Re: GBIS and ABF

From: Bjoern Olausson (namdlist_at_googlemail.com)
Date: Wed Mar 23 2011 - 09:32:43 CDT

Hej Branko,

thanks for your guess on my variable transfer question.

Regarding "appylBias":
"applyBias" defaults to "yes" so you can as well kick it if you set it to yes
;)

Acceptable Values: boolean
Default Value: yes
Description: If this is set to no, the calculation proceeds normally but the
adaptive biasing force is not applied. Data is still collected to compute the
free energy gradient. This is mostly intended for testing purposes, and should
not be used in routine simulations.
-- Quoted from the NAMD userguide --

I never tried GBIS without ABF, but what happens if you set "applyBias" to
"no", is the GBIS simulation it still slowed down?

I tried the latest snapshot CVS-2011-03-23 with the same setting and it works
flawless and I found no impact on speed whether "ABF"/"COLVARS" is on or off.
I guess you are aware that the number of Atoms in a colvar can have an impact
on the simulation speed
http://www.ks.uiuc.edu/Research/namd/2.7/ug/node49.html#SECTION000122310000000000000
(Thanks to Ajasja LjubetiÄ, who made me aware of this)

Here's a benchmark with the CVS-2011-03-23 version.
2440 Atoms, 12 Cores (2x Six-Core AMD Opteron 2427)
NAMD CVS-2011-03-23 for Linux-x86_64-ibverbs-net-linux-x86_64-ibverbs-icc

GBIS (colvars off)
Benchmark:
[s/step (CPU)] [s/step (WALL)] [days/ns] [MB memory]
XXXXXXXXX 0.0272864 0.157908 91.5742
Timing (after 14000 Steps):
0.0268379 0.0268443 XXXXXXXX 93.605469

GBIS+COLVARS (applyBias no)
Benchmark:
[s/step (CPU)] [s/step (WALL)] [days/ns] [MB memory]
XXXXXXXXX 0.0283311 0.163952 92.0156
Timing (after 14000 Steps):
0.0271339 0.0271504 XXXXXXXX 93.941406

GBIS+COLVARS+ABF
Benchmark:
[s/step (CPU)] [s/step (WALL)] [days/ns] [MB memory]
XXXXXXXXX 0.0278928 0.161417 91.6836
Timing (after 14000 Steps):
0.0273758 0.0273889 XXXXXXXX 93.843750

Example output for "GBIS+COLVARS (applyBias no)"

