search

njoy / NJOY2016

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

Bug in acefc.90 #188

Closed emiliomendozacembranos closed 2 years ago

emiliomendozacembranos commented 3 years ago

Dear NJOY developers,

I have found a bug in NJOY2016. It is located in line 10235: llx=lld+6 should be replaced with: llx=lld+6+(ig-1)(2+2nmu)

The reason is that the index llx has to go throug a full matrix. As it is now, it is going several times through the first row/column, without moving to the others.

Best regards,

Emilio

whaeck commented 3 years ago

I have added these changes to the fix/ciemat branch.

As I indicated in my email, I would like to diagnose the underlying problem in more detail to see if any of our processing is impacted by this. Once I fully understand what is going on, we can proceed with pulling this into master or develop - depending on the urgency to deploy these changes.

I would also like to add a test to the test suite to detect this in the future.

whaeck commented 3 years ago

@emiliomendozacembranos Any news on releasing that evaluation? I really would like to add this to the test suite.

emiliomendozacembranos commented 3 years ago

Dear whaeck, I have contacted the authors of the evaluation, but I am still wainting for their answer. I will keep you updated. Regards, Emilio

whaeck commented 3 years ago

@emiliomendozacembranos I've been looking through ENDF/B-VIII.0 evaluations and I saw Be9 uses LAW=7 for MT16. Can you confirm that the issue can be reproduced using this evaluation? It is the only evaluation in ENDF/B-VIII.0 that uses LAW=7.

emiliomendozacembranos commented 3 years ago

Dear whaek,

This format is not used very often for neutron incident data libaries. Indeed, for neutron incident, LAW=7 is only used in Be-9 in ENDF/B-VIII.0. In JEFF-3.3 and JENDL-4.0 only in H-2 and Be-9. However, the issue cannot be reproduced with neutron incident data libraries. The piece of the NJOY code which contains the bug is used only for charged particle data libraries.

Still no answer from the authors of the evaluation.

Regards, Emilio

whaeck commented 2 years ago

@emiliomendozacembranos I know it has been a year - almost to the day - but have you had any reponse from the evaluators on providing the evaluation. I really need a test case before I can put it into an official release.

emiliomendozacembranos commented 2 years ago

Dear whaeck,

Yes. I got a positive answer. I already senty you an email with all the info and files, to you LANL email account. The date of that email is 30/4/21, i.e almost one year ago. Maybe you missed it, but I can send it to you again.

Regards, Emilio

whaeck commented 2 years ago

Oops. Must have missed that. I'll check right away.

From: emiliomendozacembranos @.***> Sent: Wednesday, March 16, 2022 2:25:15 AM To: njoy/NJOY2016 Cc: Haeck, Wim; Comment Subject: [EXTERNAL] Re: [njoy/NJOY2016] Bug in acefc.90 (#188)

Dear whaeck,

Yes. I got a positive answer. I already senty you an email with all the info and files, to you LANL email account. The date of that email is 30/4/21, i.e almost one year ago. Maybe you missed it, but I can send it to you again.

Regards, Emilio

— Reply to this email directly, view it on GitHubhttps://github.com/njoy/NJOY2016/issues/188#issuecomment-1068858154, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AENOHQCM3J6EK5QSFTE4FHLVAGLGXANCNFSM4XPI3DRQ. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub. You are receiving this because you commented.Message ID: @.***>

whaeck commented 2 years ago

Follow-up notes: this issue should also present itself when using LAW=1 with interpolation types 11 to 15. The reason why this also happens when using LAW=7 is that under specific circumstances (newfor=1 in the acer input) this LAW gets transformed into a LAW=1 using an interpolation of 10 + the original interpolation type to indicate a fully tabulated E,Eprime,mu (or ACELAW=61).

I'm going to look for an ENDF/B-VIII.0 evaluation that uses LAW=1 LANG=1x.

whaeck commented 2 years ago

I have checked JENDL40, ENDF/B-VII.0, ENDF/B-VII.1, ENDF/B-VIII.0 and JEFF33 and none of these libraries seem to use the LAW=1 LANG=1x option. However, there are just a few files that use LAW=7 (e.g. ENDF/B-VIII.0 n-Be9).