io_nemo_f(filename, slen, MAXBODY, param, ....)
character * filename(*);/* input/output file name */
integer slen; /* #characters in ’infile’ */
integer MAXBODY; /* maximun bodies allowed */
character * param(*); /* parameters list */
Descriptionio_nemo_f is a high level C function which allows the user to
perform I/O operations on NEMO snapshot files from a FORTRAN program. The
use of this function is very simple and doesn’t require any knowledge of
how to program in NEMO.
Parameters
filename=character variable Input/Output file name of the NEMO snapshot
that you want to read/save. (this CHARACTER string must be a VARIABLE, not
a constant !) slen=integer Lenght of the file name. (We need this information
because of the interface bettween FORTRAN and C language) MAXBODY=integer
Size of array that you use to get/put the data. Match the size of the array
to the maximum number of bodies that you are able to read/save. param=character
Param is a string in which you specify what you want to do with the NEMO
file. Each choice is defined with a flag separated with ’,’. There are two
kind of flags. (1) Information flags allow to specify some actions during
the I/O. (2) Variables flags allow to specify what you want to get/put from/into
the NEMO file. The FLAGS list is described below.
(1) Information flagss
| save Specify that you want to save data to a NEMO file. r | read Specify
that you want to read data from a NEMO file. real4|single Specify that the
variables that you use to get/put data from/into NEMO files have been declared
in single precision. (real * 4). real8|double Specify that the variables you
use to get/put data from/into NEMO files have been declared in double precision.
(real * 8). f | n3 Assumes that you declared yours two-dimensionnal FORTRAN
arrays (pos, vel, acc) using the number of bodies for the first dimension
and the number of space coordinates (always 3D) for the second dimension.
(example : real * 4 pos(10000,3) ). c | 3n Assumes that you declared yours
two-dimensionnal FORTRAN arrays (pos, vel, acc) using the number of space
coordinates (always 3D) for the first dimension and the number of bodies
for the second dimension. (example : real * 4 vel(3,10000) ). info | diag
Gives some informations during the runtime execution.
(2) Variables flags
n|nbody Match to the number of bodies. t|time Match to the snapshot time. m|mass
Match to the particle masses. x|pos Match to the particle positions. v|vel
Match to the particle velocities. p|pot Match to the particle potentials.
a|acc Match to the particle accelerations. st Select snapshot’s times. You
have to put the selected time in a string limited by the character ’#’. Example
"100:150#" select the time beween 100 and 150. The variable flags specify
which data you will get/put from/into the NEMO file, hence after the selection
string param, you must insert all the selected variables in the same order
with which they have been declared in param. ExampleHere is a FORTRAN program
to illustrate the use of the function io_nemo_f(). The program reads all
the time steps into a NEMO a file, and saves the data in another NEMO file.
C-----------------------------------------------------------
program nemo_fortran
implicit none
c maximum # of bodies
integer MAXBODY
parameter (MAXBODY=100000)
real*8
c particle masses.
+ mass(1:MAXBODY),
c particle positions and velocities
+ pos(3,MAXBODY), vel(3,MAXBODY),
c snapshot time
+ ts
c file names
character infile*80,outfile*80
integer i,j,k,nbody,io_nemo_f,close_io_nemo_f
c get input snapshot filename
write(*,*)’Infile name : ’
read(*,’(a80)’)infile
c get output snapshot filename
write(*,*)’outfile name : ’
read(*,’(a80)’)outfile
i = 1
do while (i.gt.0)
c read the snapshot up to the end of file ( i <= 0 )
i=io_nemo_f(infile,80,MAXBODY,"real8,3n,read,n,m,x,v,t,info",
$ nbody,mass,pos,vel,ts)
c save the snapshot
if (i.gt.0) then
j=io_nemo_f(outfile,80,MAXBODY,"real8,3n,save,n,m,x,v,t,
$ info",nbody,mass,pos,vel,ts)
endif
end do
c close the snapshot ’outfile’
k= close_io_nemo_f(outfile,80)
end
C-----------------------------------------------------------
Important Thingsa) Notice in the example above, that in the parameter list,
’n’ matches ’nbody’, ’m’ matches ’mass’, ’x’ matches ’pos’, ’v’ matches ’vel’, ’t’ matches
’ts’. All the variables are in the same order that they have been declared
in the param list. b) You must declare all the two-dimensionnal arrays in
the same way. That means all the dimensions must be the same for all the
arrays, moreover both one-dimensional and two-dimensional array must have
the same size for the maximum of bodies. c) All the arrays must be declared
in the same floating type. d) During a "read" operation, the function io_nemo_f()
return ’0’ if it is the end of the NEMO file. That means that no new values
have been read.
FFLAGS = -fno-second-underscore -fwritable-strings
# ----------------------------------------
# MAKEFILE to use IO_NEMO_F
#
# ----------------------------------------
# path for NEMO Library and IO_NEMO_F library
LIBS = -L$(NEMOLIB) -L/usr/local/lib
FFLAGS = $(OPT) -fno-second-underscore -fwritable-strings
nemo_fortran : nemo_fortran.o
$(FC) -o $@ nemo_fortran.o $(LIBS) \
-lio_nemo_f -lm
# ----------------------------------------
nemo(1NEMO), snapshot(5NEMO).
15-Jun-95 V1.0 : created JCL 21-Jun-95 V1.1 : bugs fixes JCL 12-Dec-95 V1.2 : possibility to close file JCL 11-Mar-96 V1.3 : acceleration I/O added JCL 04-Apr-97 V1.4 : generic real format JCL 07-Apr-97 V1.41: manual created JCL 09-Apr-97 V1.42: manual corrections AS