[gmx-users] restarting of shifted potential, 4.6.7 bug?

[Johnny Lu]

The system has 30k water and a protein with ~170 amino acid.
Here are the PME tuning:

npt14:
DD  step 4 load imb.: force 58.1%

step   10: timed with pme grid 80 80 80, coulomb cutoff 1.000: 37.1 M-cycles
step   20: timed with pme grid 64 64 64, coulomb cutoff 1.173: 32.6 M-cycles
step   30: timed with pme grid 56 56 56, coulomb cutoff 1.341: 34.5 M-cycles
step   40: timed with pme grid 48 48 48, coulomb cutoff 1.564: 40.7 M-cycles
step   50: timed with pme grid 72 72 72, coulomb cutoff 1.043: 33.9 M-cycles
step   60: timed with pme grid 64 64 64, coulomb cutoff 1.173: 32.2 M-cycles
step   70: timed with pme grid 60 60 60, coulomb cutoff 1.251: 32.8 M-cycles
step   80: timed with pme grid 56 56 56, coulomb cutoff 1.341: 32.2 M-cycles
step   90: timed with pme grid 52 52 52, coulomb cutoff 1.444: 32.2 M-cycles
step  100: timed with pme grid 72 72 72, coulomb cutoff 1.043: 33.8 M-cycles
step  110: timed with pme grid 64 64 64, coulomb cutoff 1.173: 40.3 M-cycles
step  120: timed with pme grid 60 60 60, coulomb cutoff 1.251: 32.4 M-cycles
step  130: timed with pme grid 56 56 56, coulomb cutoff 1.341: 33.0 M-cycles
step  140: timed with pme grid 52 52 52, coulomb cutoff 1.444: 32.1 M-cycles
              optimal pme grid 52 52 52, coulomb cutoff 1.444
DD  step 4999 load imb.: force 55.3%

npt15:

DD  step 4 load imb.: force 41.2%

step   10: timed with pme grid 80 80 80, coulomb cutoff 1.000: 36.6 M-cycles
step   20: timed with pme grid 64 64 64, coulomb cutoff 1.181: 32.9 M-cycles
step   30: timed with pme grid 56 56 56, coulomb cutoff 1.350: 33.6 M-cycles
step   40: timed with pme grid 48 48 48, coulomb cutoff 1.575: 33.2 M-cycles
step   50: timed with pme grid 44 44 44, coulomb cutoff 1.718: 35.8 M-cycles
step   60: timed with pme grid 40 40 40, coulomb cutoff 1.890: 40.8 M-cycles
step   70: timed with pme grid 80 80 80, coulomb cutoff 1.000: 36.5 M-cycles
step   80: timed with pme grid 72 72 72, coulomb cutoff 1.050: 33.8 M-cycles
step   90: timed with pme grid 64 64 64, coulomb cutoff 1.181: 32.0 M-cycles
step  100: timed with pme grid 60 60 60, coulomb cutoff 1.260: 32.3 M-cycles
step  110: timed with pme grid 56 56 56, coulomb cutoff 1.350: 31.6 M-cycles
step  120: timed with pme grid 52 52 52, coulomb cutoff 1.454: 32.1 M-cycles
step  130: timed with pme grid 48 48 48, coulomb cutoff 1.575: 32.3 M-cycles
step  140: timed with pme grid 44 44 44, coulomb cutoff 1.718: 35.3 M-cycles
step  150: timed with pme grid 42 42 42, coulomb cutoff 1.800: 37.7 M-cycles
step  160: timed with pme grid 72 72 72, coulomb cutoff 1.050: 33.5 M-cycles
step  170: timed with pme grid 64 64 64, coulomb cutoff 1.181: 32.0 M-cycles
step  180: timed with pme grid 60 60 60, coulomb cutoff 1.260: 31.9 M-cycles
step  190: timed with pme grid 56 56 56, coulomb cutoff 1.350: 31.4 M-cycles
step  200: timed with pme grid 52 52 52, coulomb cutoff 1.454: 31.6 M-cycles
step  210: timed with pme grid 48 48 48, coulomb cutoff 1.575: 32.0 M-cycles
              optimal pme grid 56 56 56, coulomb cutoff 1.350
DD  step 4999 load imb.: force 41.0%





On Tue, Nov 11, 2014 at 9:44 AM, Johnny Lu <johnny.lu128 at gmail.com> wrote:

> by the way, the machine has 4 Tesla K40m gpu, 32 xeon cpu with avx, and i
> ran: nohup ../../mdrun -deffnm npt15 -cpt 1 -cpnum &
>
> On Tue, Nov 11, 2014 at 9:38 AM, Johnny Lu <johnny.lu128 at gmail.com> wrote:
>
>> Hi.
>>
>> I ran 30ns NPT with Potential-shift-Verlet on both potentials, continuing
>> from a NPT without Potential-shift-Verlet. The energy dropped from -3e5 to
>> -4e5, which is fine.
>>
>> After that, I continue the 30ns NPT with the same mdp file, except with a
>> longer run time (300ns).
>>
>> The total energy was -4e5 at step 0, and then -3e5 at step 5000, and now
>> at about 25 ns, it goes back to -3.5e5. Why the energy suddenly changed
>> after step 0?
>>
>> I used the same machine and same mixed precision gromacs 4.6.7 mdrun and
>> grompp for both simulations.
>>
>> The plot of total energies for the two simulations:
>> http://oi60.tinypic.com/ddmejd.jpg
>> The two mdp files: http://oi58.tinypic.com/11980si.jpg
>>
>> The commands that I used for continuing the simulations:
>>
>> The previous simulation (npt14): ../../sofware/gromacs-4.6.7/bin/grompp
>> -f npt14.mdp -c npt13.tpr -t npt13_step22620750.cpt -o npt14.tpr
>>
>> The current simulation (npt15): ../../sofware/gromacs-4.6.7/bin/grompp -f
>> npt15.mdp -c npt14.tpr -t npt14_step30000000.cpt -o npt15.tpr
>>
>> Thanks.
>>
>>
>