Porting to QVF6.5

Porting to QVF6.5

Im porting a program written in fortran 90/95 unix to QVF6.5 pc. After fixing some small problems i am able to compile and run the program. The problem is that the calculated results differ from the old version. My question is if there is any known conversion errors, let's say datatypes that aren't the same or other things that can cause wrong calculations but without causing compiler errors?

(the program is BIG, and I haven't written the original)

2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

There are lots of potential reasons. One is data alignment of derived types and commons. Another is if the program makes any assumption about floating point or other binary formats that might be different between the original and new architectures. One thing you can do is turn on all warnings and see if the compiler can flag anything that might assist you in locating this sort of thing. Otherwise you'll probably want to have the program output intermediate results so you can pinpoint where the results start to diverge.

James

Login to leave a comment.