Each MPI tread is supposed to take a single GPU. You can run a REMD simulation using a single GPU, but this is a terrible idea because this would kill you performance.
VinÃcius Wilian D Cruzeiro
PhD Candidate
Department of Chemistry, Physical Chemistry Division
University of Florida, United States
Voice: +1(352)846-1633<tel:+1(352)846-1633>
On Feb 16, 2018, at 7:41 PM, James Kress <jimkress_58.kressworks.org<mailto:jimkress_58.kressworks.org>> wrote:
So, Replica exchange simulations cannot be run on a system with only 1 GPU?
Jim Kress
-----Original Message-----
From: Cruzeiro,Vinicius Wilian D [mailto:vwcruzeiro.ufl.edu]
Sent: Friday, February 16, 2018 6:38 AM
To: AMBER Mailing List <amber.ambermd.org<mailto:amber.ambermd.org>>
Subject: Re: [AMBER] 3 GPUs and replica exchange MD
Hello Nikolay,
Replica exchange simulations require an even number of GPUs. This is because the number of replicas needs to be even.
Best,
VinÃcius Wilian D Cruzeiro
PhD Candidate
Department of Chemistry, Physical Chemistry Division University of Florida, United States
Voice: +1(352)846-1633<tel:+1(352)846-1633>
On Feb 16, 2018, at 4:28 AM, Nikolay N. Kuzmich <nnkuzmich.gmail.com<mailto:nnkuzmich.gmail.com><mailto:nnkuzmich.gmail.com>> wrote:
Dear AMBER users and developers,
I would like to ask you if the pmemd.cuda (pmemd.cuda.mpi) code can use all 3 GPUs at once for a replica exchange MD simulation.
I mean the situation when a workstation has exactly 3 GPUs.
Does it depend on the number of replicas employed?
Kind regards,
Nick
_______________________________________________
AMBER mailing list
AMBER.ambermd.org<mailto:AMBER.ambermd.org><mailto:
AMBER.ambermd.org>
https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwICAg&c=pZJPUDQ3SB9JplYbifm4nt2lEVG5pWx2KikqINpWlZM&r=LIQu8OlVNKmzfbMg9_5FnKrt9-DrdQBJXyFyocKAWXc&m=qPny4TQgnr8yaNcTs44DbrbBp0LuImT-Ai92mhnbeVs&s=7c28oECd5GVRhneEnjGF-UoBR34TxLMPzFYVi3CJBcg&e=
_______________________________________________
AMBER mailing list
AMBER.ambermd.org<mailto:
AMBER.ambermd.org>
https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=pZJPUDQ3SB9JplYbifm4nt2lEVG5pWx2KikqINpWlZM&r=LIQu8OlVNKmzfbMg9_5FnKrt9-DrdQBJXyFyocKAWXc&m=Q3l9JsMIB1ciy8p3sQ9uTopz2Gf0Uklhj2pud4rkMjU&s=n3neoBNXPGor3dUGvpgnpZhie4gY7Wqe7xrjFeezlog&e=
_______________________________________________
AMBER mailing list
AMBER.ambermd.org<mailto:
AMBER.ambermd.org>
https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=pZJPUDQ3SB9JplYbifm4nt2lEVG5pWx2KikqINpWlZM&r=LIQu8OlVNKmzfbMg9_5FnKrt9-DrdQBJXyFyocKAWXc&m=Q3l9JsMIB1ciy8p3sQ9uTopz2Gf0Uklhj2pud4rkMjU&s=n3neoBNXPGor3dUGvpgnpZhie4gY7Wqe7xrjFeezlog&e=
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Feb 16 2018 - 17:00:03 PST
