[gmx-users] Too many LINCS warnings with CaDA model, User-defined potential, Stochastic dynamics integrator

[Srinath Kompella]

Dear All,
  Using GROMACS-5.1.4, I am trying to reproduce the results of
a metal-ligand coordination-directed self-assembly study. The metal ion
Pd(II) is represented using a CaDA model where four identical dummy atoms
each with charge +0.5 are coplanarly attached to the Pd with charge 0.0. A
united-atom approach is applied for modeling the organic neutral ligand.
Since the system is not charge neutral, generalized reaction field method
is used for electrostatics. To take the solvent effects (treated implicitly
as a continuum medium) into account, SD integrator, different near- and
far-field dielectric constant values are set, and WCA potential is applied
for ligand-ligand interactions. All bonds are constrained using LINCS, and
the timestep is 5 fs.

  The authors of the original work had modified the source code
to employ WCA interaction. I am using GROMACS user-defined potential tables
for all the vdW interactions, i.e., WCA between the ligand atoms and LJ for
the rest. I have modified their topology file and made a .mdp file
accordingly (attached here).

  After energy minimizing the system, NVT equilibration crashes
with the error "too many LINCS warnings". From the log file, and through
visual inspection, I noticed that the bonds and angles of the cations
(Pd(II) CaDA model) change in a bizarre fashion and cause the error. I've
experimented with smaller timesteps and slow ramping of temperature, but
only for timestep of 3 fs or smaller, the simulations run without any LINCS
error. Since a timestep of 5 fs was used in the original work, and I would
like to use the same, kindly help me understand where I am doing wrong and
correct it.

  Thank you in advance.

-- 
Srinath Kompella
Ph.D. Research Scholar
Molecular Simulations Laboratory,
Chemistry and Physics of Materials Unit (CPMU),
Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR),
Jakkur - 560 064, Bangalore,
Karnataka State, India.
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

[gmx-users] (no subject)

[Soham Sarkar]

Hello Everyone,
   I want to calculate some properties of the solvent molecules within
6angstrom from the protein backbone, for that I need a index group which
only contain that number of solvent molecules in the index.ndx file so that
I can specifically select that index to calculate different parameters.
Please any one help me how would I do that?
Soham
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

[gmx-users] (no subject)

[Soham Sarkar]

Hello Everyone,
   I want to calculate some properties of the solvent molecules within
6angstrom from the protein backbone, for that I need a index group which
only contain that number of solvent molecules in the index.ndx file so that
I can specifically select that index to calculate different parameters.
Please any one help me how would I do that?
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

[gmx-users] pdb2gmx error due to terminal capping of peptide

[Neena Susan Eappen]

Hi GMX users,


  1.  I drew the structure of my target peptide on Pymol, added amide capping 
group (NHH) at C terminus of peptide
  2.  Typed in the command: gmx pdb2gmx –f A15K.pdb –o A15K.gro –ter –ignh
  3.  Chose AMBER99sb-ildn force field which has parameters for NH2 capping 
group
  4.  Fatal error
  5.  Changed naming of amide group in .rtp file of AMBER99sb-ildn from NH2 to 
NHH so that naming is similar to .pdb file.
  6.  Still fatal error.
The residues in the chain ALA16--NHH32 do not have a consistent type. The first 
residue has type 'Protein', while residue NHH32 is of type 'Other'. Either 
there is a mistake in your chain, or it includes nonstandard residue names that 
have not yet been added to the residuetypes.dat file in the GROMACS library 
directory.

Any insight would be appreciated,
Thank you,
Neena
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

Re: [gmx-users] Density of a droplet in spherical coordinate

[paul buscemi]

You MIGHT be able to use the vmd plugin  membplugin to measure the density 
https://sourceforge.net/p/membplugin/wiki/Home/ 
  treating the sphere as a 
membrane or if you know the total density  and the radius you should be able to 
construct an integral to fit the data  —   similar to the following 
https://math.boisestate.edu/~jaimos/classes/m175-45-summer2014/notes/notes1-4.pdf
 

 

or - since you say each molecule type - make a block or sphere of each one and 
use gmx energy

Paul

> On Jan 24, 2019, at 11:25 AM, Alex  wrote:
> 
> Dear gmx user,
> I have a droplet of some short molecules (it is not perfectly spherical
> though), I was wondering how to calculate the density of each molecule type
> respect to center of the droplet. In other words the question is how the
> density of each molecule type varies in the spherical coordinate along
> radius of droplet (r), theta and phi. I am not sure about gmx density as
> the output is just a function of r apparently.
> Any help or idea is highly appreciated.
> 
> Thanks
> Alex
> -- 
> Gromacs Users mailing list
> 
> * Please search the archive at 
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
> 
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> 
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
> mail to gmx-users-requ...@gromacs.org.

