[gmx-users] MPI segmentation Fault
Justin Lemkul
jalemkul at vt.edu
Wed Jan 2 21:48:37 CET 2013
On 1/2/13 3:42 PM, Xu Dong Huang wrote:
> @Justin, here are the details of my NVT and NPT .mdp, (All ran without position restraint) , + details of the NVT run
>
> For NVT.mdp:
> integrator = md
> tinit = 0.0
> dt = 0.02
> nsteps = 10000
> nstcomm = 1
> comm-grps =
>
> nstxout = 0
> nstvout = 0
> nstfout = 0
> nstlog = 1000
> nstenergy = 100
> nstxtcout = 1000
> xtc_precision = 100
> xtc-grps =
> energygrps = system
> nstlist = 10
> ns_type = grid
> pbc = xyz
> rlist = 1.4
> coulombtype = Shift
> rcoulomb_switch = 0.0
> rcoulomb = 1.2
> epsilon_r = 15
> vdw_type = Shift
> rvdw_switch = 0.9
> rvdw = 1.2
> DispCorr = No
> tcoupl = V-rescale
> tc-grps = system
> tau_t = 1.0
> ref_t = 300
> Pcoupl = no
> Pcoupltype = isotropic
> tau_p = 1.0
> compressibility = 3e-4
> ref_p = 1.0
>
> gen_vel = yes
> gen_temp = 300
> gen_seed = 473529
>
> constraints = none
> constraint_algorithm = Lincs
> unconstrained_start = no
> lincs_order = 4
> lincs_warnangle = 30
>
> For NPT.mdp
>
> integrator = md
> tinit = 0.0
> dt = 0.02
> nsteps = 1000000
> nstcomm = 1
> comm-grps =
>
> nstxout = 0
> nstvout = 0
> nstfout = 0
> nstlog = 1000
> nstenergy = 100
> nstxtcout = 1000
> xtc_precision = 100
> xtc-grps =
> energygrps = system
> nstlist = 10
> ns_type = grid
> pbc = xyz
> rlist = 1.4
> coulombtype = Shift
> rcoulomb_switch = 0.0
> rcoulomb = 1.2
> epsilon_r = 15
> vdw_type = Shift
> rvdw_switch = 0.9
> rvdw = 1.2
> DispCorr = No
> tcoupl = V-rescale
> tc-grps = system
> tau_t = 1.0
> ref_t = 300
> Pcoupl = Parrinello-Rahman
> Pcoupltype = isotropic
> tau_p = 5.0
> compressibility = 3e-4
> ref_p = 1.0
>
Parrinello-Rahman is a really bad choice for the first phase of NPT
equilibration for all but the most robust systems. It allows for rather wide
box oscillations if the system is far from equilibrium. Try a weak coupling
method (Berendsen) to see if it runs, then move to P-R if things are stable.
What all is in the system? Is this that star-shaped structure you posted about
previously? Any solvent? A full description would be really helpful.
-Justin
> gen_vel = no
> gen_temp = 300
> gen_seed = 473529
>
> constraints = none
> constraint_algorithm = Lincs
> unconstrained_start = no
> lincs_order = 4
> lincs_warnangle = 30
>
> OUTPUT log FOR NVT RUN:
>
> Writing checkpoint, step 9760 at Wed Jan 2 06:43:55 2013
>
>
> Step Time Lambda
> 10000 200.00000 0.00000
>
> Writing checkpoint, step 10000 at Wed Jan 2 06:56:58 2013
>
>
> Energies (kJ/mol)
> Bond LJ (SR) Coulomb (SR) Potential Kinetic En.
> 1.27815e+03 -2.21982e+07 0.00000e+00 -2.21969e+07 2.99756e+06
> Total Energy Conserved En. Temperature Pressure (bar)
> -1.91994e+07 -2.18426e+07 2.99676e+02 7.09077e+02
>
> <====== ############### ==>
> <==== A V E R A G E S ====>
> <== ############### ======>
>
> Statistics over 10001 steps using 1001 frames
>
> Energies (kJ/mol)
> Bond LJ (SR) Coulomb (SR) Potential Kinetic En.
