lagrangian multipliers & multiple time step (RESPA)

89 views

Skip to first unread message

Jinggang Lan

unread,

Oct 26, 2021, 4:40:46 PM10/26/21

to cp2k

When the blue-moon ensemble combines with RESPA, the Lagrangian multipliers are incorrect. The SHAKE Lagrangian Multipliers from RESPA are much smaller than the normal constrained MD.

Vladimir Rybkin

unread,

Oct 26, 2021, 5:45:11 PM10/26/21

to cp2k

Hi Jingagang,

I believe the issue is that although the constraint (shake/rattle) is in principle implemented for RESPA (src/motion/integrator.F) it's only applied to the inner loop of the algorithm which likely uses the high-low force_eval force difference as far as I can see from the code.

Yours,

Vladimir

вторник, 26 октября 2021 г. в 22:40:46 UTC+2, Jinggang Lan:

Reply all

Reply to author

Forward

0 new messages