Fortran Maximum Number of Continuation Lines

Continuation lines in f90 (trouble?)

 Continuation lines in f90 (trouble?)

Author	Message
	Continuation lines in f90 (trouble?) Let say you want to build a module containing large constant valued arrays: (Could be multidimensional integration quadratures or Chebyshev coefficients) double precision, dimension(400), parameter :: x =                  & (/ 1.834734464564556d0, -4.235642445677834d0, 3.952866543169867d0, & 6.332453511246747d0, ................ /) This statement can NOT fit into the f90 requirements: a) Maximum 132 characters on each line b) Maximum 39 continuation lines 1) Why does f90 include the latter restriction ??? 2) How should I circumvent this restriction in order to define the constant valued array x(400) ??? Best Regards, Jens B. Helmers
Tue, 06 May 1997 22:44:57 GMT
Craig Burl #2 / 7	Continuation lines in f90 (trouble?)    Let say you want to build a module containing large constant valued arrays: (Could be multidimensional integration quadratures or Chebyshev coefficients)    double precision, dimension(400), parameter :: x =                  & (/ 1.834734464564556d0, -4.235642445677834d0, 3.952866543169867d0, & 6.332453511246747d0, ................ /)    This statement can NOT fit into the f90 requirements: a) Maximum 132 characters on each line b) Maximum 39 continuation lines    2) How should I circumvent this restriction in order to define the constant valued array x(400) ??? Hmm, maybe you could build it up using smaller named constant arrays? Does seem kind of a drag to have to do that, though.  Too bad more "sophisticated" things like DATA cannot be used to specify values for named constants (parameters). --
Fri, 09 May 1997 13:39:52 GMT
Len Mo #3 / 7	Continuation lines in f90 (trouble?) Quote: > This statement can NOT fit into the f90 requirements: >     a) Maximum 132 characters on each line >     b) Maximum 39 continuation lines > 1) Why does f90 include the latter restriction ??? > 2) How should I circumvent this restriction in order to define the constant >    valued array x(400) ??? These maximum restrictions on programs correspond to minimum requirements on processors, and thus enhance the portability of programs that adhere to them. Processors are always permitted to extend the standard, and many probably do so with these restrictions in particular. To circumvent this limit, you could, as suggested by Craig Burley, build them up from smaller named constant arrays.  For example,       DOUBLE PRECISION, PARAMETER, DIMENSION(25) :: X_01 = & & (/ ... DOUBLE PRECISION, PARAMETER, DIMENSION(25) :: X_02 = & & (/ ... . . . DOUBLE PRECISION, PARAMETER, DIMENSION(25) :: X_16 = & & (/ ...       DOUBLE PRECISION, PARAMETER, DIMENSION(400) :: X =   & & (/ X_01, X_02,... X_16 /) -- necessarily Stanford Linear Accelerator Center, MS 97 | reflect those of SLAC, P.O. Box 4349; Stanford, CA  94309        | Stanford or the DOE
Sat, 10 May 1997 02:14:57 GMT
IanMaclu #4 / 7	Continuation lines in f90 (trouble?) Quote: >   Let say you want to build a module containing large constant valued arrays: >   (Could be multidimensional integration quadratures or Chebyshev coefficients) >   double precision, dimension(400), parameter :: x =                  & >    (/ 1.834734464564556d0, -4.235642445677834d0, 3.952866543169867d0, & >       6.332453511246747d0, ................ /) >   This statement can NOT fit into the f90 requirements: >       a) Maximum 132 characters on each line >       b) Maximum 39 continuation lines >   2) How should I circumvent this restriction in order to define the constant >      valued array x(400) ??? >Hmm, maybe you could build it up using smaller named constant arrays? >Does seem kind of a drag to have to do that, though.  Too bad >more "sophisticated" things like DATA cannot be used to specify >values for named constants (parameters). >-- Of course your compiler may allow you to override the line continuation limit as well. Alternatively use an implied do: C234567 INTEGER   TI(10) C DATA      (TI(I),I=1,5) &              /       1, &                      2, &                      3, &                      4, &                      5       / C DATA      (TI(I),I=6,10) &              /       6, &                      7, &                      8, &                      9, &                      10      / The price you may be reduced portability but on the other hand, at least in fortran 77 I've not had problems on a number ( admittedly not large ) of platforms using a number of compilers. IBM -- ################ No Times Like The Maritimes, Eh! ###################### ########## Opinions expressed here are mine, mine, mine. ###############