> 1.33393e+03 -2.22049e+07 0.00000e+00 -2.22036e+07 2.98604e+06
> Total Energy Conserved En. Temperature Pressure (bar)
> -1.92176e+07 -2.18431e+07 2.98524e+02 7.06152e+02
>
> Total Virial (kJ/mol)
> -9.41186e+05 7.05498e+01 6.65566e+02
> 7.05612e+01 -9.40675e+05 -3.05521e+00
> 6.65543e+02 -3.06027e+00 -9.44794e+05
>
> Pressure (bar)
> 7.05835e+02 -4.93053e-02 -2.47918e-01
> -4.93095e-02 7.05573e+02 -4.79105e-03
> -2.47909e-01 -4.78920e-03 7.07047e+02
>
> Total Dipole (D)
> 0.00000e+00 0.00000e+00 0.00000e+00
>
>
>
>
> Xu Dong Huang
> Chemical & Biochemical Engineering
> Rutgers School of Engineering
> xudongh at eden.rutgers.edu
>
> On Jan 2, 2013, at 3:20 PM, Xu Dong Huang <xudongh at eden.rutgers.edu> wrote:
>
>> Dear gromac users,
>>
>> after examining my nvt run outcome, I see that my molecule has half of the arm stuck outside of the box (I guess my run wasn't very successful). But I thought the jumping outside of the box would be corrected by periodic boundary? Should I run my nvt for longer since the structure itself did not seem to change much in the solute? (My initial
>> Xu Dong Huangnvt nstep=10000, should I increase it to resolve the problem? if that resolves the problem at all that is)
>>
>> Thanks for any input,
>>
>> Chemical & Biochemical Engineering
>> Rutgers School of Engineering
>> xudongh at eden.rutgers.edu
>>
>> On Jan 2, 2013, at 3:08 PM, Xu Dong Huang <xudongh at eden.rutgers.edu> wrote:
>>
>>> Dear gromac users,
>>>
>>> after running a successful NVT, I was going to run NPT. (All ran with MPI), however when I attempt to run the NPT, I get the following segmentation error:
>>>
>>> Getting Loaded...
>>> Reading file npt.tpr, VERSION 4.5.3 (single precision)
>>> Loaded with Money
>>>
>>>
>>> WARNING: This run will generate roughly 3890 Mb of data
>>>
>>> starting mdrun 'star'
>>> 1000000 steps, 20000.0 ps.
>>>
>>> step 0[Ultranode02:26884] *** Process received signal ***
>>> [Ultranode02:26884] Signal: Segmentation fault (11)
>>> [Ultranode02:26884] Signal code: Address not mapped (1)
>>> [Ultranode02:26884] Failing at address: 0x1f5acf020
>>> [Ultranode02:26884] [ 0] /lib/libpthread.so.0(+0xf8f0) [0x2b21605168f0]
>>> [Ultranode02:26884] [ 1] mdrun_mpi() [0x745509]
>>> [Ultranode02:26884] *** End of error message ***
>>> /usr/local/SGE/default/spool/Ultranode02/job_scripts/39346: line 38: 26884 Segmentation fault mdrun_mpi -v -deffnm npt
>>>
>>>
>>> -----Is my system blowing up? Or there is something wrong with MPI ? (I requested 4 processors)
>>>
>>> Xu Dong Huang
>>> Chemical & Biochemical Engineering
>>> Rutgers School of Engineering
>>> xudongh at eden.rutgers.edu
>>>
>>> --
>>> gmx-users mailing list gmx-users at gromacs.org
>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/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/Support/Mailing_Lists
>>
>> --
>> gmx-users mailing list gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/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/Support/Mailing_Lists
>
--
========================================
Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
More information about the gromacs.org_gmx-users
mailing list