-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

Re: [gmx-users] Density of a droplet in spherical coordinate

[Dallas Warren]

gmx spatial, then covert the g(xy,z) from the output file into the d(r,
*ϴ,ϕ) you require?*

Catch ya,

Dr. Dallas Warren
Drug Delivery, Disposition and Dynamics
Monash Institute of Pharmaceutical Sciences, Monash University
381 Royal Parade, Parkville VIC 3052
dallas.war...@monash.edu
-
When the only tool you own is a hammer, every problem begins to resemble a
nail.


On Fri, 25 Jan 2019 at 04:26, Alex  wrote:

> Dear gmx user,
> I have a droplet of some short molecules (it is not perfectly spherical
> though), I was wondering how to calculate the density of each molecule type
> respect to center of the droplet. In other words the question is how the
> density of each molecule type varies in the spherical coordinate along
> radius of droplet (r), theta and phi. I am not sure about gmx density as
> the output is just a function of r apparently.
> Any help or idea is highly appreciated.
>
> Thanks
> Alex
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
>
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

[gmx-users] Density of a droplet in spherical coordinate

[Alex]

Dear gmx user,
I have a droplet of some short molecules (it is not perfectly spherical
though), I was wondering how to calculate the density of each molecule type
respect to center of the droplet. In other words the question is how the
density of each molecule type varies in the spherical coordinate along
radius of droplet (r), theta and phi. I am not sure about gmx density as
the output is just a function of r apparently.
Any help or idea is highly appreciated.

Thanks
Alex
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

Re: [gmx-users] Bennet error in FEP-calculations for charged ligands.

[Artem Shekhovtsov]

Michael, Thank you so much for such a quick and detailed response.
I tried to increase the number of intermediate lambda windows, but
unfortunately, it was not possible to reduce the error to an acceptable
level.

coul-lambdas  = 0.0 0.02 0.05 0.08 0.12 0.16 0.2 0.3 0.4 0.6 0.7 0.8 0.84
0.88 0.92 0.95 0.98 1.0

point  0 -  1,   DG 47.68 +/-  0.13
point  1 -  2,   DG 34.31 +/-  0.56
point  2 -  3,   DG 19.42 +/-  0.97
point  3 -  4,   DG 15.56 +/-  1.07
point  4 -  5,   DG  8.45 +/-  0.03
point  5 -  6,   DG  3.53 +/-  0.02
point  6 -  7,   DG  4.99 +/-  0.02
point  7 -  8,   DG  2.30 +/-  0.02
point  8 -  9,   DG -0.02 +/-  0.03
point  9 - 10,   DG -2.79 +/-  0.31
point 10 - 11,   DG -5.74 +/-  0.72
point 11 - 12,   DG -6.58 +/-  0.45
point 12 - 13,   DG -8.02 +/-  0.03
point 13 - 14,   DG -8.60 +/-  0.03
point 14 - 15,   DG -15.92 +/-  0.04
point 15 - 16,   DG -33.08 +/-  1.38
point 16 - 17,   DG -47.34 +/-  1.22

total  0 - 17,   DG  8.15 +/-  1.69

As for the three-level transformation, I agree that it should help, but it
seems I don’t quite understand how to implement this behavior in GROMACS.
I don't know by which options of the .mdp file it should be implemented and
how to build topology files for such transformations.
If you provided me with a small example of an mdp and itp file for such a
transformation, I would be very grateful to you.