Charmrun> IBVERBS version of charmrun
Charm++> scheduler running in netpoll mode.
Charm++> cpu affinity enabled.
Charm++> Running on 1 unique compute nodes (12-way SMP).
Charm++> cpu topology info is gathered in 0.006 seconds.
Info: NAMD CVS-2011-03-23 for Linux-x86_64-ibverbs-net-linux-x86_64-ibverbs-
icc
Info:
Info: Please visit http://www.ks.uiuc.edu/Research/namd/
Info: for updates, documentation, and support information.
Info:
Info: Please cite Phillips et al., J. Comp. Chem. 26:1781-1802 (2005)
Info: in all publications reporting results obtained with NAMD.
Info:
Info: Based on Charm++/Converse 60303 for net-linux-x86_64-ibverbs-icc
Info: Built Wed Mar 23 13:18:48 CET 2011 by blub on mrbird
Info: 1 NAMD CVS-2011-03-23 Linux-x86_64-ibverbs-net-linux-x86_64-ibverbs-
icc 12 node012 blub
Info: Running on 12 processors, 12 nodes, 1 physical nodes.
Info: CPU topology information available.
Info: Charm++/Converse parallel runtime startup completed at 0.0626321 s
Info: 84.9297 MB of memory in use based on /proc/self/stat
Info: Configuration file is gbis-abf.conf
Info: Working in the current directory /home/blub/work/prot/wildtype/namd-
gbis/test
TCL: Suspending until startup complete.
Warning: The following variables were set in the
Warning: configuration file but will be ignored:
Warning: LangevinPistonTarget (LangevinPiston)
Warning: LangevinPistonPeriod (LangevinPiston)
Warning: LangevinPistonDecay (LangevinPiston)
Warning: LangevinPistonTemp (LangevinPiston)
Warning: consref (constraints)
Warning: conskfile (constraints)
Warning: conskcol (constraints)
Warning: constraintScaling (constraints)
Warning: IMDport (IMDon)
Warning: IMDfreq (IMDon)
Warning: IMDwait (IMDon)
Warning: IMDignore (IMDon)
Info: EXTENDED SYSTEM FILE run000/prot_wt.restart.xsc
Info: SIMULATION PARAMETERS:
nfo: TIMESTEP 2
Info: NUMBER OF STEPS 500000
Info: STEPS PER CYCLE 10
Info: PERIODIC CELL BASIS 1 95 0 0
Info: PERIODIC CELL BASIS 2 0 95 0
Info: PERIODIC CELL BASIS 3 0 0 95
Info: PERIODIC CELL CENTER 0 0 0
Info: WRAPPING ALL CLUSTERS AROUND PERIODIC BOUNDARIES ON OUTPUT.
Info: LOAD BALANCER Centralized
Info: LOAD BALANCING STRATEGY New Load Balancers -- DEFAULT
Info: LDB PERIOD 2000 steps
Info: FIRST LDB TIMESTEP 50
Info: LAST LDB TIMESTEP -1
Info: LDB BACKGROUND SCALING 1
Info: HOM BACKGROUND SCALING 1
Info: MIN ATOMS PER PATCH 40
Info: VELOCITY FILE run000/prot_wt.restart.vel
Info: CENTER OF MASS MOVING INITIALLY? NO
Info: DIELECTRIC 1
Info: EXCLUDE SCALED ONE-FOUR
Info: 1-4 ELECTROSTATICS SCALED BY 1
Info: MODIFIED 1-4 VDW PARAMETERS WILL BE USED
Info: DCD FILENAME prot_wt.dcd
Info: DCD FREQUENCY 500
Info: DCD FIRST STEP 500500
Info: DCD FILE WILL CONTAIN UNIT CELL DATA
Info: XST FILENAME prot_wt.xst
Info: XST FREQUENCY 500
Info: NO VELOCITY DCD OUTPUT
Info: NO FORCE DCD OUTPUT
Info: OUTPUT FILENAME prot_wt
Info: BINARY OUTPUT FILES WILL BE USED
Info: RESTART FILENAME prot_wt.restart
Info: RESTART FREQUENCY 500
Info: BINARY RESTART FILES WILL BE USED
Info: SWITCHING ACTIVE
Info: SWITCHING ON 15
Info: SWITCHING OFF 16
Info: PAIRLIST DISTANCE 17
Info: PAIRLIST SHRINK RATE 0.01
Info: PAIRLIST GROW RATE 0.01
Info: PAIRLIST TRIGGER 0.3
Info: PAIRLISTS PER CYCLE 2
Info: PAIRLISTS ENABLED
Info: MARGIN 0
Info: HYDROGEN GROUP CUTOFF 2.5
Info: PATCH DIMENSION 19.5
Info: ENERGY OUTPUT STEPS 100
Info: CROSSTERM ENERGY INCLUDED IN DIHEDRAL
Info: TIMING OUTPUT STEPS 1000
Info: PRESSURE OUTPUT STEPS 100
Info: GBIS GENERALIZED BORN IMPLICIT SOLVENT ACTIVE
Info: GBIS SOLVENT DIELECTRIC: 80
