[gmx-users] Low GPU utilization

[Mark Abraham]

Hi,

That sounds like it could just be normal for what you'd expect on a
small-ish system if your GPU is better than your CPU. Sharing a .log file
via a file-sharing service gives you a better chance of useful feedback.

Mark

On Mon, Feb 1, 2016 at 7:43 AM Hovakim Grabski <hovakim_grabski at yahoo.com>
wrote:

> Dear Gromacs users,
>
> I'm running a simulation between LASR receptor and Quercetin on Q6600, GTX
> 960, CPU utilization is about 90%, but GPU is only 25-30%(Nvidia X server
> tools).The system is only 28643 atoms and dt = 2 fs.
> This is the mdp settings:
> ; Run parametersdefine      = -DPOSRES_B ; Lets seeintegrator  = md
>  ; leap-frog integratornsteps      = 5000000  ; 2 * 500000 = 1000 ps (1
> ns)dt          = 0.002     ; 2 fs; Output controlnstxout     = 0         ;
> suppress .trr output nstvout     = 0         ; suppress .trr
> outputnstenergy   = 1000      ; save energies every 2 psnstlog      = 1000
>      ; update log file every 2 psnstxout-compressed  = 1000      ; write
> .xtc trajectory every 2 pscompressed-x-grps   = Systemenergygrps  = Protein
> HSL; Bond parameterscontinuation    = yes           ; first dynamics
> runconstraint_algorithm = lincs    ; holonomic constraints constraints
> = all-bonds     ; all bonds (even heavy atom-H bonds) constrainedlincs_iter
>      = 1             ; accuracy of LINCSlincs_order     = 4             ;
> also related to accuracy; Neighborsearchingcutoff-scheme   = Verletns_type
>     = grid      ; search neighboring grid cellsnstlist     = 10         ;
> 20 fsrcoulomb    = 1.4rvdw        = 1.4       ; short-range van der Waals
> cutoff (in nm); Electrostaticscoulombtype     = PME       ; Particle Mesh
> Ewald for long-range electrostaticspme_order       = 4         ; cubic
> interpolationfourierspacing  = 0.16      ; grid spacing for FFT;
> Temperature coupling is ontcoupl      = V-rescale                     ;
> modified Berendsen thermostattc-grps     = Protein_HSL Water_and_ions    ;
> two coupling groups - more accuratetau_t       = 0.1   0.1
>     ; time constant, in psref_t       = 300   300                     ;
> reference temperature, one for each group, in K; Pressure coupling is
> offpcoupl      = Parrinello-Rahman             ; pressure coupling is on
> for NPTpcoupltype  = isotropic                     ; uniform scaling of box
> vectorstau_p       = 2.0                           ; time constant, in
> psref_p       = 1.0                           ; reference pressure, in
> barcompressibility = 4.5e-5                    ; isothermal compressibility
> of water, bar^-1; Periodic boundary conditionspbc         = xyz       ; 3-D
> PBC; Dispersion correctionDispCorr    = EnerPres  ; account for cut-off vdW
> scheme; Velocity generationgen_vel     = no       ; assign velocities from
> Maxwell distributiongen_temp    = 300       ; temperature for Maxwell
> distributiongen_seed    = -1        ; generate a random
> seedrefcoord_scaling = com
>
> Best regards,Hovakim Grabski
> --
> 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-request at gromacs.org.