Re: Re: Charm++ fatal error with QM-MM MOPAC

From: Marcelo C. R. Melo (melomcr_at_gmail.com)
Date: Mon Jan 16 2017 - 15:52:41 CST

Hi,

The first point I want to make is that, when executing independent NAMD
QM/MM runs, make sure each one is using a different qmBaseDir, otherwise
different runs will read and write on the same folder, which could cause
all the issues you are experiencing right now. If you are not using
different folders in your simultaneous NAMD runs, try that now to see if it
solves your problem.

To answer your questions, we have not tried open shell systems with mopac.
NAMD reads the gradient output that is written in the *.aux file when MOPAC
is selected. The keywords "GRAD" and "1SCF" should be placed in the
qmConfigLine so that one SCF calculation is done and so that the gradients
are always printed. The keyword "QMMM" will tell MOPAC to read point
charges from the classical system (which are written by NAMD) and the
keyword "AUX" makes sure the file *.aux is used to print the gradient
information. Make sure these keywords are always used. As for MOZYME, this
is indicated for biomolecules, along with "XYZ". There are several ways to
fine-tune and speed up your calculations with MOPAC, but you should refer
to their documentation <http://openmopac.net/manual/allkeys.html>for that.

The error NAMD is returning indicates that it either cannot find the *.aux
file (ERROR: Could not find QM output file!) or that the calculation did
not end properly and gradients were not written ( Not all data could be
read!).

If the problem persists, could you provide the qmConfigLine you are using?
Also, the full output of your *.aux file?