Info: GBIS PROTEIN DIELECTRIC: 1
Info: GBIS COULOMB RADIUS OFFSET: 0.09 Ang
Info: GBIS ION CONCENTRATION: 0.3 M
Info: GBIS DEBYE SCREENING LENGTH: 5.62399 Ang
Info: GBIS DELTA: 1
Info: GBIS BETA: 0.8
Info: GBIS GAMMA: 4.85
Info: GBIS BORN RADIUS CUTOFF: 14 Ang
Info: GBIS MAX BORN RADIUS: 30 Ang
Info: COLLECTIVE VARIABLES CALCULATION REQUESTED
Info: COLLECTIVE VARIABLES CONFIGURATION colvars.in
Info: LANGEVIN DYNAMICS ACTIVE
Info: LANGEVIN TEMPERATURE 303
Info: LANGEVIN DAMPING COEFFICIENT IS 5 INVERSE PS
Info: LANGEVIN DYNAMICS NOT APPLIED TO HYDROGENS
Info: USING VERLET I (r-RESPA) MTS SCHEME.
Info: C1 SPLITTING OF LONG RANGE ELECTROSTATICS
Info: PLACING ATOMS IN PATCHES BY HYDROGEN GROUPS
Info: RIGID BONDS TO HYDROGEN : ALL
Info: ERROR TOLERANCE : 1e-08
Info: MAX ITERATIONS : 100
Info: RIGID WATER USING SETTLE ALGORITHM
Info: RANDOM NUMBER SEED 1300889106
Info: USE HYDROGEN BONDS? NO
Info: COORDINATE PDB /home/blub/work/prot/wildtype/charmm-
gui/step1_pdbreader.pdb
Info: STRUCTURE FILE /home/blub/work/prot/wildtype/charmm-
gui/step1_pdbreader.xplor.psf
Info: PARAMETER file: CHARMM format!
Info: PARAMETERS
/cvos/shared/apps/charmm/c34b2/toppar/par_all27_prot_lipid.prm
Info: USING ARITHMETIC MEAN TO COMBINE L-J SIGMA PARAMETERS
Info: BINARY COORDINATES run000/prot_wt.restart.coor
Info: FIRST TIMESTEP 500000
Info: SUMMARY OF PARAMETERS:
Info: 180 BONDS
Info: 447 ANGLES
Info: 566 DIHEDRAL
Info: 46 IMPROPER
Info: 6 CROSSTERM
Info: 119 VDW
Info: 0 VDW_PAIRS
Info: TIME FOR READING PSF FILE: 0.0781538
Info: TIME FOR READING PDB FILE: 0.00650382
Info:
Info: Reading from binary file run000/prot_wt.restart.coor
Info: ****************************
Info: STRUCTURE SUMMARY:
Info: 2440 ATOMS
Info: 2453 BONDS
Info: 4481 ANGLES
Info: 6546 DIHEDRALS
Info: 404 IMPROPERS
Info: 148 CROSSTERMS
Info: 0 EXCLUSIONS
Info: 1248 RIGID BONDS
Info: 6072 DEGREES OF FREEDOM
Info: 1192 HYDROGEN GROUPS
Info: 4 ATOMS IN LARGEST HYDROGEN GROUP
Info: 1192 MIGRATION GROUPS
Info: 4 ATOMS IN LARGEST MIGRATION GROUP
Info: TOTAL MASS = 16953.6 amu
Info: TOTAL CHARGE = -0.999998 e
Info: MASS DENSITY = 0.032836 g/cm^3
Info: ATOM DENSITY = 0.0028459 atoms/A^3
Info: *****************************
Info:
Info: Entering startup at 0.16531 s, 86.2148 MB of memory in use
Info: Startup phase 0 took 0.000216007 s, 86.2148 MB of memory in use
Info: Startup phase 1 took 0.013021 s, 87.8672 MB of memory in use
Info: Startup phase 2 took 0.000271082 s, 87.8672 MB of memory in use
Info: Startup phase 3 took 0.000101089 s, 87.8672 MB of memory in use
Info: PATCH GRID IS 3 (PERIODIC) BY 4 (PERIODIC) BY 4 (PERIODIC)
Info: PATCH GRID IS 1-AWAY BY 1-AWAY BY 1-AWAY
Info: Reading from binary file run000/prot_wt.restart.vel
Info: REMOVING COM VELOCITY 0.0108206 -0.18671 -0.131359
Info: LARGEST PATCH (27) HAS 527 ATOMS
Info: Startup phase 4 took 0.00324392 s, 88.8164 MB of memory in use
Info: Startup phase 5 took 0.000789881 s, 88.9922 MB of memory in use
Info: Startup phase 6 took 0.000650167 s, 88.9922 MB of memory in use
LDB: Central LB being created...
Info: Startup phase 7 took 0.400699 s, 88.9922 MB of memory in use
Info: CREATING 984 COMPUTE OBJECTS
Info: useSync: 1 useProxySync: 0
Info: NONBONDED TABLE R-SQUARED SPACING: 0.0625
Info: NONBONDED TABLE SIZE: 833 POINTS
Info: ABSOLUTE IMPRECISION IN FAST TABLE ENERGY: 2.71051e-20 AT 16.4905
Info: RELATIVE IMPRECISION IN FAST TABLE ENERGY: 1.15453e-16 AT 15.998
Info: ABSOLUTE IMPRECISION IN FAST TABLE FORCE: 8.47033e-22 AT 15.998
Info: RELATIVE IMPRECISION IN FAST TABLE FORCE: 1.14081e-16 AT 15.998
