implicit") in most cases.

    Arrays

    ------

    When passing arrays in and out of a subroutine/function, use

    the following pattern for 1D arrays (it is called `assumed-shape`)::

        subroutine f(r)

        real(dp), intent(out) :: r(:)

        integer :: n, i

        n = size(r)

        do i = 1, n

            r(i) = 1.0_dp / i**2

        enddo

        end subroutine

    2D arrays::

        subroutine g(A)

        real(dp), intent(in) :: A(:, :)

        ...

        end subroutine

    and call it like this::

        real(dp) :: r(5)

        call f(r)

    No array copying is done above. It has the following

    advantages:

    * the shape and size of the array is passed in automatically

    * the shape is checked at compile time, the size optionally at runtime

    * allows to use strides and all kinds of array

      arithmetic without actually copying any data.

    This should always be your default

    way of passing arrays in and out of subroutines. However

    in the following cases one can (or has to) use `explicit-shape` arrays:

    * returning an array from a function

    * interfacing with C code or legacy Fortran (like Lapack)

    * operating on arbitrary shape array with the given function (however there are

      also other ways to do that, see :ref:`elemental` for more information)

    To use `explicit-shape` arrays, do::

        subroutine f(n, r)

        integer, intent(in) :: n

        real(dp), intent(out) :: r(n)

        integer :: i

        do i = 1, n

            r(i) = 1.0_dp / i**2

        enddo

        end subroutine

    2D arrays::

        subroutine g(m, n, A)

        integer, intent(in) :: m, n

        real(dp), intent(in) :: A(m, n)

        ...

        end subroutine

    and call it like this::

        real(dp) :: r(5)

        call f(size(r), r)

    In order to return an array from a function, do::

        function f(n) result(r)

        integer, intent(in) :: n

        real(dp) :: r(n)

        integer :: i

        do i = 1, n

            r(i) = 1.0_dp / i**2

        enddo

        end function

    If you want to enforce/check the size of the arrays, put at the beginning of

    the function::

        if (size(r) /= 4) stop "Incorrect size of 'r'"

    To initialize an array, do::

        integer :: r(5)

        r = [1, 2, 3, 4, 5]

    This syntax is valid since the Fortran 2003 standard and it is the preferred

    syntax (the old syntax ``r = (/ 1, 2, 3, 4, 5 /)`` should only be used if you

    cannot use Fortran 2003).

    In order for the array to start with different index than 1, do::

        subroutine print_eigenvalues(kappa_min, lam)

        integer, intent(in) :: kappa_min

        real(dp), intent(in) :: lam(kappa_min:)

        integer :: kappa

        do kappa = kappa_min, ubound(lam, 1)

            print *, kappa, lam(kappa)

        end do

        end subroutine

    Multidimensional Arrays

    -----------------------

    Always access slices as ``V(:, 1)``, ``V(:, 2)``, or ``V(:, :, 1)``, e.g. the colons should be on the left. That way the stride is contiguous and it will be fast. So when you need some slice in your algorithm, always setup the array in a way, so that you call it as above. If you put the colon on the right, it will be slow.

    Example::

        dydx = matmul(C(:, :, i), y) ! fast

        dydx = matmul(C(i, :, :), y) ! slow

    In other words, the "fortran storage order" is: smallest/fastest changing/innermost-loop index first, largest/slowest/outermost-loop index last ("Inner-most are left-most."). So the elements of a 3D array ``A(N1,N2,N3)`` are stored, and thus most efficiently accessed, as::

        do i3 = 1, N3

            do i2 = 1, N2

                do i1 = 1, N1

                    A(i1, i2, i3)

                end do

            end do

        end do

    Associated array of vectors would then be most efficiently accessed as::

        do i3 = 1, N3

            do i2 = 1, N2

                A(:, i2, i3)

            end do

        end do

    And associated set of matrices would be most efficiently accessed as::

        do i3 = 1, N3

            A(:, :, i3)

        end do

    Storing/accessing as above then accesses always contiguous blocks of memory, directly adjacent to one another; no skips/strides.

