mexFunction return valur error after doing clear all

Hello,

I'm running C code as mexFunction and the result returned was correct before I added 'clear all'. After adding 'clear all' at the beginning of the compile .m script, certain values of the output matrix became infinitely large, which I guess are addresses instead of the calculated values. Then after omitting 'clear all', the output matrix was correct again.

       What could have caused the error? If without 'clear all', the mexFunction gave the correct answer, can I conclude that the C code contains no error?

clear all;

load('parametersForKalmanFilter.mat');

      blaslib = fullfile('C:\Program Files\MATLAB\R2011b\extern\lib\win64\microsoft\libmwblas.lib');
      lapacklib = fullfile('C:\Program Files\MATLAB\R2011b\extern\lib\win64\microsoft\libmwlapack.lib');
      mex('-largeArrayDims', 'kalmanFilter.c', 'mathFunction64.c', 'softConstrainedAddFunctions.c', blaslib, lapacklib)

      [output1] = kalmanFilter(varA, varB, varC, varD);

Answer 1

[Kaustubha Govind]

Could you replace "clear all" with "clear mex" and see if have similar behavior? That is, having "clear mex" causes issues, but commenting it out gets the right result. If the issue is "clear mex", you could look for things like static/persistent variables in your C code and examine how they are being used.

[Jane Jean]

clear mex worked once and did not work afterwards...

[Jane Jean]

Sorry, my bad. There were some errors in the .mat file I used. >.<