Thank you,
Shekhovtsov Artem

On Wed, Jan 23, 2019 at 7:00 PM Michael Shirts  wrote:

> As free energies get larger, then the error gets less accurate. So if it is
> reporting 18.51 +/-  2.95 for one of the intervals, that likely suggest
> there is very little overlap in that area.
> A total free energy difference of  in -6.12 +/-  3.19 for the
> transformation indicates that the result is not very certain; you're within
> two standard deviations, and again, the Bennett error formula is inaccurate
> for larger error. I would suggest adding 1-2 more intermediates between
> those points.
>
> https://github.com/MobleyLab/alchemical-analysis provides some of the
> tools
> to check overlap.
>
> I would also suggest doing a multi stage transformation to increase overlap
> - for atoms that are disappearing or appearing that are charged, turn off
> charges off, change vdw, then turn charges back on. In some cases,
> especially when the atoms involved have different sizes, sc-coul can lead
> to some semi-pathological cases when vdw softcore and coul softcore result
> in some very low/high potentials at intermediates.
>
> On Wed, Jan 23, 2019 at 8:21 AM Artem Shekhovtsov <
> job.shekhovt...@gmail.com>
> wrote:
>
> > Hi all!
> > I encountered an error in my relative free energy calculations and do not
> > know how to fix it.
> > Molecules for which I want to carry out calculations contain a carboxyl
> > group.
> > To validate the protocol, I tried to run the fep-calculations of
> symmetric
> > molecules for which the change in free energy will be zero.
> > During validation, I was faced with the fact that the convergence error
> for
> > charged ligands greatly exceeds that for neutral ones.
> > For example, for the case of m-methyl benzoic acid [O-]C(= O)c1cc(C)ccc1:
> > Mutation (-CH3 + H) for one meta position (-H + CH3) for another meta
> > position relative to the carboxyl group.
> >
> > Ionized acid (solvent leg, 5 ns):
> > point  0 -  1,   DG 28.74 +/-  0.02
> > point  1 -  2,   DG 53.79 +/-  0.41
> > point  2 -  3,   DG 27.76 +/-  2.45
> > point  3 -  4,   DG 18.51 +/-  2.95
> > point  4 -  5,   DG  8.80 +/-  0.79
> > point  5 -  6,   DG  4.04 +/-  0.07
> > point  6 -  7,   DG -0.06 +/-  0.04
> > point  7 -  8,   DG -2.62 +/-  0.33
> > point  8 -  9,   DG -8.95 +/-  0.27
> > point  9 - 10,   DG -23.89 +/-  1.61
> > point 10 - 11,   DG -29.32 +/-  1.29
> > point 11 - 12,   DG -54.15 +/-  0.08
> > point 12 - 13,   DG -28.79 +/-  0.03
> >
> > total  0 - 13,   DG -6.12 +/-  3.19
> >
> > Unionized acid (solvent leg, 5 ns):
> > point  0 -  1,   DG  0.08 +/-  0.01
> > point  1 -  2,   DG -6.88 +/-  0.10
> > point  2 -  3,   DG -14.69 +/-  0.05
> > point  3 -  4,   DG -21.07 +/-  0.03
> > point  4 -  5,   DG -13.21 +/-  0.02
> > point  5 -  6,   DG -7.46 +/-  0.02
> > point  6 -  7,   DG -0.09 +/-  0.03
> > point  7 -  8,   DG  7.37 +/-  0.02
> > point  8 -  9,   DG 13.24 +/-  0.03
> > point  9 - 10,   DG 21.15 +/-  0.06
> > point 10 - 11,   DG 14.72 +/-  0.03
> > point 11 - 12,   DG  6.98 +/-  0.07
> > point 12 - 13,   DG -0.05 +/-  0.01
> >
> > total  0 - 13,   DG  0.09 +/-  0.22
> >
> > For a neutral molecule containing a charged carboxyl and amino groups
> > ([O-]C(=O)c1cc(C)c([NH3+])cc1) the error is small:
> 

Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

[Maity, Sayan]

Hi Mark,


Thanks. Though I understood about gen-seed which is irrelevant once the 
velocity has been generated (from Boltzmann distribution) but still I am a bit 
confused about ld-seed.


Is this ld-seed has a dependence to the configuration of the machine i.e. no of 
cores and processor...?


