search

njoy / NJOY2016

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

Some updates to PURR made by Bob MacFarlane. #177

Closed jlconlin closed 3 years ago

jlconlin commented 3 years ago

This was created by Bob MacFarlane who wrote:

In working with Red on the La-139 problem, I made several changes to the PURR module. First, Red wanted to run at sigma0=0. That is an easy two line fix. Second, I carelessly ran with an ascending sigma0 list (against the input instructions). That destroyed the normalization. I rewrote it to allow either ascending or descending lists. Third, the nunx option in the input didn't work, so I fixed that.

jlconlin commented 3 years ago

This will require changes to the tests as well as an update to version number. Let's take a look at this first and see what's going on first.

kahlerac commented 3 years ago

All, Let's be careful here. It is my understanding that Bob was working with NJOY2012.0 when he "upgraded" his purr.f90 module. Make sure you don't lose all the other improvements made over the years since then when merging his update! Skip

On Thu, Oct 22, 2020 at 6:15 PM Jeremy Lloyd Conlin < notifications@github.com> wrote:

This will require changes to the tests as well as an update to version number. Let's take a look at this first and see what's going on first.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/njoy/NJOY2016/pull/177#issuecomment-714792107, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEHJISOGEAQC6A6NY5ZTROTSMCVHDANCNFSM4S3ZE3UA .

-- Dr. A. C. (Skip) Kahler Kahler Nuclear Data Services, LLC kahler3ac@gmail.com +1 321 368 3645

whaeck commented 3 years ago

@kahlerac You're right, we should proceed with extreme caution. I notified Jeremy as soon as I saw this about the fact that this is NJOY2012 code, and not NJOY2016. I'm trying to compare the file with the original NJOY2012 source code we have here and try and figure out what exactly has changed.

To be honest, I would like more information of what the underlying issues are, get a test case for them and then make the modifications one by one.

jchsublet commented 3 years ago

@kahlerac apart from some Fortran semantics, there is no difference to speak off between purr.f90 in NJOY2016/src and the NJOY2012.149 version. The issue was that Red wanted sig0=0 at a given energy, , sig0 ascending, ...and convinced Bod to mend that so the nunx revelation, ...what is much more dangerous (excuse the term) in purr is the renormalisation before feeding Ace but then it also open the question does MCNP really do a good job out of Purr tables?

kahlerac commented 3 years ago

Yes, but Bob's (partial) email to me said "... making some PURR calculations for him using my obsolete NJOY2012.0, and I found some problems ...". Hence a blind acceptance of his modified purr.f90 will lose 2012 changes in updates 6, 14, 22, 27, 55, 70, 93, 103, 107, 120, 128, 129 plus any other njoy2016 changes that haven't been back-fitted into njoy2012. Skip

On Fri, Oct 23, 2020 at 5:24 AM jchsublet notifications@github.com wrote:

@kahlerac https://github.com/kahlerac apart from some Fortran semantics, there is no difference to speak off between purr.f90 in NJOY2016/src and the NJOY2012.149 version. The issue was that Red wanted sig0=0 at a given energy, , sig0 ascending, ...and convinced Bod to mend that so the nunx revelation, ...what is much more dangerous (excuse the term) in purr is the renormalisation before feeding Ace but then it also open the question does MCNP really do a good job out of Purr tables?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/njoy/NJOY2016/pull/177#issuecomment-715219249, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEHJISNKHRJOCR7P6A62ZT3SMFDTLANCNFSM4S3ZE3UA .

-- Dr. A. C. (Skip) Kahler Kahler Nuclear Data Services, LLC kahler3ac@gmail.com +1 321 368 3645

jchsublet commented 3 years ago

Hi Skip, Bob now has NJOY12.149 and I hope I have convinced him to work, put his modifications using upd/upn in this one!! This should make any move forward so much more safer and recordable

whaeck commented 3 years ago

I have been looking into the fixes Bob proposed, I'm currently on the simplest one of them (the nunx issue) but it doesn't appear that "simple". The proposed fix is lacking in a number of places (e.g. the wrong unresolved resonance energy is being assigned as well as the wrong heating cross sections when writing MF2 MT152 and MT153), and as a result ACER and VIEWR get in a bind downstream (VIEWR seem to be getting into an infinite loop). For more information: please see https://github.com/njoy/NJOY2016/issues/178

I'd prefer working with Bob directly to solve the underlying issues. Can someone ask him if that is OK for him (I don't want to contact him since he's retired)? I'd also like to know what the La-139 problem is he's referring to. I'd also like to point out that NJOY2012's PURR module should be fixed to properly calculate the probability tables for nuclides with an average level spacing larger than 500 eV (see the corresponding fix in NJOY2016: https://github.com/njoy/NJOY2016/pull/59).

Since this is not in any state to be pulled in, I'll be closing this pull request by the end of the day.

I'd also like to note that I strongly advise against using NJOY2012 for any production work since it does not contain the changes and updates we've made to NJOY2016 (the purr fix mentioned above is just one of these).

jchsublet commented 3 years ago

Vim, you can contact him directly, cc or tell him that i think it is for the best. I’ve stayed in touch, and he raise the issue

jchsublet commented 3 years ago

Sorry Wim

whaeck commented 3 years ago

@jchsublet It was as if I was still in France where my name is sometimes pronounced that way. You know I don't have any problems with that, no need to apologise ;-)