search

njoy / NJOY2016

Nuclear data processing with legacy NJOY
https://www.njoy21.io/NJOY2016
Other
98 stars 86 forks source link

Feature/updated constants #99

Closed jlconlin closed 6 years ago

jlconlin commented 6 years ago

In this Pull Request, I've updated the physical constants to something a bit more modern. Of course, this changes virtually all of the reference tapes. Exactly what to do with this is a matter of discussion.

jlconlin commented 6 years ago

Note: @kahlerac did most of the work on this and gave me the upn file before he retired. I incorporated the changes into git and had to make a few minor changes to get it to compile.

coveralls commented 6 years ago

Coverage Status

Coverage remained the same at ?% when pulling 5c7c6f5132cad260ffa0c9dd1d520eee7ef759e0 on feature/updatedConstants into d72bbb6250bf65b29a6ce835f65c3860e118f498 on master.

whaeck commented 6 years ago

So I looked at the changes in the output files. Can you rerun the tests using a tolerance of 1e-5? That'll allow us to quantify the overall impact of changing the values of the physical constants.

jlconlin commented 6 years ago

I tested with 1E-5 relative tolerance. This didn't change the test results.

When the comparison is done, the first thing that is done is to count the number of lines in the the generated file and the reference file. If the number of lines are not the same, then no other checking is done. In fact, additional checking can't be done since we do everything line-by-line.

jlconlin commented 6 years ago

Now, perhaps we should change the way that we compare references files. There is an argument to be made there.

whaeck commented 6 years ago

Well, we do test non regression, so if the number of lines is different there is definitely something going on. In those cases a by hand verification is warranted.

I assume that in this case, no files actually get additional lines?

jlconlin commented 6 years ago

No, there are plenty that get additional lines.

jlconlin commented 6 years ago

I've updated the constants even more. I double-checked the values put in previously. I used the constants in the ENDF manual and from NIST (when not in the ENDF manual).

whaeck commented 6 years ago

By the way, do not forget to update the version number and the date.