Sayan

--

Sayan Maity

Workgroup of Prof. Dr. Ulrich Kleinekathöfer

Ph.D. Student/Research Associate

Jacobs University Bremen gGmbH, Germany


https://www.jacobs-university.de/directory/maity


From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se 
 on behalf of Mark Abraham 

Sent: Thursday, January 24, 2019 10:28:59 AM
To: gmx-us...@gromacs.org
Subject: Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

Hi,

On Thu, 24 Jan 2019 at 10:22 Maity, Sayan 
wrote:

> Hi Mark,
>
>
> So basically, in order to reproduce the simulation, I need to fix both
> seed numbers in the .mdp file right ...??
>

No, I said the seeds were fixed *after* you run grompp. That is, the .tpr
is in principle reproducible, caveat the usual things
http://manual.gromacs.org/current/user-guide/managing-simulations.html#reproducibility.
That's because the velocities have already been generated (if requested
with gen-vel, and so gen-seed is then irrelevant), and ld-seed has a
definite value (as you can see with gmx dump, or in the subsequent mdrun
log file).


> However, I can find the ld-seed number from the .log file but I didn't
> find any gen-seed number. So in which file I could get the gen-seed number
> if the previous simulation was done considering gen-seed=-1??
>

grompp reported it in its terminal output. But you never need it again.

Another question is that,
>
>
> Is ld-seed number is also taking random velocity from Boltzmann
> distribution...?


No, some algorithm that you haven't mentioned may be using numbers from a
random number generator that was seeded with that value :-)

So, if gen-seeds are same but ld-seeds are different, can I call it as a
> different simulation for the same starting structure...?
>

The trajectories will diverge soon.

Mark


> Thanks.
>
> Sayan
>
>
> --
>
> Sayan Maity
>
> Workgroup of Prof. Dr. Ulrich Kleinekathöfer
>
> Ph.D. Student/Research Associate
>
> Jacobs University Bremen gGmbH, Germany
>
>
> https://www.jacobs-university.de/directory/maity
>
> 
> From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se <
> gromacs.org_gmx-users-boun...@maillist.sys.kth.se> on behalf of Mark
> Abraham 
> Sent: Thursday, January 24, 2019 9:20:41 AM
> To: gmx-us...@gromacs.org
> Subject: Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.
>
> Hi,
>
> Both these seeds are fixed after you run grompp. That has no relationship
> to whether velocities are present from any input file. There is also no
> difference between NVT and NPT.
>
> I assume that your non-reproducible cases are each running grompp with one
> of these seed values set to -1, in which case grompp can picks a different
> value each time you run it. Otherwise, please show your command lines :-)
>
> Mark
>
> On Wed, 23 Jan 2019 at 16:55 Maity, Sayan 
> wrote:
>
> > Dear All,
> >
> >
> > I have a doubt regarding the ld-seed no and .cpt file. If you repeat a
> > simulation with the same .cpt file but different ld-seed no, my guess I
> > should get the same trajectory running in the serial node (1node+1core ),
> > as because of all the velocities are stored in the .cpt file. I did this
> > test but the trajectories are different unless I fixed the ld-seed no.
> So,
> > my point is what is the significance of using .cpt file when you are
> doing
> > a production run from an equilibration as velocity is already there in
> the
> > .gro file?? and why ld-seed no is controlling the dynamics not the
> > gen-seed? I checked that for nvt I need to specify both gen-seed and
> > ld-seed in order to reproduce the trajectory, but from npt, I only need
> the
> > ld-seed to reproduce the trajectory. Could you please clarify my
> doubts...??
> >
> >
> > Thanks.
> >
> >
> > Regards,
> >
> > Sayan
> >
> > --
> >
> > Sayan Maity
> >
> > Workgroup of Prof. Dr. Ulrich Kleinekathöfer
> >
> > Ph.D. Student/Research Associate
> >
> > Jacobs University Bremen gGmbH, Germany
> >
> >
> > https://www.jacobs-university.de/directory/maity
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > send a mail to gmx-users-requ...@gromacs.org.
> >
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe 

Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

[Mark Abraham]

Hi,

On Thu, 24 Jan 2019 at 10:22 Maity, Sayan 
wrote:

> Hi Mark,
>
>
> So basically, in order to reproduce the simulation, I need to fix both
> seed numbers in the .mdp file right ...??
>

No, I said the seeds were fixed *after* you run grompp. That is, the .tpr
is in principle reproducible, caveat the usual things
http://manual.gromacs.org/current/user-guide/managing-simulations.html#reproducibility.
That's because the velocities have already been generated (if requested
with gen-vel, and so gen-seed is then irrelevant), and ld-seed has a
definite value (as you can see with gmx dump, or in the subsequent mdrun
log file).


> However, I can find the ld-seed number from the .log file but I didn't
> find any gen-seed number. So in which file I could get the gen-seed number
> if the previous simulation was done considering gen-seed=-1??
>

grompp reported it in its terminal output. But you never need it again.