    When not sure, always rewrite (in your head) the algorithm to use strides, for example the first loop would become::

        do i3 = 1, N3

            Ai3 = A(:, :, i3)

            do i2 = 1, N2

                Ai2i3 = Ai3(:, i2)

                do i1 = 1, N1

                    Ai2i3(i1)

                end do

            end do

        end do

    the second loop would become::

        do i3 = 1, N3

            Ai3 = A(:, :, i3)

            do i2 = 1, N2

                Ai3(:, i2)

            end do

        end do

    And then make sure that all the strides are always on the left. Then it will be fast.

    .. _elemental:

    Element-wise Operations on Arrays Using Subroutines/Functions

    -------------------------------------------------------------

    There are three approaches:

    * ``elemental`` subroutines

    * `explicit-shape` arrays

    * implementing the operation for vectors and write simple wrapper subroutines

      (that use ``reshape`` internally) for each array shape

    In the first approach,

    one uses the ``elemental`` keyword to create a function like this::

        real(dp) elemental function nroot(n, x) result(y)

        integer, intent(in) :: n

        real(dp), intent(in) :: x

        y = x**(1._dp / n)

        end function

    All arguments (in and out) must be scalars. You can then use

    this function with arrays of any (compatible) shape, for example::

        print *, nroot(2, 9._dp)

        print *, nroot(2, [1._dp, 4._dp, 9._dp, 10._dp])

        print *, nroot(2, reshape([1._dp, 4._dp, 9._dp, 10._dp], [2, 2]))

        print *, nroot([2, 3, 4, 5], [1._dp, 4._dp, 9._dp, 10._dp])

        print *, nroot([2, 3, 4, 5], 4._dp)

    The output will be::

      3.0000000000000000

      1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

      1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

      1.0000000000000000        1.5874010519681994        1.7320508075688772        1.5848931924611136

      2.0000000000000000        1.5874010519681994        1.4142135623730951        1.3195079107728942

    In the above, typically ``n`` is a parameter and ``x`` is the array of an

    arbitrary shape, but as you can see, Fortran does not care as long as the final

    operation makes sense (if one argument is an array, then the other arguments

    must be either arrays of the same shape or scalars). If it does not, you will

    get a compiler error.

    The ``elemental`` keyword implies the ``pure`` keyword, so the subroutine must

    be pure (can only use ``pure`` subroutines and have no side effects).

    If the elemental function's algorithm can be made faster using array operations

    inside, or if for some reason the arguments must be arrays of incompatible

    shapes,

    then one should use the other two approaches.

    One can make ``nroot`` operate

    on a vector and write a simple wrappers for other array shapes::

        function nroot(n, x) result(y)

        integer, intent(in) :: n

        real(dp), intent(in) :: x(:)

        real(dp) :: y(size(x))

        y = x**(1._dp / n)

        end function

        function nroot_0d(n, x) result(y)

        integer, intent(in) :: n

        real(dp), intent(in) :: x

        real(dp) :: y

        real(dp) :: tmp(1)

        tmp = nroot(n, [x])

        y = tmp(1)

        end function

        function nroot_2d(n, x) result(y)

        integer, intent(in) :: n

        real(dp), intent(in) :: x(:, :)

        real(dp) :: y(size(x, 1), size(x, 2))

        y = reshape(nroot(n, reshape(x, [size(x)])), [size(x, 1), size(x, 2)])

        end function

    And use as follows::

        print *, nroot_0d(2, 9._dp)

        print *, nroot(2, [1._dp, 4._dp, 9._dp, 10._dp])

        print *, nroot_2d(2, reshape([1._dp, 4._dp, 9._dp, 10._dp], [2, 2]))

    This will print::

        3.0000000000000000

        1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

        1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

    Or one can use `explicit-shape` arrays as follows::

        function nroot(n, k, x) result(y)

        integer, intent(in) :: n, k

        real(dp), intent(in) :: x(k)

        real(dp) :: y(k)

        y = x**(1._dp / n)

        end function

    Use as follows::

        print *, nroot(2, 1, [9._dp])

        print *, nroot(2, 4, [1._dp, 4._dp, 9._dp, 10._dp])

        print *, nroot(2, 4, reshape([1._dp, 4._dp, 9._dp, 10._dp], [2, 2]))

    The output is the same as before::

          3.0000000000000000

          1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

          1.0000000000000000        2.0000000000000000        3.0000000000000000        3.1622776601683795

    Allocatable Arrays

    ------------------

    When using allocatable arrays (as opposed to pointers), Fortran manages the

    memory automatically and it is not possible to create memory leaks.

