@Katetc, thanks for taking care of this. I think this branch would be a good test for the refurbished LIVVkit. We could do something like this:
(1) Build "main" prior to these three commits, with the old python2 scripts. Run a LIVVkit suite and make this our baseline case.
(2) Rebuild with the first commit ("Making CISM python3 compatible and bit for bit"). Run LIVVkit and verify that answers are BFB as advertised.
(3) Rebuild with the second commit ("Ismip-hom no longer bit for bit). Run LIVVkit and confirm that ismip-hom answers have changed as expected.
(4) Rebuild with the third commit ("Additional updates..."). Run LIVVkit and confirm that the answers are BFB compared to the previous commit. Make this our new LIVVkit baseline for further commits (like the basal_physics3 branch).
It will take some extra time for us to get familiar with LIVVkit and do these tests. But I think it would be a useful exercise, since this is how we'd like to test many of our future PRs.
@Katetc, thanks for taking care of this. I think this branch would be a good test for the refurbished LIVVkit. We could do something like this: (1) Build "main" prior to these three commits, with the old python2 scripts. Run a LIVVkit suite and make this our baseline case. (2) Rebuild with the first commit ("Making CISM python3 compatible and bit for bit"). Run LIVVkit and verify that answers are BFB as advertised. (3) Rebuild with the second commit ("Ismip-hom no longer bit for bit). Run LIVVkit and confirm that ismip-hom answers have changed as expected. (4) Rebuild with the third commit ("Additional updates..."). Run LIVVkit and confirm that the answers are BFB compared to the previous commit. Make this our new LIVVkit baseline for further commits (like the basal_physics3 branch). It will take some extra time for us to get familiar with LIVVkit and do these tests. But I think it would be a useful exercise, since this is how we'd like to test many of our future PRs.