Another question is that,
>
>
> Is ld-seed number is also taking random velocity from Boltzmann
> distribution...?


No, some algorithm that you haven't mentioned may be using numbers from a
random number generator that was seeded with that value :-)

So, if gen-seeds are same but ld-seeds are different, can I call it as a
> different simulation for the same starting structure...?
>

The trajectories will diverge soon.

Mark


> Thanks.
>
> Sayan
>
>
> --
>
> Sayan Maity
>
> Workgroup of Prof. Dr. Ulrich Kleinekathöfer
>
> Ph.D. Student/Research Associate
>
> Jacobs University Bremen gGmbH, Germany
>
>
> https://www.jacobs-university.de/directory/maity
>
> 
> From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se <
> gromacs.org_gmx-users-boun...@maillist.sys.kth.se> on behalf of Mark
> Abraham 
> Sent: Thursday, January 24, 2019 9:20:41 AM
> To: gmx-us...@gromacs.org
> Subject: Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.
>
> Hi,
>
> Both these seeds are fixed after you run grompp. That has no relationship
> to whether velocities are present from any input file. There is also no
> difference between NVT and NPT.
>
> I assume that your non-reproducible cases are each running grompp with one
> of these seed values set to -1, in which case grompp can picks a different
> value each time you run it. Otherwise, please show your command lines :-)
>
> Mark
>
> On Wed, 23 Jan 2019 at 16:55 Maity, Sayan 
> wrote:
>
> > Dear All,
> >
> >
> > I have a doubt regarding the ld-seed no and .cpt file. If you repeat a
> > simulation with the same .cpt file but different ld-seed no, my guess I
> > should get the same trajectory running in the serial node (1node+1core ),
> > as because of all the velocities are stored in the .cpt file. I did this
> > test but the trajectories are different unless I fixed the ld-seed no.
> So,
> > my point is what is the significance of using .cpt file when you are
> doing
> > a production run from an equilibration as velocity is already there in
> the
> > .gro file?? and why ld-seed no is controlling the dynamics not the
> > gen-seed? I checked that for nvt I need to specify both gen-seed and
> > ld-seed in order to reproduce the trajectory, but from npt, I only need
> the
> > ld-seed to reproduce the trajectory. Could you please clarify my
> doubts...??
> >
> >
> > Thanks.
> >
> >
> > Regards,
> >
> > Sayan
> >
> > --
> >
> > Sayan Maity
> >
> > Workgroup of Prof. Dr. Ulrich Kleinekathöfer
> >
> > Ph.D. Student/Research Associate
> >
> > Jacobs University Bremen gGmbH, Germany
> >
> >
> > https://www.jacobs-university.de/directory/maity
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > send a mail to gmx-users-requ...@gromacs.org.
> >
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
>
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit

Re: [gmx-users] Nonsense timing accounting with OpenCL in Haswell

[Schulz, Roland]

Hi Elton,

It is very unlikely that you would be able to get any speedup with OpenCL with 
that CPU. For its generation the CPU is powerful but the GPU is a lower tier. 
To get meaningful speedup compared to running on i7 CPU you want to have a GT3 
or higher GPU with at least 48EUs. Yours has 20EUs.
That the timing is nonsense might be caused by beignet. We haven’t tested 
beignet with GROMACS.
Other than for the GPU/CPU load balancing (which you could do manually) the 
timing shouldn’t affect the performance. So you can try runs with and without 
GPUs to determine what the speedup/slowdown is of using the GPU.
If you are interested to find out what exactly goes wrong with the timing, you 
might want to look at getLastRangeTime in 
src/gromacs/gpu_utils/gpuregiontimer_ocl.h. Maybe add some extra debug printing 
to see whether you notice a few abnormal timings or whether all are nonsensical.

Roland


From: Elton Carvalho mailto:elto...@gmail.com>>
Date: Thu, 24 Jan 2019 at 04:06
Subject: [gmx-users] Nonsense timing accounting with OpenCL in Haswell
To: Discussion list for GROMACS users 
mailto:gmx-us...@gromacs.org>>


Greetings!