    For example you can allocate it inside a subroutine::

        subroutine foo(lam)

        real(dp), allocatable, intent(out) :: lam(:)

        allocate(lam(5))

        end subroutine

    And use somewhere else::

        real(dp), allocatable :: lam(:)

        call foo(lam)

    When the ``lam`` symbol goes out of scope, Fortran will deallocate it. If

    ``allocate`` is called twice on the same array, Fortran will abort with a

    runtime error. One can check if ``lam`` is already allocated and deallocate it

    if needed (before another allocation)::

        if (allocated(lam)) deallocate(lam)

        allocate(lam(10))

    File Input/Output

    -----------------

    To read from a file::

        integer :: u

        open(newunit=u, file="log.txt", status="old")

        read(u, *) a, b

        close(u)

    Write to a file::

        integer :: u

        open(newunit=u, file="log.txt", status="replace")

        write(u, *) a, b

        close(u)

    To append to an existing file::

        integer :: u

        open(newunit=u, file="log.txt", position="append", status="old")

        write(u, *) N, V(N)

        close(u)

    The ``newunit`` keyword argument to ``open`` is a Fortran 2008 standard, in older compilers, just replace

    ``open(newunit=u, ...)`` by::

        open(newunit(u), ...)

    where the ``newunit`` function is defined by::

        integer function newunit(unit) result(n)

        ! returns lowest i/o unit number not in use

        integer, intent(out), optional :: unit

        logical inuse

        integer, parameter :: nmin=10  ! avoid lower numbers which are sometimes reserved

        integer, parameter :: nmax=999  ! may be system-dependent

        do n = nmin, nmax

            inquire(unit=n, opened=inuse)

            if (.not. inuse) then

                if (present(unit)) unit=n

                return

            end if

        end do

        call stop_error("newunit ERROR: available unit not found.")

        end function

    .. _c_interface:

    Interfacing with C

    ------------------

    Write a C wrapper using the ``iso_c_binding`` module::

        module fmesh_wrapper

        use iso_c_binding, only: c_double, c_int

        use fmesh, only: mesh_exp

        implicit none

        contains

        subroutine c_mesh_exp(r_min, r_max, a, N, mesh) bind(c)

        real(c_double), intent(in) :: r_min

        real(c_double), intent(in) :: r_max

        real(c_double), intent(in) :: a

        integer(c_int), intent(in) :: N

        real(c_double), intent(out) :: mesh(N)

        call mesh_exp(r_min, r_max, a, N, mesh)

        end subroutine

        ! wrap more functions here

        ! ...

        end module

    You need to declare the length of all arrays (``mesh(N)``) and pass it as a

    parameter. The Fortran compiler will check that the C and Fortran types match.

    If it compiles, you can then trust it, and call it from C using the following

    declaration:

    .. code-block:: c

        void c_mesh_exp(double *r_min, double *r_max, double *a, int *N,

                double *mesh);

    use it as:

    .. code-block:: c

        int N=5;

        double r_min, r_max, a, mesh[N];

        c_mesh_exp(&r_min, &r_max, &a, &N, mesh);

    No matter if you are passing arrays in or out, always allocate them in C first, and you are (in C) responsible for the memory management. Use Fortran to fill (or use) your arrays (that you own in C).

    If calling the Fortran ``exp_mesh`` subroutine from the ``c_exp_mesh`` subroutine is a problem (CPU efficiency), you can simply implement whatever the routine does directly in the ``c_exp_mesh`` subroutine. In other words, use the ``iso_c_binding`` module as a direct way to call Fortran code from C, and you can make it as fast as needed.

