Actually (though this may be outdated advice), one might still want to use
the keyword in certain cases when using the GPU code and you know exchanges
should never occur.
At one point we observed that large energy differences in the GTI
alchemical code were tricking the PRNG routines into incorrectly accepting
exchanges when they definitely should not have been. It seemed to be a
mixture of overflow in the energy code and odd behavior in the PRNG (which
produces overly regular values in exchange for reproducible behavior across
GPU architectures).
Maybe this was fixed...
BKR
On Tue, Aug 16, 2022, 11:45 AM Adrian Roitberg via AMBER <amber.ambermd.org>
wrote:
> Yes, you are right, that is what that keyword is supposed to do.
>
> However, if you do not use it, the only thing that happens is that the
> remd will try to exchange the first and last replicas, as you mentioned.
> If they are too far apart, they will "never" exchange, which is the
> correct behavior.
>
> Bottom-line, there is nothing to gain from using it if you have an even
> number of replicas.
>
>
> adrian
>
>
> On 8/15/22 4:42 PM, Zhuoran Long wrote:
> > *[External Email]*
> >
> > Hi Adrian,
> >
> > The md can run after removing the gremd_acyc keyword. However it
> > involving exchange between the first and last replicas. In my case,
> > the generalized distance reaction coordinates are restrained to about
> > -2 Angstrom and +2 Angstrom in these two replicas.
> > The amber 2021 manual says, " gremd_acyc (H-REMD only) When set to 1,
> > the exchange between the first and the last replicas will not be
> > considered and odd numbers of replicas are allowed. Default: 0".
> > That's why I want to use this keyword.
> >
> > Thank you,
> > Zhuoran
> >
> > Adrian Roitberg via AMBER <amber.ambermd.org> 于2022年8月15日周一
> > 16:01写道：
> >
> > Hi
> >
> > As far as I know, that keyword has nothing to do with periodicity.
> > It is
> > there to enable an odd number of replicas, instead of the usual
> > even_number
> >
> > So, try to use an even number of replicas without gremd_acyc=1 and
> > see
> > what happens.
> >
> > adrian
> >
> >
> > On 8/15/22 3:48 PM, Zhuoran Long via AMBER wrote:
> > > [External Email]
> > >
> > > Hi,
> > >
> > > I‘m trying to use the H-REMD for umbrella sampling. I'm using a
> > generalized
> > > distance reaction coordinate, which is not periodic. Therefore I
> > tried the
> > > gremd_acyc keyword. However, it gives the "error in reading
> > namelist cntrl"
> > > using sander.MPI and amber version 20.11. This error disappears
> > when not
> > > including this keyword. I also tried version 20.06 and 22.00 but
> > neither
> > > version works. I need to use the semiempirical QM/MM so I can't
> > switch to
> > > pmemd.MPI.
> > > I notice that this keyword is introduced in Sept 2019. I can't
> > find any
> > > examples including the amber test codes.
> > > I paste my input file below. Any suggestions or examples are
> > appreciated.
> > >
> > > Best,
> > > Zhuoran
> > >
> > > HREMD-Umbrella sampling
> > > &cntrl
> > > imin = 0,
> > > irest = 1,
> > > iwrap = 1,
> > > ntx = 5,
> > > ntxo= 1,
> > > nstlim= 100,
> > > dt = 0.001,
> > > ntc = 1,
> > > ntf = 1,
> > > cut = 8,
> > > ntb = 2,
> > > ntp = 1,
> > > pres0 = 1.0,
> > > taup= 5.0,
> > > ntpr= 500,
> > > ntwx= 500,
> > > ntt = 3,
> > > gamma_ln = 1.0,
> > > temp0 = 300.0,
> > > ioutfm= 1,
> > > ifqnt = 1,
> > > nmropt= 1,
> > > ig = 4746,
> > > numexchg = 5,
> > > gremd_acyc=1
> > > /
> > > &qmmm
> > > qmmask=':1',
> > > qmcharge=-2,
> > > qm_theory='PM3',
> > > qmshake=1,
> > > qm_ewald=1, qm_pme=1,
> > > /
> > > &wt
> > > type='DUMPFREQ', istep1=500,
> > > /
> > > &wt
> > > type='END',
> > > /
> > > DISANG=disang_cocc_001.in
> > <
> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fdisang_cocc_001.in%2F&data=05%7C01%7Croitberg%40ufl.edu%7C69d8c9334abc498ce99408da7efeb8d9%7C0d4da0f84a314d76ace60a62331e1b84%7C0%7C0%7C637961929800309591%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ljdxwGn2r1JB9u%2B2LMRRmVBegr%2FAtbmLhLoRIuQBM5I%3D&reserved=0
> >
> > > DUMPAVE=cocc_001.dat
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > >
> >
> https://urldefense.com/v3/__https://nam10.safelinks.protection.outlook.com/?url=http*3A*2F*2Flists.ambermd.org*2Fmailman*2Flistinfo*2Famber&data=05*7C01*7Croitberg*40ufl.edu*7C61ed6b01d3e445fd25fa08da7ef763cd*7C0d4da0f84a314d76ace60a62331e1b84*7C0*7C0*7C637961898246393685*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000*7C*7C*7C&sdata=1*2B7TDF8o151aq2JZoJxZB*2FfH*2FNgP9X4IV*2FHuYiNnR2k*3D&reserved=0__;JSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSU!!BhJSzQqDqA!T5GBkXB835_rWxWnwz5PosVwB4iXL9fdgPRMgP5ryQQRwPaHckdWOI8VT64p_-6ImcFLiojeCyLbaprA$
> > <
> https://urldefense.com/v3/__https://nam10.safelinks.protection.outlook.com/?url=http*3A*2F*2Flists.ambermd.org*2Fmailman*2Flistinfo*2Famber&data=05*7C01*7Croitberg*40ufl.edu*7C61ed6b01d3e445fd25fa08da7ef763cd*7C0d4da0f84a314d76ace60a62331e1b84*7C0*7C0*7C637961898246393685*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000*7C*7C*7C&sdata=1*2B7TDF8o151aq2JZoJxZB*2FfH*2FNgP9X4IV*2FHuYiNnR2k*3D&reserved=0__;JSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSU!!BhJSzQqDqA!T5GBkXB835_rWxWnwz5PosVwB4iXL9fdgPRMgP5ryQQRwPaHckdWOI8VT64p_-6ImcFLiojeCyLbaprA$
> >
> >
> >
> > --
> > Dr. Adrian E. Roitberg (he/him/el)
> > V.T. and Louise Jackson Professor in Chemistry
> > Department of Chemistry
> > University of Florida
> > roitberg.ufl.edu
> > 352-392-6972
> >
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> >
> https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!BhJSzQqDqA!T5GBkXB835_rWxWnwz5PosVwB4iXL9fdgPRMgP5ryQQRwPaHckdWOI8VT64p_-6ImcFLiojeC6kZ0qN-$
> >
> >
> --
> Dr. Adrian E. Roitberg (he/him/el)
> V.T. and Louise Jackson Professor in Chemistry
> Department of Chemistry
> University of Florida
> roitberg.ufl.edu
> 352-392-6972
> _______________________________________________
> 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 Tue Aug 16 2022 - 11:30:03 PDT