I'm trying to set up  gromacs-2019 to use OpenCL with my Intel GPU,
integrated in a Haswell processor. The log file says it's detected as

Intel(R) HD Graphics Haswell GT2 Desktop, vendor: Intel, device version:
OpenCL 1.2 beignet 1.3, stat: compatible

I'm running beignet as the OpenCL driver because the NEO drivers don't seem
to support Haswell.

Testing with the "NADP-DEPENDENT ALCOHOL DEHYDROGENASE in water" benchmark,
as available in the "adh_cubic_vsites" directory, I get *MUCH* slower
performance and some really weird timings at the end of the logfile. Things
like:

1)  "Launch GPU ops." taking almost 90% of the run time
2) "Nonbonded F kernel" in the GPU times with nonsense readings such as "
5589922469 ms/step" in a 30-minute test run.
3) 110680465055.927 seconds of Total walltime in the GPU in a 30-minute
real-time run.

My questions are:
a) Could this nonsense timing be coming from beignet, which is not _really_
supported? If not, wherefrom?
b) How can I troubleshoot this and get sensible timings to decide whether
using OpenCL in this machine is even worth it?
c) Is Intel OpenCL even wirth it in a Haswell machine ( i7-4790)? :)

The whole log is available at
https://gist.github.com/eltonfc/dd8755bce756b627464df70faa9d3bab and the
relevant parts are below:

[ LOG FILE BEGINS ]
Command line:
  gmx mdrun -v -maxh .5 -notunepme

GROMACS version:2019
Precision:  single
Memory model:   64 bit
MPI library:thread_mpi
OpenMP support: enabled (GMX_OPENMP_MAX_THREADS = 64)
GPU support:OpenCL
SIMD instructions:  AVX2_256
FFT library:fftw-3.3.6-pl2-fma-sse2-avx-avx2-avx2_128
RDTSCP usage:   enabled
TNG support:enabled
Hwloc support:  hwloc-1.11.2
Tracing support:disabled
C compiler: /usr/bin/cc GNU 7.3.0
C compiler flags:-mavx2 -mfma -O3 -DNDEBUG -funroll-all-loops
-fexcess-precision=fast
C++ compiler:   /usr/bin/c++ GNU 7.3.0
C++ compiler flags:  -mavx2 -mfma-std=c++11   -O3 -DNDEBUG
-funroll-all-loops -fexcess-precision=fast
OpenCL include dir: /usr/include
OpenCL library: /usr/lib/libOpenCL.so
OpenCL version: 2.0


Running on 1 node with total 4 cores, 8 logical cores, 1 compatible GPU
Hardware detected:
  CPU info:
Vendor: Intel
Brand:  Intel(R) Core(TM) i7-4790 CPU @ 3.60GHz
Family: 6   Model: 60   Stepping: 3
Features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma hle htt intel
lahf mmx msr nonstop_tsc pcid pclmuldq pdcm pdpe1gb popcnt pse rdrnd rdtscp
rtm sse2 sse3 sse4.1 sse4.2 ssse3 tdt x2apic
  Hardware topology: Full, with devices
Sockets, cores, and logical processors:
  Socket  0: [   0   4] [   1   5] [   2   6] [   3   7]
Numa nodes:
  Node  0 (16704245760 bytes mem):   0   1   2   3   4   5   6   7
  Latency:
   0
 0  1.00
Caches:
  L1: 32768 bytes, linesize 64 bytes, assoc. 8, shared 2 ways
  L2: 262144 bytes, linesize 64 bytes, assoc. 8, shared 2 ways
  L3: 8388608 bytes, linesize 64 bytes, assoc. 16, shared 8 ways
PCI devices:
  :00:02.0  Id: 8086:0412  Class: 0x0300  Numa: 0
  :00:19.0  Id: 8086:153a  Class: 0x0200  Numa: 0
  :00:1f.2  Id: 8086:8c02  Class: 0x0106  Numa: 0
  GPU info:
Number of GPUs detected: 1
#0: name: Intel(R) HD Graphics Haswell GT2 Desktop, vendor: Intel,
device version: OpenCL 1.2 beignet 1.3, stat: compatible

[... skipping ... ]

Changing rlist from 0.935 to 0.956 for non-bonded 4x2 atom kernels

Changing nstlist from 10 to 40, rlist from 0.956 to 1.094

Using 1 MPI thread
Using 8 OpenMP threads

1 GPU auto-selected for this run.
Mapping of GPU IDs to the 1 GPU task in the 1 rank on this node:
  PP:0
PP tasks will do (non-perturbed) short-ranged interactions on the GPU
Pinning threads with an auto-selected logical core stride of 1

Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

[Maity, Sayan]

Hi Mark,


So basically, in order to reproduce the simulation, I need to fix both seed 
numbers in the .mdp file right ...??