Info: ABSOLUTE IMPRECISION IN VDWA TABLE ENERGY: 1.54074e-33 AT 21.9986
Info: RELATIVE IMPRECISION IN VDWA TABLE ENERGY: 2.06776e-20 AT 21.9986
Info: ABSOLUTE IMPRECISION IN VDWA TABLE FORCE: 1.54074e-33 AT 21.9986
Info: RELATIVE IMPRECISION IN VDWA TABLE FORCE: 2.51633e-16 AT 21.9986
colvars:
----------------------------------------------------------------------
colvars: Initializing the collective variables module, version 2011-02-03.
colvars: # analysis = off [default]
colvars: # colvarsTrajFrequency = 500
colvars: # colvarsRestartFrequency = 500
colvars: # colvarsTrajAppend = off [default]
colvars: The restart output state file will be
"prot_wt.restart.colvars.state".
colvars: The final output state file will be "prot_wt.colvars.state".
colvars: The trajectory file will be "prot_wt.colvars.traj".
colvars:
----------------------------------------------------------------------
colvars: Initializing a new collective variable.
colvars: # name = eprof
colvars: Initializing a new "dihedral" component.
colvars: # componentCoeff = 1 [default]
colvars: # componentExp = 1 [default]
colvars: # period = 0 [default]
colvars: # wrapAround = 0 [default]
colvars: # oneSiteSystemForce = off [default]
colvars: Initializing atom group "group1".
colvars: Atom group "group1" defined, 146 initialized: total mass =
928.381.
colvars: Initializing atom group "group2".
colvars: Atom group "group2" defined, 6 initialized: total mass =
39.057.
colvars: Initializing atom group "group3".
colvars: Atom group "group3" defined, 8 initialized: total mass =
41.073.
colvars: Initializing atom group "group4".
colvars: Atom group "group4" defined, 148 initialized: total mass =
941.4.
colvars: All components initialized.
colvars: # width = 1
colvars: # lowerBoundary = -180
colvars: # lowerWallConstant = 0.1
colvars: # lowerWall = -180 [default]
colvars: # upperBoundary = 0
colvars: # upperWallConstant = 0.1
colvars: # upperWall = 0 [default]
colvars: # expandBoundaries = off [default]
colvars: # extendedLagrangian = off [default]
colvars: # outputEnergy = on [default]
colvars: # outputValue = on [default]
colvars: # outputVelocity = off [default]
colvars: # outputSystemForce = on
colvars: # outputAppliedForce = on
colvars:
----------------------------------------------------------------------
colvars: Collective variables initialized, 1 in total.
colvars:
----------------------------------------------------------------------
colvars: Initializing a new "abf" instance.
colvars: # name = "abf1" [default]
colvars: # colvars = { eprof }
colvars: # applyBias = off
colvars: WARNING: ABF biases will *not* be applied!
colvars: # updateBias = on [default]
colvars: # hideJacobian = off
colvars: Jacobian (geometric) forces will be included in reported free
energy gradients.
colvars: # fullSamples = 100
colvars: # inputPrefix = [default]
colvars: # outputFreq = 500 [default]
colvars: # historyFreq = 0 [default]
colvars: Finished ABF setup.
colvars:
----------------------------------------------------------------------
colvars: Collective variables biases initialized, 1 in total.
colvars:
----------------------------------------------------------------------
colvars: Collective variables module initialized.
colvars:
----------------------------------------------------------------------
Info: Startup phase 8 took 0.011349 s, 90.0547 MB of memory in use
Info: Startup phase 9 took 0.000121117 s, 90.3047 MB of memory in use
Info: Finished startup at 0.595772 s, 90.3047 MB of memory in use