    .. _python_interface:

    Interfacing with Python

    -----------------------

    Using Cython

    ~~~~~~~~~~~~

    To wrap Fortran code in Python, export it to C first (see above) and then write

    this Cython code:

    .. code-block:: cython

        from numpy cimport ndarray

        from numpy import empty

        cdef extern:

            void c_mesh_exp(double *r_min, double *r_max, double *a, int *N,

                    double *mesh)

        def mesh_exp(double r_min, double r_max, double a, int N):

            cdef ndarray[double, mode="c"] mesh = empty(N, dtype=double)

            c_mesh_exp(&r_min, &r_max, &a, &N, &mesh[0])

            return mesh

    The memory is allocated and owned (reference counted) by Python, and a pointer is given to the Fortran code. Use this approach for both "in" and "out" arrays.

    Notice that we didn't write any C code --- we only told fortran to use the C

    calling convention when producing the ".o" files, and then we pretended in

    Cython, that the function is implemented in C, but in fact, it is linked in

    from Fortran directly. So this is the most direct way of calling Fortran from

    Python. There is no intermediate step, and no unnecessary processing/wrapping

    involved.

    Using ctypes

    ~~~~~~~~~~~~

    Alternatively, you can assign C-callable names to your Fortran

    routines like this::

        subroutine mesh_exp(r_min, r_max, a, N, mesh) bind(c, name='mesh_exp')

          real(c_double), intent(in), value :: r_min

          real(c_double), intent(in), value :: r_max

          real(c_double), intent(in), value :: a

          integer(c_int), intent(in), value :: N

          real(c_double), intent(out) :: mesh(N)

          ! ...

        end subroutine mesh_exp

    and use the builtin `ctypes `_

    Python package to dynamically load

    shared object files containing your C-callable Fortran routines and

    call them directly:

    .. code-block:: python

        from ctypes import CDLL, POINTER, c_int, c_double

        from numpy import empty

        fortran = CDLL('./libmyfortranroutines.so')

        mesh = empty(N, dtype="double")

        fortran.mesh_exp(c_double(r_min), c_double(r_max), c_double(a), c_int(N),

                        mesh.ctypes.data_as(POINTER(c_double)))

    Callbacks

    ---------

    There are two ways to implement callbacks to be used like this::

        subroutine foo(a, k)

        use integrals, only: simpson

        real(dp) :: a, k

        print *, simpson(f, 0._dp, pi)

        print *, simpson(f, 0._dp, 2*pi)

        contains

        real(dp) function f(x) result(y)

        real(dp), intent(in) :: x

        y = a*sin(k*x)

        end function f

        end subroutine foo

    The traditional approach is to simply declare the ``f`` dummy variable as a

    subroutine/function using::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function f(x)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            end function

        end interface

        s = (b-a) / 6 * (f(a) + 4*f((a+b)/2) + f(b))

        end function

        end module

    The other approach since f2003 is to first define a new type for our callback,

    and then use ``procedure(func)`` as the type of the dummy argument::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            end function

        end interface

        procedure(func) :: f

        s = (b-a) / 6 * (f(a) + 4*f((a+b)/2) + f(b))

        end function

        end module

    The new type can also be defined outside of the function (and reused), like::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        interface

            real(dp) function func(x)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            end function

        end interface

        contains

        real(dp) function simpson(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        procedure(func) :: f

        s = (b-a) / 6 * (f(a) + 4*f((a+b)/2) + f(b))

        end function

        real(dp) function simpson2(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        procedure(func) :: f

        real(dp) :: mid

        mid = (a + b)/2

        s = simpson(f, a, mid) + simpson(f, mid, b)

        end function

        end module

    .. _callbacks:

    Type Casting in Callbacks

    -------------------------

    There are essentially five different ways to do that, each

    with its own advantages and disadvantages.

    The methods I, II and V can be used both in C and Fortran.

    The methods III and IV are only available in Fortran.