However, I can find the ld-seed number from the .log file but I didn't find any 
gen-seed number. So in which file I could get the gen-seed number if the 
previous simulation was done considering gen-seed=-1??



Another question is that,


Is ld-seed number is also taking random velocity from Boltzmann 
distribution...? So, if gen-seeds are same but ld-seeds are different, can I 
call it as a different simulation for the same starting structure...?


Thanks.

Sayan


--

Sayan Maity

Workgroup of Prof. Dr. Ulrich Kleinekathöfer

Ph.D. Student/Research Associate

Jacobs University Bremen gGmbH, Germany


https://www.jacobs-university.de/directory/maity


From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se 
 on behalf of Mark Abraham 

Sent: Thursday, January 24, 2019 9:20:41 AM
To: gmx-us...@gromacs.org
Subject: Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

Hi,

Both these seeds are fixed after you run grompp. That has no relationship
to whether velocities are present from any input file. There is also no
difference between NVT and NPT.

I assume that your non-reproducible cases are each running grompp with one
of these seed values set to -1, in which case grompp can picks a different
value each time you run it. Otherwise, please show your command lines :-)

Mark

On Wed, 23 Jan 2019 at 16:55 Maity, Sayan 
wrote:

> Dear All,
>
>
> I have a doubt regarding the ld-seed no and .cpt file. If you repeat a
> simulation with the same .cpt file but different ld-seed no, my guess I
> should get the same trajectory running in the serial node (1node+1core ),
> as because of all the velocities are stored in the .cpt file. I did this
> test but the trajectories are different unless I fixed the ld-seed no. So,
> my point is what is the significance of using .cpt file when you are doing
> a production run from an equilibration as velocity is already there in the
> .gro file?? and why ld-seed no is controlling the dynamics not the
> gen-seed? I checked that for nvt I need to specify both gen-seed and
> ld-seed in order to reproduce the trajectory, but from npt, I only need the
> ld-seed to reproduce the trajectory. Could you please clarify my doubts...??
>
>
> Thanks.
>
>
> Regards,
>
> Sayan
>
> --
>
> Sayan Maity
>
> Workgroup of Prof. Dr. Ulrich Kleinekathöfer
>
> Ph.D. Student/Research Associate
>
> Jacobs University Bremen gGmbH, Germany
>
>
> https://www.jacobs-university.de/directory/maity
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
>
--
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

Re: [gmx-users] Doubts regarding gen-seed, ld-seed and .cpt file.

[Mark Abraham]

Hi,

Both these seeds are fixed after you run grompp. That has no relationship
to whether velocities are present from any input file. There is also no
difference between NVT and NPT.

I assume that your non-reproducible cases are each running grompp with one
of these seed values set to -1, in which case grompp can picks a different
value each time you run it. Otherwise, please show your command lines :-)

Mark

On Wed, 23 Jan 2019 at 16:55 Maity, Sayan 
wrote:

> Dear All,
>
>
> I have a doubt regarding the ld-seed no and .cpt file. If you repeat a
> simulation with the same .cpt file but different ld-seed no, my guess I
> should get the same trajectory running in the serial node (1node+1core ),
> as because of all the velocities are stored in the .cpt file. I did this
> test but the trajectories are different unless I fixed the ld-seed no. So,
> my point is what is the significance of using .cpt file when you are doing
> a production run from an equilibration as velocity is already there in the
> .gro file?? and why ld-seed no is controlling the dynamics not the
> gen-seed? I checked that for nvt I need to specify both gen-seed and
> ld-seed in order to reproduce the trajectory, but from npt, I only need the
> ld-seed to reproduce the trajectory. Could you please clarify my doubts...??
>
>
> Thanks.
>
>
> Regards,
>
> Sayan
>
> --
>
> Sayan Maity
>
> Workgroup of Prof. Dr. Ulrich Kleinekathöfer
>
> Ph.D. Student/Research Associate
>
> Jacobs University Bremen gGmbH, Germany
>
>
> https://www.jacobs-university.de/directory/maity
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
>
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.