[gmx-users] Fwd: LJ cut-offs

Dawid das addiw7 at googlemail.com
Wed Jan 11 13:50:09 CET 2017

Also, you could try to find something in these articles:


2017-01-11 13:48 GMT+01:00 Dawid das <addiw7 at googlemail.com>:

> You might also want to read the Default section over here:
> http://charmm.sunhwanj.com/html/nbonds.html
> where following ''equations'' give you correspondence between CHARMM and
> Gromacs keywords
> rvdw-switch == ctonnb, rlist == cutnb, rvdw == ctofnb
> 2017-01-11 13:36 GMT+01:00 Dawid das <addiw7 at googlemail.com>:
>> Hi,
>> I believe that Mohsen here did not ask about what you are writing about
>> Mark. He asks about cut-offs not the exponents
>> in the LJ formula.
>> I am no expert in that but I reviewed Andrew Leach's book on that and
>> also other resources and I did not find
>> any reason why lower and upper cut-off for LJ interactions cannot be  0.9
>> - 1.0 nm, 1.1 - 1.3 nm, etc.
>> In fact Leach gives an example of 0.9 - 1.0 nm cut-offs between which the
>> shifting function is applied.
>> Probable reason for using the difference of 0.2 nm is that this is more
>> than the length of typical chemical
>> bond, while merely 0.1 nm would not give you properly thick "skin" and
>> that choice would not be too
>> meaningful to meet reasons for which we use cut-off and
>> switching/shifting functions (which I'm sure you know).
>> You need to keep in mind however that each force field was optimized for
>> given set of cut-offs and
>> they virtually become part of that FF (just like FF formula and
>> parameters), so it is strongly recommended
>> not to change them!
>> Best regards,
>> Dawid Grabarek

More information about the gromacs.org_gmx-users mailing list