    The method VI is obsolete and should not be used.

    I: Work Arrays

    ~~~~~~~~~~~~~~~

    Pass a "work array" or two which are packed with everything needed by the

    caller and unpacked by the called routine. This is the old way -- e.g., how

    LAPACK does it.

    Integrator::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b, data) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x, data)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            real(dp), intent(inout) :: data(:)

            end function

        end interface

        procedure(func) :: f

        real(dp), intent(inout) :: data(:)

        s = (b-a) / 6 * (f(a, data) + 4*f((a+b)/2, data) + f(b, data))

        end function

        end module

    Usage::

        module test

        use types, only: dp

        use integrals, only: simpson

        implicit none

        private

        public foo

        contains

        real(dp) function f(x, data) result(y)

        real(dp), intent(in) :: x

        real(dp), intent(inout) :: data(:)

        real(dp) :: a, k

        a = data(1)

        k = data(2)

        y = a*sin(k*x)

        end function

        subroutine foo(a, k)

        real(dp) :: a, k

        real(dp) :: data(2)

        data(1) = a

        data(2) = k

        print *, simpson(f, 0._dp, pi, data)

        print *, simpson(f, 0._dp, 2*pi, data)

        end subroutine

        end module

    II: General Structure

    ~~~~~~~~~~~~~~~~~~~~~

    Define general structure or two which encompass the variations you actually

    need (or are even remotely likely to need going forward). This single structure

    type or two can then change if needed as future needs/ideas permit but won't

    likely need to change from passing, say, real numbers to, say, and

    instantiation of a text editor.

    Integrator::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson, context

        type context

            ! This would be adjusted according to the problem to be solved.

            ! For example:

            real(dp) :: a, b, c, d

            integer :: i, j, k, l

            real(dp), pointer :: x(:), y(:)

            integer, pointer :: z(:)

        end type

        contains

        real(dp) function simpson(f, a, b, data) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x, data)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            type(context), intent(inout) :: data

            end function

        end interface

        procedure(func) :: f

        type(context), intent(inout) :: data

        s = (b-a) / 6 * (f(a, data) + 4*f((a+b)/2, data) + f(b, data))

        end function

        end module

    Usage::

        module test

        use types, only: dp

        use integrals, only: simpson, context

        implicit none

        private

        public foo

        contains

        real(dp) function f(x, data) result(y)

        real(dp), intent(in) :: x

        type(context), intent(inout) :: data

        real(dp) :: a, k

        a = data%a

        k = data%b

        y = a*sin(k*x)

        end function

        subroutine foo(a, k)

        real(dp) :: a, k

        type(context) :: data

        data%a = a

        data%b = k

        print *, simpson(f, 0._dp, pi, data)

        print *, simpson(f, 0._dp, 2*pi, data)

        end subroutine

        end module

    There is only so much flexibility really

    needed. For example, you could define two structure types for this purpose, one

    for Schroedinger and one for Dirac. Each would then be sufficiently general and

    contain all the needed pieces with all the right labels.

    Point is: it needn't

    be "one abstract type to encompass all" or bust. There are natural and viable

    options between "all" and "none".

    III: Private Module Variables

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    Hide the variable arguments completely by passing in module variables.

    Integrator::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            end function

        end interface

        procedure(func) :: f

        s = (b-a) / 6 * (f(a) + 4*f((a+b)/2) + f(b))

        end function

        end module

    Usage::

        module test

        use types, only: dp

        use integrals, only: simpson

        implicit none

        private

        public foo

        real(dp) :: global_a, global_k

        contains

        real(dp) function f(x) result(y)

        real(dp), intent(in) :: x

        y = global_a*sin(global_k*x)

        end function

        subroutine foo(a, k)

        real(dp) :: a, k

        global_a = a

        global_k = k

        print *, simpson(f, 0._dp, pi)

        print *, simpson(f, 0._dp, 2*pi)

        end subroutine

        end module

    However it is best to avoid such global variables -- even though really just

    semi-global -- if possible. But sometimes it may be the simplest cleanest way.