On Wednesday 23 March 2011 13:50:08 Branko wrote:
> Bjoern,
>
> Thank you very much, I first try to download last CVS. Regarding
> variables, never try, but think that it is impossible to "transfer"
> definition from main configuration file to colvar file (from rel. 2.7
> namd first read whole configuration file than recognize ABF conf. and
> "add" this to simulation configuration - this can be seen in every out
> file), most probably Axel or Jerome can give most appropriate answer. I
> see that you did't specify "applybias" if not specified probably is
> "considered as default" so I try also in this way, also in new release
> new keyword "updatebias" exist. I have experimental data in pure EtOH,
> so consider that ion conc. should not be different from 0. Did you note
> that when applying ABF simulation are significantly slower comparing
> same simulation without abf.
>
> Sincerely
>
> Branko
>
> On 3/23/2011 12:56 PM, Bjoern Olausson wrote:
> > My "colvarsRestartFrequency" and "colvarsTrajFrequency" are both set to
> > 500 which is identical to my "restartFreq", "dcdfreq" and "xstFreq"
> > setting.
> >
> > The main difference is your "ionConcentration" (0 vs. 0.3) and colvars
> > "dihedral" vs. "distance"
> >
> > My system is a Protein with 2440 Atoms (149 Residues).
> >
> >
> > You might want to rule out those to differences. If it still crashes for
> > you it might be a bug in you cvs snapshot?
> >
> > I'll run my system with a snapshot I'll download NOW ;-)
> >
> >
> > By the way:
> > Is it possible to pass arguments (variables) from the namd config to the
> > colvars config. Lets say I define "set A 200" in the main namd config
> > file can I somehow use $A in my colvars config file? Simply using $A
> > does not work :-(
> >
> >
> > Cheers,
> > Bjoern
> >
> > On Wednesday 23 March 2011 11:01:14 you wrote:
> >> Hi Bjoern,
> >>
> >> Thank you very much for respond. Yes I also use Langevin temperature,
> >> and input for both GBIS and ABF are similar to your, still something
> >> wrong happened. Please let me know about size of your system, also are
> >> colvarsRestartFrequency same as MD restart. Below are my inputs
> >>
> >> Kind regards
> >>
> >> Branko
> >>
> >> ____
> >>
> >> small molecule in implicit EtOH
> >>
> >> #GBIS parameters
> >> GBIS on
> >> solventDielectric 24.3
> >> ionConcentration 0
> >> intrinsicRadiusOffset 0.09
> >> GBISDelta 1.0
> >> GBISBeta 0.8
> >> GBISGamma 4.85
> >> alphaCutoff 14
> >>
> >> # Other parameters
> >> colvars on
> >> colvarsConfig Hmd
> >>
> >> ___
> >> ABF
> >>
> >>
> >> colvarsTrajFrequency 1
> >> colvarsRestartFrequency 10000
> >> analysis on
> >>
> >>
> >> colvar {
> >>
> >> name distance
> >> lowerBoundary 3.40
> >> upperBoundary 12.20
> >> width 0.01
> >>
> >> distance {
> >>
> >> group1 {
> >>
> >> atomnumbers 1 16 17
> >>
> >> }
> >> group2 {
> >>
> >> atomnumbers 36
> >>
> >> }
> >>
> >> }
> >>
> >> }
> >>
> >> colvar {
> >>
> >> name distance2
> >> lowerBoundary 4.80
> >> upperBoundary 12.20
> >> width 0.01
> >>
> >> distance {
> >>
> >> group1 {
> >>
> >> atomnumbers 18 19 20 21
> >>
> >> }
> >> group2 {
> >>
> >> atomnumbers 22 23 24 25
> >>
> >> }
> >>
> >> }
> >>
> >> }
> >>
> >>
> >> abf {
> >>
> >> colvars distance distance2
> >> fullSamples 500
> >> hideJacobian no
> >> outputFreq 1000
> >> applyBias yes
> >>
> >> }
> >>
