is pair_meamc's neigh_strip() technique a best practice?

Hello,

FWIW, pair/meamc just inherited that from pair/meam. Given that we don't have to
rewrite indices with neigh_c2f/f2c anymore, I believe this step could be removed
completely? I'm not 100% sure what would happen however, I have no experience
with systems with bonds (and their internal representation).

That doesn't really answer your question, though...

Regards,

Sebastian Hütter

Hello,

FWIW, pair/meamc just inherited that from pair/meam. Given that we don’t have to
rewrite indices with neigh_c2f/f2c anymore, I believe this step could be removed
completely?

​no. most pair styles that do not handle exclusions do this step ​during the compute step. please see pair_sw.cpp or pair_tersoff.cpp and others. thus there is no “best practice”.

I’m not 100% sure what would happen however, I have no experience
with systems with bonds (and their internal representation).

​there are only a small number of rather exotic use cases, where these bits would be set. but following the “better safe than segfault” strategy, it is good practice to strip off exclusion flags always.

axel.​

OK,

Thanks all!

Ryan