Changed to 1 fs time step (dt=0.01), and it was able to overcome it. After
the first couple of MC exchanges the vlimit errors went away. I think to
continue I"ll just have to run a few short simulations on cpu first with
constant pH using a shorter time step, before switching over to pmemd.cuda
w/ a 2 fs time step.
Andrew
On Wed, Oct 11, 2017 at 12:24 PM, Andrew Schaub <aschaub.uci.edu> wrote:
> Prof Case,
>
> Thank you for your reply. I tried to run using pmemd.cuda (ntpr=1), and it
> wouldn't start. So I did as you suggested and tried sander.MPI (ntpr=1). It
> dies after 99 steps with Segmentation fault.
>
> The following is the last output of the output file using ntpr=1:
>
> NSTEP = 98 TIME(PS) = 200400.196 TEMP(K) = 302.23 PRESS =
> 0.0
> Etot = -107550.0636 EKtot = 27340.4833 EPtot =
> -134890.5469
> BOND = 1104.0381 ANGLE = 3182.9964 DIHED =
> 4089.4253
> 1-4 NB = 1395.0539 1-4 EEL = 15147.3553 VDWAALS =
> 14484.0230
> EELEC = -174293.4389 EHBOND = 0.0000 RESTRAINT =
> 0.0000
> Ewald error estimate: 0.1138E-04
> -----------------------------------------------------------
> -------------------
>
>
> NSTEP = 99 TIME(PS) = 200400.198 TEMP(K) = 302.20 PRESS
> = 0.0
> Etot = -107551.1489 EKtot = 27338.1732 EPtot =
> -134889.3221
> BOND = 1100.2361 ANGLE = 3171.3757 DIHED =
> 4089.9436
> 1-4 NB = 1395.5912 1-4 EEL = 15149.1321 VDWAALS =
> 14492.0234
> EELEC = -174287.6243 EHBOND = 0.0000 RESTRAINT =
> 0.0000
> Ewald error estimate: 0.7658E-05
> -----------------------------------------------------------
> -------------------
>
>
>
>
>
>
>
>
>
>
>
> *vlimit exceeded for step 4; vmax = 27.4499vlimit exceeded for
> step 5; vmax = 37.5738vlimit exceeded for step 6; vmax =
> 48.8041vlimit exceeded for step 7; vmax = 48.0581vlimit exceeded
> for step 7; vmax = 27.8161vlimit exceeded for step 9; vmax
> = 26.7491vlimit exceeded for step 10; vmax = 21.1430vlimit
> exceeded for step 8; vmax = 20.5914vlimit exceeded for step 13;
> vmax = 24.1324vlimit exceeded for step 14; vmax = 29.7639vlimit
> exceeded for step 11; vmax = 23.6826*
>
> My input file for the constant pH simulation is as follows:
>
> Explicit solvent constant pH MD
> &cntrl
> imin=0, irest=1, ntx=5, ntxo=2,
> ntpr=1, ntwx=200,
> nstlim=50000000, dt=0.002,
> ntt=3, tempi=300,
> temp0=300, gamma_ln=5.0, ig=-1,
> ntc=2, ntf=2, cut=8, iwrap=1,
> ioutfm=1, icnstph=2, ntcnstph=100,
> solvph=11.0, ntrelax=200, saltcon=0.1,
> /
>
> I was able to run 200 ns of free MD before switching to constant pH. I'm
> assuming ntrelax=200 is 200 femtoseconds? If my timestep is 2 fs, it looks
> like it's crashing during the proton exchange monte carlo step. All of my
> initial states are 0My constant pH input file (cpin):
>
> $ cat diketidemlc.cpin
> &CNSTPH
> CHRGDAT=-0.3479,0.2747,-0.24,0.1426,-0.0094,0.0362,0.0362,
> 0.0187,0.0103,0.0103,
> -0.0479,0.0621,0.0621,-0.0143,0.1135,0.1135,-0.3854,0.
> 34,0.34,0.34,0.7341,
> -0.5894,-0.3479,0.2747,-0.24,0.1426,-0.10961,0.034,0.034,0.06612,0.01041,
> 0.01041,-0.03768,0.01155,0.01155,0.32604,-0.03358,-0.03358,-1.03581,0.0,
> 0.38604,0.38604,0.7341,-0.5894,-0.3479,0.2747,-0.1354,0.
> 1212,-0.0414,0.081,
> 0.081,-0.0012,-0.1513,0.3866,-0.017,0.2681,-0.1718,0.3911,-0.1141,0.2317,
> 0.7341,-0.5894,-0.3479,0.2747,-0.1354,0.1212,-0.111,0.
> 0402,0.0402,-0.0266,
> -0.3811,0.3649,0.2057,0.1392,-0.5727,0.0,0.1292,0.1147,0.
> 7341,-0.5894,-0.3479,
> 0.2747,-0.1354,0.1212,-0.1012,0.0367,0.0367,0.1868,-0.
> 5432,0.0,0.1635,0.1435,
> -0.2795,0.3339,-0.2207,0.1862,0.7341,-0.5894,-0.4157,0.
> 2719,0.0213,0.1124,
> -0.1231,0.1112,0.1112,-0.3119,0.1933,0.5973,-0.5679,-0.
> 4157,0.2719,0.0213,
> 0.1124,-0.3593,0.1122,0.1122,-0.8844,0.0,0.5973,-0.5679,-0.
> 4157,0.2719,-0.0014,
> 0.0876,-0.0152,0.0295,0.0295,-0.0011,-0.1906,0.1699,-0.
> 2341,0.1656,0.3226,
> -0.5579,0.3992,-0.2341,0.1656,-0.1906,0.1699,0.5973,-0.