> >> On 3/23/2011 12:13 AM, Bjoern Olausson wrote:
> >>> On Tuesday 22 March 2011 23:16:54 Branko wrote:
> >>>> Giacomo,
> >>>>
> >>>> Sorry because a question on CVS version, but I try to combine GBIS and
> >>>> ABF (linux CVS-2011-03-20 for Linux-x86_64), and likely that bug
> >>>> exist. Simulation works when GBIS without ABF are applied, also when
> >>>> GBIS is excluded but ABF applied, simulation works on 2.7 (win), but
> >>>> no on CVS. In GBIS and ABF combination, or in "ABF alone" on CVS
> >>>> whole calculation were initialized properly, but after that crash,
> >>>> without any "error message" in out file
> >>>> Please let me know are you try with GBIS and ABF combination in some
> >>>> of CVS. I use two colvars, one of which is distance between centroid
> >>>> defined on aromatic ring, and some other 'one atom'; second colvar is
> >>>> also distance defined on 3 atoms of each of two di-phenyl-methyl
> >>>> moieties that molecule comprise (also I try there to overcome any
> >>>> symmetry, when centroids are defined,even there is no constraints).
> >>>> BC was not applied.
> >>>
> >>> Hi Branko,
> >>>
> >>> I am using NAMD CVS-2011-03-03 with GBIS, one dihedral COLVAR and ABF
> >>> so far it works fine. Furthermore I use Langevin temperature control
> >>> and PBC.
> >>>
> >>>
> >>> GBIS config:
> >>>
> >>> GBIS on
> >>> solventDielectric 80
> >>> intrinsicRadiusOffset 0.09
> >>> ionConcentration 0.3
> >>> GBISDelta 1.0
> >>> GBISBeta 0.8
> >>> GBISGamma 4.85
> >>> alphaCutoff 14 cutoff (def:15).
> >>>
> >>>
> >>> This is my colvars file:
> >>>
> >>> colvarsTrajFrequency 500
> >>> colvarsRestartFrequency 500
> >>> colvar {
> >>>
> >>> name eprof conversely,
> >>>
> >>> lowerBoundary -180
> >>> upperBoundary 180
> >>>
> >>> lowerWallConstant 0.1
> >>> upperWallConstant 0.1
> >>> outputSystemForce no
> >>> outputAppliedForce no
> >>>
> >>> dihedral {
> >>>
> >>> group1 {
> >>>
> >>> # Use only Calpha
> >>> atomnumbers {4 21 38 57 77 93 115 129 148 162 169 191 205 224
> >>> 238 257
> >>>
> >>> 272 288 305 317 328 340 354 373 388 402 418 440 450 472 491 508 520 542
> >>> 557 564 585 599 611 623 640 657 681 700 719 739 749 756 778 795 814 829
> >>> 841 848 872 886 905 916 928 949 963 982 999 1021 1036 1047 1061 1080
> >>> 1097 1116 1132 1151 1175}
> >>>
> >>> }
> >>> group2 {
> >>>
> >>> # Use only Calpha
> >>> atomnumbers {1194 1218 1225}
> >>>
> >>> }
> >>> group3 {
> >>>
> >>> # Use only Calpha
> >>> atomnumbers {1244 1251}
> >>>
> >>> }
> >>> group4 {
> >>>
> >>> # Use only Calpha
> >>> atomnumbers {1268 1285 1304 1324 1340 1362 1376 1395 1409 1416
> >>> 1438 1452
> >>>
> >>> 1471 1485 1504 1519 1535 1552 1564 1575 1587 1601 1620 1635 1649 1665
> >>> 1687 1697 1719 1738 1755 1767 1789 1804 1811 1832 1846 1858 1870 1887
> >>> 1904 1928 1947 1966 1986 1996 2003 2025 2042 2061 2076 2088 2095 2119
> >>> 2133 2152 2163 2175 2196 2210 2229 2246 2268 2283 2294 2308 2327 2344
> >>> 2363 2379 2398 2422}
> >>>
> >>> }
> >>>
> >>> }
> >>>
> >>> }
> >>>
> >>> abf {
> >>>
> >>> colvars eprof
> >>> fullSamples 100
> >>> hideJacobian no
> >>>
> >>> }
> >>>
> >>> Kind regards,
> >>> Bjoern

-- 
Bjoern Olausson
Martin-Luther-Universität Halle-Wittenberg 
Fachbereich Biochemie/Biotechnologie
Kurt-Mothes-Str. 3
06120 Halle/Saale
Phone: +49-345-55-24942

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:48 CST