Re: vmd-l: FFTK VMD plugin issue (arising from VECSCALE TCL issues of either NAMD or VMD ?)

From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Tue Jun 05 2012 - 09:22:10 CDT

Make sure you're using the latest FFTK in one of the VMD alpha builds. There should be no requirement for NAMD in the charge optimization step. I'm not sure how that would cause the problem you're having, anyway, but it's possible something else was also fixed between the version you're using and the current one.

On Jun 5, 2012, at 2:16 AM, Francesco Pietra wrote:

> Chris:
> That worked for me, too. However, I met problems either in having
> namd2 found, or in running it because of VECSCALE issues.
> I have restarted the thread in order to imply also namd and vmd in the
> thread title.
>
> I have three installations of namd on this multi-GTX-580 machine.
> While VMD callsl the GPUs, for namd I have tried both the cuda and the
> non-cuda version. With the cuda version 2.9, FFTK asks for "+idlepoll"
> on the command line, which is a correct request, however, without
> "+ifdlepoll" the executable namd2. was not found
>
> With the non-cuda version there is a problemwith vecscale, as you can
> see from the attached logs. I do not understand the origin of this TCL
> issue: is that a namd problem? Issue deriving from VMD calling the GPU
> cards? I don't have a machine with namd without GPU cards. The
> multi-CPU cluster has been dismissed to save on the electricity bill.
>
> At each trial I have to reset all, in order that the namd libraries
> are set bcorrectly.
>
> Very sorry for so much troubles from my side.
>
> LOGs from the non-cuda version of namd:
>
> ==========================================
> Charge Optimization GUI Debugging Output
> ==========================================
> INPUT SECTION
> psfPath: /home/francesco/..................../CLV.psf
> pdbPath: /home/francesco/........./CLV_optimized.pdb
> resName: CLV
> parList:
> /home/francesco/........./par_all27_prot_lipid.prm
> /home/francesco/....../CLV_gaffchanged.prm
> /home/francesco/......../CLV_fftk.prm
> namdbin: /usr/local/namd_2.9b3_multicore/bin/namd2
> log file: ChargeOpt.log
> -------------------------------------------
> CHARGE CONSTRAINTS SECTION
> chargeGroups: {O4 O5}
> chargeInit: -0.8023
> chargeBounds: {-1.0 1.0}
> chargeSum: -1.0030
> -------------------------------------------
> QM TARGET DATA SECTION
> baseLog: /home/francesco/...../single-point_energy.log
> watLog: /home/francesco/......./single-point_energy_water.log
> logFileList:
> /home/francesco/............./CLV-ACC-O1.gau.log
> /home/francesco/............/CLV-ACC-O1-m120.gau.log
> /home/francesco/........../CLV-ACC-O1-p120.gau.log
> /home/francesco/............/CLV-ACC-O4.gau.log
> /home/francesco/.............../CLV-ACC-O4-m120.gau.log
> /home/francesco/............../CLV-ACC-O4-p120.gau.log
> /home/francesco/................../CLV-ACC-O5.gau.log
> /home/francesco/................../CLV_ACC-O5-m120.gau.log
> /home/francesco/..................../CLV-ACC-O5-p120.gau.log
> /home/francesco/................./CLV-DON-H8.gau.log
> atomList: O1 O1 O1 O4 O4 O4 O5 O5 O5 H8
> indWeights: 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
> -------------------------------------------
> ADVANCED SETTINGS SECTION
> start: -0.4
> end: 0.4
> delta: 0.05
> end: 0.4
> offset: -0.2
> scale: 1.16
> tol: 0.005
> dWeight: 1.0
> Optimization mode: downhill
> Simulated Annealing Parameters: Temp. 25, Steps 20, Iterations 15
> Override ReChargeFromPSF: 0
> Override Charges:
> debug: 1
> ==========================================
>
> Single Point Energies Parsed. QME(water):
> -47665.049422800854 QME(cmpd): -461756.39140976494
> Cmpd atoms retyped to: KY KY Z K W K3 K2 K2 WS K W W K3 WJ JW JK JK J2
> J1 JA J1 J1
> Cmpd atoms recharged to: -0.22939999401569366 0.21549999713897705
> -0.3328000009059906 0.6535000205039978 -0.5724999904632568
> -0.08609999716281891 0.10090000182390213 -0.3012000024318695
> -0.3549000024795532 0.9506000280380249 -0.802299976348877
> -0.802299976348877 0.18960000574588776 -0.6248000264167786
> 0.40700000524520874 0.07620000094175339 0.07620000094175339
