[gmx-users] Problems extending a calculation (in steps and Benchmark)

[Justin A. Lemkul]

Rebeca García Fandiño wrote:
> Hello,
> 
> I am doing a simulation of a membrane + protein using ffG53a6 in Gromacs.
> 
> First I equilibrated at constant volume and then I changed to 
> semi-isotropic pressure.
> 
>  
> 
> This is my mdp file:
> 
> title           = EQUILIBRADO
> 
> cpp             = /usr/bin/cpp
> 
> include         = -I../top
> 
> define          =
> 
> constraints     = all-bonds
> 
> integrator      = md
> 
> dt              = 0.002
> 
> nsteps          = 100000
> 
> nstxout         = 1000
> 
> nstvout         = 1000
> 
> nstlog          = 1000
> 
> nstenergy       = 1000
> 
> nstxtcout       = 1000
> 
> nstlist         = 10
> 
> ns_type         = grid
> 
> rlist           = 1.0
> 
> coulombtype     = PME
> 
> rcoulomb        = 1.0
> 
> vdwtype         = cut-off
> 
> rvdw            = 1.4
> 
> pme_order       = 4
> 
> ewald_rtol      = 1e-5
> 
> optimize_fft    = yes
> 
> ;Berendsen temperature coupling is on
> 
> Tcoupl          = berendsen
> 
> tau_t           = 0.1     0.1    0.1
> 
> tc-grps         = Protein DOPC SOL_NA+_CL-
> 
> ref_t           = 310     310    310
> 
> ; Pressure coupling
> 
> Pcoupl                   = Parrinello-Rahman
> 
> Pcoupltype               = Semiisotropic
> 
> ; Time constant (ps), compressibility (1/bar) and reference P (bar)
> 
> tau-p                    = 1.0 1.0
> 
> compressibility          = 4.6E-5 4.6E-5
> 
> ref-p                    = 1.0 1.0
> 
> ;Generate velocities is on at 310 K
> 
> gen_vel         = yes
> 
> gen_temp        = 310
> 
> gen_seed        = 173529
> 
>  
> 
> W
> 
> hen I did
> 
>  
> 
> grompp -f equilibrado3.mdp -c equilibrado2.gro -p ionized_system.top -o 
> equilibrado3.tpr -n index.ndx
> 
> mdrun -v -deffnm equilibrado3
> 
>  
> 
> I got a Benchmark of :
> 
>  
> 
>                NODE (s)   Real (s)      (%)
> 
>        Time:  38960.000  38960.000    100.0
> 
>                        10h49:20
> 
>                (Mnbf/s)   (GFlops)   (ns/day)  (hour/ns)
> 
> Performance:     93.815     25.746      0.444     54.111
> 
>  
> 
> (My system is very big, so the Benchmark is reasonable)
> 
>  
> 
> The next I did was extending the calculation:
> 
>  
> 
> tpbconv -s equilibrado3.tpr -f equilibrado3.trr -e equilibrado3.edr -n 
> index.ndx -o equilibrado4.tpr -extend 200000
> 
>  
> 
> Surpringsingly, the calculation reached 504.722 steps (it should have 
> stopped at 300.000 since I had done 100.000 and extended to 200.000 
> more!) and if I look at the benchmark

What is the time given on the initial frame (from the logfile?)  What does 
gmxcheck tell you about your .trr and .edr files (are they the times you think 
they are?).  If you're seeing a time indicator of 504.722, that would be the 
time (in ps), not the step number, so it's almost as if your new run started 
over from 0 and ran for a short time.

What happens if you use grompp to extend your run, instead of tpbconv?  It might 
be interesting to see if there's a difference.  Also, you could try tpbconv 
-time 100000 -until 300000, also to see if it would make a difference.  This 
will test to see if your tpbconv executable is functioning correctly or not.

-Justin

> 
> I see:
> 
>                NODE (s)   Real (s)      (%)
> 
>        Time: 172782.000 172782.000    100.0
> 
>                        1d23h59:42
> 
>                (Mnbf/s)   (GFlops)   (ns/day)  (hour/ns)
> 
> Performance:     85.694     23.514    100.010      0.240
> 
>  
> 
> What is obviously incorrect.
> 
>  
> 
> Anybody knows what could be happening?
> 
>  
> 
> Thank you very much in advance,
> 
>  
> 
> Rebeca García Fandiño
> 
> Parc Cientific of Barcelona
> 
> 
> ------------------------------------------------------------------------
> ¡Es tiempo de inaugurar tu piscina! Entra en Eventos e invita a todos 
> tus amigos, crea paneles de discusión y sube las fotos más divertidas 
> Eventos <http://www.vivelive.com/eventos>
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

-- 
========================================

Justin A. Lemkul
Graduate Research Assistant
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================