    However, with a bit of thought, usually there is a better, safer, more explicit

    way along the lines of II or IV.

    IV: Nested functions

    ~~~~~~~~~~~~~~~~~~~~

    Integrator::

        module integrals

        use types, only: dp

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            end function

        end interface

        procedure(func) :: f

        s = (b-a) / 6 * (f(a) + 4*f((a+b)/2) + f(b))

        end function

        end module

    Usage::

        subroutine foo(a, k)

        use integrals, only: simpson

        real(dp) :: a, k

        print *, simpson(f, 0._dp, pi)

        print *, simpson(f, 0._dp, 2*pi)

        contains

        real(dp) function f(x) result(y)

        real(dp), intent(in) :: x

        y = a*sin(k*x)

        end function f

        end subroutine foo

    V: Using type(c_ptr) Pointer

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    In C, one would use the ``void *`` pointer. In Fortran, one

    can use ``type(c_ptr)`` for exactly the same purpose.

    Integrator::

        module integrals

        use types, only: dp

        use iso_c_binding, only: c_ptr

        implicit none

        private

        public simpson

        contains

        real(dp) function simpson(f, a, b, data) result(s)

        real(dp), intent(in) :: a, b

        interface

            real(dp) function func(x, data)

            use types, only: dp

            implicit none

            real(dp), intent(in) :: x

            type(c_ptr), intent(in) :: data

            end function

        end interface

        procedure(func) :: f

        type(c_ptr), intent(in) :: data

        s = (b-a) / 6 * (f(a, data) + 4*f((a+b)/2, data) + f(b, data))

        end function

        end module

    Usage::

        module test

        use types, only: dp

        use integrals, only: simpson

        use iso_c_binding, only: c_ptr, c_loc, c_f_pointer

        implicit none

        private

        public foo

        type f_data

            ! Only contains data that we need for our particular callback.

            real(dp) :: a, k

        end type

        contains

        real(dp) function f(x, data) result(y)

        real(dp), intent(in) :: x

        type(c_ptr), intent(in) :: data

        type(f_data), pointer :: d

        call c_f_pointer(data, d)

        y = d%a * sin(d%k * x)

        end function

        subroutine foo(a, k)

        real(dp) :: a, k

        type(f_data), target :: data

        data%a = a

        data%k = k

        print *, simpson(f, 0._dp, pi, c_loc(data))

        print *, simpson(f, 0._dp, 2*pi, c_loc(data))

        end subroutine

        end module

    As always, with the advantages of such re-casting, as Fortran lets you

    do if you really want to, come also the disadvantages that fewer compile- and

    run-time checks are possible to catch errors; and with that, inevitably more

    leaky, bug-prone code. So one always has to balance the costs and benefits.

    Usually, in the context of scientific programming, where the main thrust

    is to represent and solve precise mathematical formulations (as opposed to

    create a GUI with some untold number of buttons, drop-downs, and other

    interface elements), simplest, least bug-prone, and fastest is to use one of

    the previous approaches.

    VI: transfer() Intrinsic Function

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    Before Fortran 2003, the only way to do type casting

    was using the ``transfer`` intrinsic function. It is functionally equivalent to

    the method V, but more verbose and more error prone.

    It is now obsolete and one should use the method V instead.

    Examples:

    http://jblevins.org/log/transfer

    http://jblevins.org/research/generic-list.pdf

    http://www.macresearch.org/advanced_fortran_90_callbacks_with_the_transfer_function

    VII: Object Oriented Approach

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    The module::

        module integrals

        use types, only: dp

        implicit none

        private

        public :: integrand, simpson

        ! User extends this type

        type, abstract :: integrand

        contains

            procedure(func), deferred :: eval

        end type

        abstract interface

            function func(this, x) result(fx)

            import :: integrand, dp

            class(integrand) :: this

            real(dp), intent(in) :: x

            real(dp) :: fx

            end function

        end interface

        contains

        real(dp) function simpson(f, a, b) result(s)

        class(integrand) :: f

        real(dp), intent(in) :: a, b

        s = ((b-a)/6) * (f%eval(a) + 4*f%eval((a+b)/2) + f%eval(b))

        end function

        end module

    The abstract type prescribes exactly what the integration routine

    needs, namely a method to evaluate the function, but imposes nothing

    else on the user.  The user extends this type, providing a concrete

    implementation of the eval type bound procedure and adding necessary

    context data as components of the extended type.