> 0.1137000024318695 0.08969999849796295 0.15800000727176666
> 0.03620000183582306 0.03620000183582306
> QM data loaded into VMD frames
> Frame QMEn QMDist
> 0 -9.145678592275653 2.969399929046631
> 1 -8.706491567064367 2.957278251647949
> 2 -10.400772799696131 3.028169870376587
> 3 -14.60439354257687 2.8175883293151855
> 4 -15.590309332930772 2.79134464263916
> 5 -18.000161877039645 2.819929838180542
> 6 -13.093001427206035 2.828505754470825
> 7 -2.3093492929896455 6.457061290740967
> 8 -18.266234190648245 2.815380573272705
> 9 -16.085382215360003 2.0513501167297363
> Beginning Optimization
> opt setup line: optimization -downhill -tol 0.005 -function
> ::ForceFieldToolKit::ChargeOpt::optCharges
> opt configure line: configure -bound
> opt initsimplex line: initsimplex
>
>
>
>
>
>
>
> RunOptimization.log:
>
> vecscale: parameters must have data
> vecscale: parameters must have data
> while executing
> "vecscale $psum [expr {1.0/double($ndim+1)}]"
> (procedure "downhill_simplex_optimization" line 39)
> invoked from within
> "downhill_simplex_optimization"
> (procedure "::Optimize::Opt0::handle" line 135)
> invoked from within
> "$opt start"
> (procedure "::ForceFieldToolKit::ChargeOpt::optimize" line 196)
> invoked from within
> "::ForceFieldToolKit::ChargeOpt::optimize"
> (procedure "::ForceFieldToolKit::gui::coptRunOpt" line 55)
> invoked from within
> "::ForceFieldToolKit::gui::coptRunOpt "
> invoked from within
> ".fftk_gui.hlf.nb.chargeopt.runOpt invoke "
> invoked from within
> ".fftk_gui.hlf.nb.chargeopt.runOpt instate {pressed !disabled} {
> .fftk_gui.hlf.nb.chargeopt.runOpt state !pressed;
> .fftk_gui.hlf.nb.chargeopt.runOpt in..."
> (command bound to event)
>
>
> Unfortunately, I dont know how to comply with the request "vescale:
> parameters must have data"
>
> Thanks
> francesco
>
>
> On Mon, Jun 4, 2012 at 8:45 PM, Mayne, Christopher G
> <cmayne2_at_illinois.edu> wrote:
>> Francesco,
>>
>> For the m120/p120 leave the filename as it, but when adding the log files to the QM Target Data box, manually remove the m120/p120 from the atom name. Duplicate atom names for target data is perfectly fine.
>>
>> Chris
>>
>>
>> On Jun 4, 2012, at 1:28 PM, Francesco Pietra wrote:
>>
>>> Chris:
>>> I am very sorry for having posted an old trial of mine. Actually, I
>>> had corrected the atom name (following your mail that you refer
>>> below), by simply deleting "gau". In my hands, extensions m120/p120
>>> were not accepted, so that I did not know how to make distinctions.
>>> The logs to which I intended to refer are seen below.
>>>
>>> Aside from that, I have doubly checked that extension ".gau" was
>>> removed throughout. Nonetheless, it still appear for H8.
>>>
>>> ==========================================
>>> Charge Optimization GUI Debugging Output
>>> ==========================================
>>> INPUT SECTION
>>> psfPath: /home/francesco/.........../CLV.psf
>>> pdbPath: /home/francesco/............./CLV_optimized.pdb
>>> resName: CLV
>>> parList:
>>> /home/francesco/............../par_all27_prot_lipid.prm
>>> /home/francesco/............./CLV_gaffchanged.prm
>>> /home/francesco/............/CLV_fftk.prm
>>> namdbin: namd2
>>> log file: ChargeOpt.log
>>> -------------------------------------------
>>> CHARGE CONSTRAINTS SECTION
>>> chargeGroups: {O4 O5}
>>> chargeInit: -0.8023
>>> chargeBounds: {-1.0 1.0}
>>> chargeSum: -1.0030
>>> -------------------------------------------
>>> QM TARGET DATA SECTION
>>> baseLog: /home/francesco/............../single-point_energy.log
>>> watLog: /home/francesco/............./single-point_energy_water.log
>>> logFileList:
>>> /home/francesco/............/CLV-ACC-O1.gau.log
>>> /home/francesco/.........../CLV-ACC-O1-m120.gau.log
>>> /home/francesco/.........../CLV-ACC-O1-p120.gau.log
>>> /home/francesco/.........../CLV-ACC-O4.gau.log
>>> /home/francesco/.........../CLV-ACC-O4-m120.gau.log
>>> /home/francesco/............../CLV-ACC-O4-p120.gau.log
>>> /home/francesco/................../CLV-ACC-O5.gau.log
>>> /home/francesco/............../CLV-ACC-O5-m120.gau.log
>>> /home/francesco/............/CLV-ACC-O4-p120.gau.log
>>> /home/francesco/ /CLV-DON-H8.gau.log
