I am writing this F90 program to compute a function in fortran which takes the input from a .mat file and save the results in another .mat file.
I followed this answer to get the code compiled and correctly linked. This is my makefile command:
gfortran -g -fcheck=all binhkorn_mat.F90 -I/usr/local/MATLAB/R2015b/extern/include/ -L/usr/local/MATLAB/R2015b/bin/glnxa64 -cpp -o binhkorn_mat -lmat -lmx -Wl,-rpath /usr/local/MATLAB/R2015b/bin/glnxa64/
The output file is apparently correctly compiled, but then once I run the program the following SF appears (I'm working on LINUX Ubuntu 14.04 LTS):
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x7FD650063F27
#1 0x7FD6500644F4
#2 0x7FD64FCBCD3F
#3 0x7FD64FDD7AF6
#4 0x400A3F in binhkorn_mat at binhkorn_mat.F90:17 (discriminator 2)
./binhkorn_mat: Segmentation fault
I can't figure out if there's an error with the compiler or if I did something wrong with the pointers/functions definitions. Here's the code (binhkorn_mat.F90):
#include "fintrf.h"
PROGRAM binhkorn_mat
IMPLICIT NONE
mwPointer matOpen, matGetVariable, matPutVariable
mwPointer mpin, mpX, mpout, mpcf
INTEGER :: i,j
REAL*8, DIMENSION(2) :: x
REAL*8, DIMENSION(4) :: cf
!input/output through .mat f
mpin = matOpen('X.mat', 'u')
mpX = matGetVariable(mpin, 'X')
CALL mxCopyPtrToReal8(mpX, x, 2)
CALL matClose(mpin)
!fitness functions
cf(1) = ((x(1)-2)**2 + (x(2)-1)**2 + 2)
cf(2) = (9*x(1) + (x(2)-1)**2)
!constraints
cf(3) = x(1)*x(1) + x(2)*x(2) - 225
cf(4) = x(1) - 3*x(2) + 10
!output file created
CALL mxCopyReal8ToPtr(cf, mpcf, 4)
mpout = matOpen('cf.mat', 'w')
mpcf = matPutVariable(mpout, 'cf', mpcf)
CALL matClose(mpout)
END PROGRAM
The X.mat file is correctly created by an external Matlab script and contains a variable named X
which is a 2-element row vector.
I basically misunderstood how to use of many functions. The pointers i supplied as input to some of them were not the correct ones. I post here the working solution:
#include "fintrf.h"
PROGRAM binhkorn_mat
IMPLICIT NONE
mwPointer matOpen, matGetVariable!, matPutVariable
mwPointer mxGetData, mxGetNumberOfElements, mxCreateNumericArray
mwPointer mpin, mpX, mpout, mpcf
mwSize ndim
mwSize dims(2)
INTEGER :: s
INTEGER*4 mxClassIDFromClassName
CHARACTER (LEN = 6) :: classname
REAL*8, DIMENSION(2) :: x
REAL*8, DIMENSION(4) :: cf
!input/output through .mat f
mpin = matOpen('X.mat', 'r')
mpX = matGetVariable(mpin, 'X')
CALL mxCopyPtrToReal8(mxGetData(mpX), x, mxGetNumberOfElements(mpX))
!CALL matClose(mpin)
!fitness functions
cf(1) = ((x(1)-2)**2 + (x(2)-1)**2 + 2)
cf(2) = (9*x(1) + (x(2)-1)**2)
!constraints
cf(3) = x(1)*x(1) + x(2)*x(2) - 225
cf(4) = x(1) - 3*x(2) + 10
!output .mat file created and filled
s = size(cf)
ndim = 2
classname = 'double'
dims(1) = 1
dims(2) = s
mpcf = mxCreateNumericArray(ndim, dims, mxClassIDFromClassName(classname), 0)
CALL mxCopyReal8ToPtr(cf, mxGetData(mpcf), mxGetNumberOfElements(mpcf))
mpout = matOpen('cf.mat', 'w')
CALL matPutVariable(mpout, 'cf', mpcf)
!CALL matClose(mpout)
END PROGRAM
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.