> 5679,-0.4157,0.2719,
> -0.0014,0.0876,-0.0858,0.019,0.019,-0.213,-0.103,0.132,-0.
> 498,0.132,0.777,
> -0.814,0.0,-0.498,0.132,-0.103,0.132,0.5973,-0.5679,
> PROTCNT=3,2,2,1,1,1,0,1,0,
> RESNAME='System: Unknown','Residue: LYS 41','Residue: LYS 43',
> 'Residue: HIP 57','Residue: HIP 147','Residue: CYS 183','Residue: HIP
> 243',
> 'Residue: HIP 252','Residue: TYR 254','Residue: LYS 255','Residue: HIP
> 289',
> RESSTATE=0,0,0,0,0,0,0,0,0,0,
> STATEINF(0)%FIRST_ATOM=635, STATEINF(0)%FIRST_CHARGE=0,
> STATEINF(0)%FIRST_STATE=0, STATEINF(0)%NUM_ATOMS=22,
> STATEINF(0)%NUM_STATES=2,
> STATEINF(1)%FIRST_ATOM=677, STATEINF(1)%FIRST_CHARGE=0,
> STATEINF(1)%FIRST_STATE=0, STATEINF(1)%NUM_ATOMS=22,
> STATEINF(1)%NUM_STATES=2,
> STATEINF(2)%FIRST_ATOM=937, STATEINF(2)%FIRST_CHARGE=44,
> STATEINF(2)%FIRST_STATE=2, STATEINF(2)%NUM_ATOMS=18,
> STATEINF(2)%NUM_STATES=3,
> STATEINF(3)%FIRST_ATOM=2371, STATEINF(3)%FIRST_CHARGE=44,
> STATEINF(3)%FIRST_STATE=2, STATEINF(3)%NUM_ATOMS=18,
> STATEINF(3)%NUM_STATES=3,
> STATEINF(4)%FIRST_ATOM=2929, STATEINF(4)%FIRST_CHARGE=98,
> STATEINF(4)%FIRST_STATE=5, STATEINF(4)%NUM_ATOMS=11,
> STATEINF(4)%NUM_STATES=2,
> STATEINF(5)%FIRST_ATOM=3784, STATEINF(5)%FIRST_CHARGE=44,
> STATEINF(5)%FIRST_STATE=2, STATEINF(5)%NUM_ATOMS=18,
> STATEINF(5)%NUM_STATES=3,
> STATEINF(6)%FIRST_ATOM=3923, STATEINF(6)%FIRST_CHARGE=44,
> STATEINF(6)%FIRST_STATE=2, STATEINF(6)%NUM_ATOMS=18,
> STATEINF(6)%NUM_STATES=3,
> STATEINF(7)%FIRST_ATOM=3960, STATEINF(7)%FIRST_CHARGE=120,
> STATEINF(7)%FIRST_STATE=7, STATEINF(7)%NUM_ATOMS=21,
> STATEINF(7)%NUM_STATES=2,
> STATEINF(8)%FIRST_ATOM=3981, STATEINF(8)%FIRST_CHARGE=0,
> STATEINF(8)%FIRST_STATE=0, STATEINF(8)%NUM_ATOMS=22,
> STATEINF(8)%NUM_STATES=2,
> STATEINF(9)%FIRST_ATOM=4521, STATEINF(9)%FIRST_CHARGE=44,
> STATEINF(9)%FIRST_STATE=2, STATEINF(9)%NUM_ATOMS=18,
> STATEINF(9)%NUM_STATES=3,
> STATENE=-0.845507,0.000000,0.000000,-11.711591,-16.243597,
> 89.288609,0.000000,
> 0.000000,-78.199991,
> TRESCNT=10,CPHFIRST_SOL=5883, CPH_IGB=2, CPH_INTDIEL=1.0,
> /
>
> It initializes my cpout file, but it simply has one line:
>
> Solvent pH: 11.00000
>
> Best Regards,
>
> Andrew
>
> On Wed, Oct 4, 2017 at 5:44 AM, David A Case <david.case.rutgers.edu>
> wrote:
>
>> On Tue, Oct 03, 2017, Andrew Schaub wrote:
>> >
>> > Error: an illegal memory access was encountered launching kernel
>> > kClearForces
>> > cudaFree GpuBuffer::Deallocate failed an illegal memory access was
>> > encountered
>>
>> Messages like these confirm that bad things happened, but don't provide
>> any more information. Usual first debugging step is to re-run on a CPU,
>> which
>> sometimes give better error messages. Consider short runs with ntpr=1 to
>> get the best chance of finding out what is going on.
>>
>> ....dac
>>
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
>
>
> --
> Andrew Schaub
> Graduate Program in Chemical & Structural Biology
> Tsai Lab, http:///www.tsailabuci.com/ <http://www.tsailabuci.com/>
> Luo Lab, http://rayl0.bio.uci.edu/html/
> University of California, Irvine
> Irvine, CA 92697-2280
> 949-824-8829 <(949)%20824-8829> (lab)
> 949-877-9380 <(949)%20877-9380> (cell)
> aschaub.uci.edu <http://www.linkedin.com/pub/andrew-schaub/9a/907/382/>
>
>
--
Andrew Schaub
Graduate Program in Chemical & Structural Biology
Tsai Lab, http:///www.tsailabuci.com/ <http://www.tsailabuci.com/>
Luo Lab, http://rayl0.bio.uci.edu/html/
University of California, Irvine
Irvine, CA 92697-2280
949-824-8829 (lab)
949-877-9380 (cell)
aschaub.uci.edu <http://www.linkedin.com/pub/andrew-schaub/9a/907/382/>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Oct 11 2017 - 13:30:02 PDT
