I think you cannot fix it. As stated, PMEMD needs more than a single atom (I am assuming H residue is just that). Also, I am curious what you expect to learn from simulation of a single atom. If you ran MD on this, the atom would just move in straight line based on initial velocity (assuming a T scaling algorithm) ... with nothing to collide into to change course, it would just drift forever. Depending on code / force accuracy, with T coupling it could speed up over time (slowly, assuming energy drain although since single atom, no worry about SHAKE accuracy), but not change course since nothing to collide with. With Langevin you may get kicks out of the straight line, but still unclear what you learn.
--tec3
________________________________________
From: Huimin Tian <201921150072.mail.bnu.edu.cn>
Sent: Thursday, March 24, 2022 8:44 PM
To: amber.ambermd.org
Subject: [AMBER] PMEMD does not support single H residues!
Dear all,
I encountered a problem a problem when running Molecular Dynamics with GPU of AMBER2020(PMEMD).
I have a single H residue in my system and When I use GPU for dynamics simulation, it prompts an error that PMEMD does not support single H residues!
Below is my input file：“export CUDA_VISIBLE_DEVICES=0”
“pmemd.cuda -O -i 0-5ns.in -o md0-5ns.out -p NI_R_4U9H_solv.prmtop -c equ.rst -r md0-5ns.rst -x md0-5ns.mdcrd&”
How can I fix this error?
Thanks and Regards,
Tian
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Mar 24 2022 - 20:30:02 PDT