Sun, 11 May 1997 00:02:55 GMT
Len Mo #5 / 7	Continuation lines in f90 (trouble?) Quote: > Of course your compiler may allow you to override the line continuation > limit as well. Alternatively use an implied do: The original poster was asking, I believe, about large _constant_ arrays, i.e., with the PARAMETER attribute in Fortran 90.  Such arrays cannot be initialized via the DATA statement and thus cannot make use of an implied-do on their indexes (though if the initialization values are regularly spaced, an implied-do in the array constructor may help). -- necessarily Stanford Linear Accelerator Center, MS 97 | reflect those of SLAC, P.O. Box 4349; Stanford, CA  94309        | Stanford or the DOE
Sun, 11 May 1997 05:41:18 GMT
Michel Olagnon, Ifremer DITI GO, 98.22.41. #6 / 7	Continuation lines in f90 (trouble?) Quote: >   Let say you want to build a module containing large constant valued arrays: >   (Could be multidimensional integration quadratures or Chebyshev coefficients) >   double precision, dimension(400), parameter :: x =                  & >    (/ 1.834734464564556d0, -4.235642445677834d0, 3.952866543169867d0, & >       6.332453511246747d0, ................ /) >   This statement can NOT fit into the f90 requirements: >       a) Maximum 132 characters on each line >       b) Maximum 39 continuation lines >   2) How should I circumvent this restriction in order to define the constant >      valued array x(400) ??? Since you should not care much about readability, just use double precision, parameter :: x00 =  1.834734464564556d0 double precision, parameter :: x01 = -4.235642445677834d0 double precision, parameter :: x02 =  3.952866543169867d0 double precision, parameter :: x03 =  6.332453511246747d0 ! ... double precision, parameter :: xzz =  ! whatever ! double precision, dimension(400), parameter :: x =  & (/x00,x01,x02,..........,xzy,xzz/) With 3 character names like this, you can handle up to about dimension 1250. If it is not sufficient, define for instance x1to1250 (1250), x1251to2500 (1250) and double precision, dimension(2500), parameter :: x =  & (/ (x1to1250 (i), i=1,1250), (x1250to2500 (i), i=1,1250) /) Michel --- | IFREMER: Institut Francais de Recherches pour l'Exploitation de la Mer|
Sun, 11 May 1997 17:05:50 GMT
Bill Lo #7 / 7	Continuation lines in f90 (trouble?) Quote: >    Let say you want to build a module containing large constant valued arrays: >    (Could be multidimensional integration quadratures or Chebyshev coefficients) >    double precision, dimension(400), parameter :: x =                  & >     (/ 1.834734464564556d0, -4.235642445677834d0, 3.952866543169867d0, & >        6.332453511246747d0, ................ /) >    This statement can NOT fit into the f90 requirements: >        a) Maximum 132 characters on each line >        b) Maximum 39 continuation lines >    2) How should I circumvent this restriction in order to define the constant >       valued array x(400) ??? The same problem comes up if you try to initialize a non-parameter array as part of the initial specification statement.  In the latter case a sequence of DATA statements provides a way out. I would question the purpose of making such a large object a parameter in the first place. It is true that parameter status would prevent its reassignment in the program. But beyond that there is probably little practical difference between the parameter object above and a simple array which is initialized with data statements.  It is quite unlikely that the compiler would actually completely unroll loops involving x and directly insert the constants one-by-one. Cheers, Bill Long Department of Physics University of Wisconsin-Madison
Thu, 15 May 1997 11:27:36 GMT

Powered by phpBB® Forum Software

cheekeforl1975.blogspot.com

Source: http://computer-programming-forum.com/49-fortran/a79ab45ec5afd1a3.htm

Share this post