    Usage::

        module example_usage

        use types, only: dp

        use integrals, only: integrand, simpson

        implicit none

        private

        public :: foo

        type, extends(integrand) :: my_integrand

            real(dp) :: a, k

        contains

            procedure :: eval => f

        end type

        contains

        function f(this, x) result(fx)

        class(my_integrand) :: this

        real(dp), intent(in) :: x

        real(dp) :: fx

        fx = this%a*sin(this%k*x)

        end function

        subroutine foo(a, k)

        real(dp) :: a, k

        type(my_integrand) :: my_f

        my_f%a = a

        my_f%k = k

        print *, simpson(my_f, 0.0_dp, 1.0_dp)

        print *, simpson(my_f, 0.0_dp, 2.0_dp)

        end subroutine

        end module

    Complete Example of void * vs type(c_ptr) and transfer()

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    Here are three equivalent codes: one in C using ``void *`` and two codes in

    Fortran using ``type(c_ptr)`` and ``transfer()``:

    ========  ===============  ===============================

    Language  Method            Link

    ========  ===============  ===============================

    C        ``void *``        https://gist.github.com/1665641

    Fortran  ``type(c_ptr)``  https://gist.github.com/1665626

    Fortran  ``transfer()``    https://gist.github.com/1665630

    ========  ===============  ===============================

    The C code uses the standard C approach for writing extensible libraries that

    accept callbacks and contexts. The two Fortran codes show how to do the same.

    The ``type(c_ptr)`` method is equivalent to the C version and that is the

    approach that should be used.

    The ``transfer()`` method is here for completeness only (before Fortran 2003,

    it was the only way) and it is a little cumbersome, because the user needs to

    create auxiliary conversion functions for each of his types.

    As such, the ``type(c_ptr)`` method should be used instead.

    .. _parallel:

    Parallel programming

    --------------------

    OpenMP

    ~~~~~~

    `OpenMP `_ should be compatible with

    non-openMP compilers. This can be enforced by prepending all

    OpenMP-specific calls by ``!\$``. Regular compilers will consider these

    lines as comments and ignore them. For OpenMP compilers, these lines

    will be considered as regular Fortran code. The following code ::

        program test_openmpi

          !\$ use omp_lib

          implicit none

          integer :: nthreads

          nthreads = -1

          !\$ nthreads = omp_get_num_threads()

          ! will print the number of running threads when compiled with OpenMP, else will print -1

          print*, "nthreads=", nthreads

        end program

    will print the number of threads used when compiled with OpenMP. It will print by default -1 if compiled without OpenMP.

    MPI

    ~~~

    There are three ways of including MPI in a fortran program:

    =============== ==================== ====================================================================

    Fortran version  Method              Comments

    =============== ==================== ====================================================================

    Fortran 08      ``use mpi_f08``      Consistent with F08 standard, good type-checking; highly recommended

    Fortran 90      ``use mpi``          Not consistent with standard, so-so type-checking; not recommended

    Fortran 77      ``include "mpif.h"`` Not consistent with standard, no type-checking; strongly discouraged

    =============== ==================== ====================================================================

    On infrastructures where ``use mpi_f08`` is not available, one should

    fallback to ``use mpi``. The use of ``include "mpif.h"`` is strongly

    discouraged, as it does not check at all the types of the argument or

    that the function calls provide the good arguments. For example, you

    don’t get any compiler warnings if you call a subroutine and forget a

    parameter, add an extra parameter, or pass a parameter of the wrong

    type. It may also lead to silent data corruption.

    相关文章

      网友评论

          本文标题:2018-04-16

          本文链接:https://www.haomeiwen.com/subject/tanykftx.html