>>> atomList: O1 O1-m120 O1-p120 O4 O4-m120 O4-p120 O5 O5-m120 O4-p120 H8.gau
>>> indWeights: 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
>>> -------------------------------------------
>>> ADVANCED SETTINGS SECTION
>>> start: -0.4
>>> end: 0.4
>>> delta: 0.05
>>> end: 0.4
>>> offset: -0.2
>>> scale: 1.16
>>> tol: 0.005
>>> dWeight: 1.0
>>> Optimization mode: downhill
>>> Simulated Annealing Parameters: Temp. 25, Steps 20, Iterations 15
>>> Override ReChargeFromPSF: 0
>>> Override Charges:
>>> debug: 1
>>> ==========================================
>>>
>>> Single Point Energies Parsed. QME(water):
>>> -47665.049422800854 QME(cmpd): -461756.39140976494
>>> Cmpd atoms retyped to: KY KY Z K W K3 K2 K2 WS K W W K3 WJ JW JK JK J2
>>> J1 JA J1 J1
>>> Cmpd atoms recharged to: -0.22939999401569366 0.21549999713897705
>>> -0.3328000009059906 0.6535000205039978 -0.5724999904632568
>>> -0.08609999716281891 0.10090000182390213 -0.3012000024318695
>>> -0.3549000024795532 0.9506000280380249 -0.802299976348877
>>> -0.802299976348877 0.18960000574588776 -0.6248000264167786
>>> 0.40700000524520874 0.07620000094175339 0.07620000094175339
>>> 0.1137000024318695 0.08969999849796295 0.15800000727176666
>>> 0.03620000183582306 0.03620000183582306
>>>
>>>
>>>
>>>
>>> RunOptimozation.log:
>>> atomselect: cannot parse selection text: name O1-m120 and resname CLV
>>> atomselect: cannot parse selection text: name O1-m120 and resname CLV
>>> while executing
>>> "atomselect top "name [lindex $atomList $i] and resname $resName" frame $i"
>>> (procedure "::ForceFieldToolKit::ChargeOpt::optimize" line 152)
>>> invoked from within
>>> "::ForceFieldToolKit::ChargeOpt::optimize"
>>> (procedure "::ForceFieldToolKit::gui::coptRunOpt" line 55)
>>> invoked from within
>>> "::ForceFieldToolKit::gui::coptRunOpt "
>>> invoked from within
>>> ".fftk_gui.hlf.nb.chargeopt.runOpt invoke "
>>> invoked from within
>>> ".fftk_gui.hlf.nb.chargeopt.runOpt instate {pressed !disabled} {
>>> .fftk_gui.hlf.nb.chargeopt.runOpt state !pressed;
>>> .fftk_gui.hlf.nb.chargeopt.runOpt in..."
>>> (command bound to event)
>>>
>>>
>>> francesco
>>>
>>> *******************************************************
>>>
>>>
>>>
>>> On Mon, Jun 4, 2012 at 4:41 PM, Mayne, Christopher G
>>> <cmayne2_at_illinois.edu> wrote:
>>>> Francesco,
>>>>
>>>> As I have told you previously in direct correspondence, the problem is due to improper atom names as shown in the debugging file here:
>>>>
>>>>> QM TARGET DATA SECTION
>>>>> baseLog: /home/francesco/work_3LEZ_charmm27FF/mod2-CLV_FFTK_CHARMM/single-point_energy.log
>>>>> watLog: /home/francesco/work_3LEZ_charmm27FF/mod2-CLV_FFTK_CHARMM/single-point_energy_water.log
>>>>> logFileList:
>>>>> /home/francesco/....../CLV-ACC-O1.gau.log
>>>>> /home/francesco/...../CLV-ACC-O1-m120.gau.log
>>>>> /home/francesco/...../CLV-ACC-O1-p120.gau.log
>>>>> /home/francesco/..../CLV-ACC-O4.gau.log
>>>>> /home/francesco/...../CLV-ACC-O4-m120.gau.log
>>>>> /home/francesco/........./CLV-ACC-O4-p120.gau.log
>>>>> /home/francesco/......../CLV-ACC-O5.gau.log
>>>>> /home/francesco/......../CLV-ACC-O5-m120.gau.log
>>>>> /home/francesco/.........../CLV-ACC-O5-p120.gau.log
>>>>> /home/francesco/........./CLV-DON-H8.gau.log
>>>>> atomList: O1.gau O1-m120.gau O1-p120.gau O4.gau O4-m120.gau
>>>>> O4-p120.gau O5.gau O5-m120.gau O5-p120.gau H8.gau
>>>>> indWeights: 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
>>>>
>>>> Upon loading the log files into the QM target data section, ffTK tries to automatically fill out the "atom name" column by parsing the relevant information from the filename. It expects log filenames to follow the form: <resname>-<DON or ACC>-<atom name>.log. The .gau.log m120/p120 extensions included in your filenames fully compliant with what the parser expects. This yields incorrect settings for the atom names, as evidenced from the atom list and the resulting error message thrown by ffTK. You need to manually edit the "Atom Name" properly in the qm target data section to give the proper atom names (e.g., O1 rather than O1.gau).
>>>>
>>>>
>>>> Chris
>>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:37 CST