Best,
Marcelo

---
Marcelo Cardoso dos Reis Melo
PhD Candidate
Luthey-Schulten Group
University of Illinois at Urbana-Champaign
crdsdsr2_at_illinois.edu
+1 (217) 244-5983
On 15 January 2017 at 01:55, Francesco Pietra <chiendarret_at_gmail.com> wrote:
> Hi Marcelo:
>
> As I wrote, MOPAC halted because MOZYME can't be used for open shell
> systems. Removing the key word MOZYME lets MOPAC2016 working correctly with
> open shell systems. Following the complex case of my first mail, I have now
> run a simpler system at either multiplicit 3 or 7. In both cases MOPAC
> ended without error messages. In the following read please abou the mult 7
> case from the /0 folder:
>
> From qmmm_0.input.ou
>>
>>  *                    *
>>  * JOB ENDED NORMALLY *
>>  *                    *
>>  **********************
>>
>>  TOTAL JOB TIME:            10.40 SECONDS
>>
>>  == MOPAC DONE ==
>>
>>
>
> From qmmm_0.input.aux:
> BETA_M.O.SYMMETRY_LABELS[0020]=
>  109A    110A    111A    112A    113A    114A    115A    116A    117A
> 118A
>  119A    120A    121A    122A    123A    124A    125A    126A    127A
> 128A
>  ALPHA_M.O.SYMMETRY_LABELS[0020]=
>  109A    110A    111A    112A    113A    114A    115A    116A    117A
> 118A
>  119A    120A    121A    122A    123A    124A    125A    126A    127A
> 128A
>  ALPHA_EIGENVALUES[0020]=
>     0.000    0.000    0.000    0.000    0.000    0.000    0.000
> 0.000    0.000    0.000
>     0.000    0.000    0.000    0.000    0.000    0.000    0.000
> 0.000    0.000    0.000
>  BETA_EIGENVALUES[0020]=
>       NaN      NaN      NaN      NaN      NaN      NaN      NaN
> NaN      NaN      NaN
>       NaN      NaN      NaN      NaN      NaN      NaN      NaN
> NaN      NaN      NaN
>  ALPHA_MOLECULAR_ORBITAL_OCCUPANCIES[00020]=
>  1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0
>  BETA_MOLECULAR_ORBITAL_OCCUPANCIES[00020]=
>  1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>  CPU_TIME:SECONDS[1]=       10.40
>  CPU_TIME:ARBITRARY_UNITS[1]=       14.79
>  END OF MOPAC FILE
>
>
>  However, NAMD12 procedure stopped with error message
>
> TCL: Minimizing for 100 steps
>> Info: List of ranks running QM simulations: 0.
>>           Sat Jan 14 23:09:08 2017  Job: '/dev/shm/NAMD_4IEV/0/qmmm_0.input'
>> started successfully
>>
>>
>>           MOPAC Job: "/dev/shm/NAMD_4IEV/0/qmmm_0.input" ended normally
>> on Jan 14, 2017, at 23:09.
>>
>> FATAL ERROR: Error reading QM forces file. Not all data could be read!
>> ------------- Processor 0 Exiting: Called CmiAbort ------------
>> Reason: FATAL ERROR: Error reading QM forces file. Not all data could be
>> read!
>>
>
> May I ask whether you already applied NAM12-MOPAC to open shell system
> (probably not because of MOZYME). At any event, could you suggest how to
> set the system to have QM-MM running? As I asked initially, what does
> charmm++ expects to read? Forces in which format? Or the above charmm++
> message is a fake error message and the problem lies elsewhere?
>
> Thanks
>
> francesco pietra
>
> On Tue, Jan 10, 2017 at 3:11 PM, Marcelo C. R. Melo <melomcr_at_gmail.com>
> wrote:
>
>> Hello!
>>
>> In NAMD, the keyword qmMult is only used when building the input for
>> ORCA. When using MOPAC, all options and keywords are to be passed using the
>> qmConfigLine keyword.
>>
>> Without more information, it seems that MOPAC is not accepting the
>> requested spin-state. You could find more info in the output file inside
>> the execution folder:  "/dev/shm/NAMD_4ztt_C/0/"
>> MOPAC will print its standard output and errors in the file:
>> qmmm_0.input.aux
>>
>> There is nothing in NAMD that would check the options you passed in
>> qmConfigLine and prevent you from using UHF or RHF, so there must be a
>> conflict somewhere between the keywords and your specific system.
>>
>> I hope this could be of help,
>> Marcelo
>>
>> ---
>> Marcelo Cardoso dos Reis Melo
>> PhD Candidate
>> Luthey-Schulten Group
>> University of Illinois at Urbana-Champaign
>> crdsdsr2_at_illinois.edu
>> +1 (217) 244-5983 <(217)%20244-5983>
>>
>> On 10 January 2017 at 01:56, Francesco Pietra <chiendarret_at_gmail.com>
>> wrote:
>>
>>> Hello:
>>>
>>> PROBLEM:
>>> I am trying to apply QM-MM with MOPAC2016 to a protein system with
>>> electron unpaired QM part. The procedure ends without errors by using the
>>> default configuration
>>>
>>> qmMult          "1 x"
>>>
>>>
>>> qmConfigLine    "PM7 XYZ T=2M 1SCF MOZYME CUTOFF=9.0 AUX LET GRAD QMMM
>>> GEO-OK"
>>>
>>> with "x" as the supposed total multiplicity (guessed in various modes,
>>> even by supposing that electrons from different unpaired species will get
>>> mixed). However any output from such a configugaration is highly suspicious
>>> to my eyes.
>>>
>>> I tried by inserting either "UHF" alone or "UHF MS=y" (with "y"
>>> representing the supposed multiplicity in MOPAC mode, commenting out, or
>>> not, the  "qmMult  "1 x"" line. UHF was made to replace 1SCF, or follow
>>> it .
>>>
>>> In all cases, I got the error
>>>
>>> MOPAC Job: "/dev/shm/NAMD_4ztt_C/0/qmmm_0.input" ended normally on Jan
>>>> 9, 2017, at 22:21.
>>>>
>>>> FATAL ERROR: Error reading QM forces file. Not all data could be read!
>>>> ------------- Processor 0 Exiting: Called CmiAbort ------------
>>>> Reason: FATAL ERROR: Error reading QM forces file. Not all data could
>>>> be read!
>>>>
>>>> Charm++ fatal error:
>>>> FATAL ERROR: Error reading QM forces file. Not all data could be read!
>>>>
>>>
>>> QUESTION:
>>> I would appreciate any indication where to look for  the QM forces file
>>> that CHARMM++ expects. Or any other suggestion that could allow running
>>> QM-MM/MOPAC under UHF.
>>>
>>> thanks
>>>
>>> francesco pietra
>>>
>>>
>>
>

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:59 CST