[gmx-users] 100 step md run: Infinite giga flopses error

[Justin A. Lemkul]

Sai Pooja wrote:
> Hi,
> 
> I am running a very short md run for 100 steps and I get the Infinite 
> giga flopses error. I stored info at every step to check if the 
> simulation is running. And the log file has all the variables for each 
> time-step with reasonable  values. I don't understand the meaning of the 
> message at the end of the log file. 
> 
> nodetime = 0! Infinite Giga flopses!
> Finished mdrun on node 0 Tue Jun  8 10:16:31 2010
> 
> This is not the first time I have encountered this problem with short md 
> runs.
> 

This isn't really a "problem," per se.  Gromacs is fast.  If you look at the 
benchmarks in the Gromacs 4 paper, you can get several hundred steps of MD per 
second for a relatively small system.  So essentially your run probably finished 
in less than a second, and mdrun recorded that as "zero" time.  Division by zero 
= infinite performance :)

If you think there's something else actually wrong, please at least provide a 
description of how large your system is, how long the run actually took to 
complete, if you're running in parallel, etc.

-Justin

> Parameter file
> 
> ; VARIOUS PREPROCESSING OPTIONS  
> title                    =NPT simulation for a 2D WCA system
> cpp                      =/lib/cpp
> 
> ; RUN CONTROL PARAMETERS
> integrator               = md
> dt                       = 0.0001
> nsteps                   = 100
> 
> ; OUTPUT CONTROL OPTIONS
> nstxout                  = 0    ; No output, except for last frame 
> (coordinates)
> nstvout                  = 0    ; No output, except for last frame 
> (velocities)
> nstfout                  = 0    ; No output, except for last frame (forces)
> nstlog                   = 1   ; Write every nth step to the log
> nstenergy                = 0     ; Write energies at every n step
> nstxtcout                = 0     ; Do not write a compressed tr
> 
> ; NEIGHBORSEARCHING PARAMETERS
> nstlist                  = 10
> ns-type                  = Grid
> pbc                      = xyz
> rlist                    = 0.9
> 
> ; OPTIONS FOR ELECTROSTATICS AND VDW
> coulombtype              = Cut-off
> rcoulomb                 = 1.1225
> vdw-type                 = Cut-off
> rvdw                     = 1.1225
> 
> ; Temperature coupling
> tcoupl                   = v-rescale
> tc-grps                  = AR
> tau_t                    = 0.1
> ref_t                    = 0.1
> ld_seed                   =-1
> 
> ; Pressure coupling
> pcoupl                   = Berendsen
> pcoupltype               = semiisotropic
> tau-p                    = 0.1 0.1
> compressibility          = 1e-5  0   ;Dont know
> ref-p                    = 58 0
> 
> 
> ;Generate velocities for startup run
> gen_vel                  =yes
> gen_temp                 =1
> gen_seed                 =-1
> 
> ;Non equilibrium MD steps
> freezegrps              =AR
> freezedim               =N N Y
> 
> Error: At the end of log file
> 
> nodetime = 0! Infinite Giga flopses!
> Finished mdrun on node 0 Tue Jun  8 10:16:31 2010
> 
> 
> Regards
> Pooja
> 
> -- 
> Quaerendo Invenietis-Seek and you shall discover.
> 

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

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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