Re: question about CUDA-enabled

From: Abhishek TYAGI (atyagiaa_at_connect.ust.hk)
Date: Sun Nov 30 2014 - 02:57:53 CST

Hi,

As Axel mentioned this topic is discussed several times you may follow the following links, and their are many more you can find the answers:

http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2010-2011/3979.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/2898.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2010-2011/3647.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2008-2009/0295.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2007-2008/3576.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2008-2009/1573.html
http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2008-2009/0300.html

your conf file looks OK, i suggest if you reduce timestep to 1, which is 2 in your file will help you, moreover, you can think about the value of piston period and piston decay. As Axel mentioned in NPT system "

This question has been asked and answered probably hundreds of times. It has nothing to do with CUDA. Please do a careful search of the mailing list archive. Occasionally people give bad advice, because they don't understand where the message is coming from (your box has shrunk too much and thus a new domain decomposition has to be done, which cannot be done on the fly)." is right.

________________________________
From: maryam atabay <matabay2010_at_gmail.com>
Sent: Sunday, November 30, 2014 3:41 PM
To: Abhishek TYAGI
Subject: Re: namd-l: question about CUDA-enabled

Hi Abhi
I had done you'r suggestion previosly and I reduced ts to 0.1 even but It not execute. It's intresting this conf file execute with NAMD-molticore but not with CUDA!
this is the md conf file I used it
####################################################################
#ENERGY MINIMIZATION AND EQUILIBRATION in NPT ensemble
####################################################################
#Molecular system
####################################################################
numsteps 25000000
structure ionized.psf
coordinates ionized.pdb
set outputname md
firsttimestep 0
set temperature 310
#####################################################################
# Input
#####################################################################
paraTypeCharmm on
parameters par_all36_cgenff.prm
parameters par_all27_prot_lipid_na.inp
#temperature $temperature
bincoordinates md.coor
extendedSystem md.xsc
binvelocities md.vel
######################################################################
# Force-Field Parameters
#####################################################################
exclude scaled1-4
1-4scaling 1.0
cutoff 12
switching on
switchdist 10
pairlistdist 14
margin 3
########################################################################
# Integrator Parameters
########################################################################
timestep 2.0
rigidBonds all
nonbondedFreq 1
fullElectFrequency 2
Stepspercycle 10
##########################################################################
#Position-restrained (based on chapter 2.4)
#############################################################################
fixedAtoms on
fixedAtomsFile fixed-atoms-bnn.pdb
fixedAtomsCol B
fixedAtomsForces off
################################################################################
# Constant Temperature Control
#################################################################################
langevin on # do langevin dynamics
langevinDamping 5 # damping coefficient (gamma) of 5/ps
langevinTemp $temperature
langevinHydrogen off # don't couple langevin bath to hydrogens
##################################################################################
# Periodic Boundary Conditions
####################################################################################
#cellBasisVector1 56 0.0 0.0
#cellBasisVector2 0.0 60 0.0
#cellBasisVector3 0.0 0.0 56
#cellOrigin -11.63747787475586 6.021169185638428 23.122079849243164
wrapWater on
wrapAll on
wrapnearest on
COMmotion no
#########################################################################################
# PME (for full-system periodic electrostatics)
#use numbers with small integer factor: 2,3,5
###########################################################################################
PME yes
PMEGridSizeX 72
PMEGridSizeY 72
PMEGridSizeZ 72
#########################################################################################
# Constant Pressure Control (variable volume)
##############################################################################################
langevinPiston on
langevinPistonTarget 1.01325 # in bar -> 1 atm
langevinPistonPeriod 100
langevinPistonDecay 50
langevinPistonTemp $temperature # sets the "noise" temp in K
#############################################################################################
# Output
#########################################################################################
outputName $outputname
binaryoutput yes
binaryrestart yes
dcdfile $outputname.dcd
xstFile $outputname.xst
dcdfreq 12500
xstFreq 12500
restartFreq 12500
outputEnergies 12500
outputpressure 12500
outputTiming 12500
############################################################################################
regards:Maryam

On Sat, Nov 29, 2014 at 8:34 PM, Abhishek Tyagi <atyagiaa_at_connect.ust.hk<mailto:atyagiaa_at_connect.ust.hk>> wrote:
I think you must reduce your tilmestep to 1 and for NPT ensemble, try this, and please post your config file

cheers
Abhi

On 30 Nov, 2014, at 12:19 am, maryam atabay <matabay2010_at_gmail.com<mailto:matabay2010_at_gmail.com>> wrote:

Dear NAMD users:
Hi
I want to simulate protein and nanosheet in water with using cuda-enabled in the NPT ensamble with fixed-coordinated nanosheet. The minimization and heating stages was executed without any error but in the md stage I encounter with following error:

FATAL ERROR: Periodic cell has become too small for original patch grid!
Possible solutions are to restart from a recent checkpoint,
increase margin, or disable useFlexibleCell for liquid simulation.
please help me. I will appreciate the oncoming helps.
regards:Maryam

--
Maryam Atabay
Ph.D in Nano Computational Chemistry,
Molecular Simulation Lab, Azarbaijan Shahid Madani University, Tabriz,Iran.
Email: Atabay_at_azaruniv.edu<mailto:Atabay_at_azaruniv.edu>
--
Maryam Atabay
Ph.D in Nano Computational Chemistry,
Molecular Simulation Lab, Azarbaijan Shahid Madani University, Tabriz,Iran.
Email: Atabay_at_azaruniv.edu<mailto:Atabay_at_azaruniv.edu>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:23:04 CST