Home>
I am writing a program to find the product of matrices in FORTRAN77.
It compiles, but I can't run it.
The error that came out is as follows.
The NOSTRICT option (default at OPT (3)) has the potential to alter the semantics of a program. Please refer to documentation on the STRICT/NOSTRICT option for more information.
what's the meaning of this?
Why can't I compile it but not?

Answer # 1
Related articles
 about errors in python input
 python  about variable errors
 ruby  about integration test errors
 about regular expression errors in javascript
 python  about errors in skorch
 about python errors
 centos  about rpm build errors that appear in rpmbuild ta slurm20025tarbz2
 about errors in python matplotlib
 about errors in dockercomposeyml file
 c++  about errors encountered while solving atcoder problems
 ruby  about errors related to migration files in rails development
 about android studio errors
 about errors in games made with python
 about errors when building ruby on rails environment
 java  about runtime errors
 about scraping errors in google spread sheet
 linux  about errors when running npm run dev
 html  about errors when inserting images
 about code errors in unity
Related questions
 A book that describes LDD (Lightly Doped Drain)
 reading csv file with mixed character string and numerical value with fortran
 i want to read and write a continuous string with fortran by specifying the number of digits
 read csv data with fortran
 fortran  disadvantages of recursive
 fortran90 fails to compile with an argument error
 compile  the result is always output as zero
 numeric value is wrong when passed to fortran function
 fortran90/95 debuted
 how to read when fortran is leftjustified and the number of digits is different
If the optimization level is 3, it is a warning that the calculation is not always performed according to the formula in the program, and mathematically equivalent transformation may be performed, so first lower the optimization level and opt (0) or debug. It is recommended to try it in a mode.
If that doesn't work, it's a problem other than this warning.
Even if the formulas are mathematically equivalent on real numbers (complex numbers), the formulas may not be equivalent in the case of floatingpoint arithmetic, so the calculation results may not always match. The compiler warns me just in case.