7
$\begingroup$

I am trying to perform a coupled cluster calculation at the CCSD(T) level using OpenMOLCAS. My input file is the following, in which I am using default values for most of the key inputs:

&GATEWAY
XBAS
C.STO-3G
H.STO-3G
O.STO-3G
End of basis
ZMAT
H1
C1 1 1.097
O1 2 1.343 1 111
H2 3 0.972 2 106.3 1 180
O2 2 1.202 3 124.9 4 0
End of z-matrix
&SEWARD
&SCF
 UHF
&RASSCF
  OUTOrbitals= Canonical
  LumOrb
  OUTOrbitals= Canonical
&MOTRA
  JobIph
&CCSDT
  CCT

Upon inspecting the output in the .log file, I am getting the total energy after 20 iterations:

      Total energy (diff) :    -186.39362934      -0.00000006
      Correlation energy  :      -0.1784309132990
      Reference energy    :    -186.2151984277011
      E1aa   contribution :       0.00000000
      E1bb   contribution :       0.00000000
      E2aaaa contribution :      -0.01326086
      E2bbbb contribution :      -0.01326086
      E2abab contribution :      -0.15190920

However, the code is not finishing the job correctly at the end of &CCSDT module. The following error message is printed to the log file:

Work space requirements :                17769
  Allocation of work space : Done
--- Stop Module: ccsdt at Sun Mar 10 21:07:35 2024 /rc=-6 ---
.########################.
.# Non-zero return code #.
.########################.

When examining the .err file, I am finding the following aborting signal:

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:
#0  0x7fc0cfa1f171 in ???
#1  0x7fc0cfa1e313 in ???
#2  0x7fc0cea4eb1f in ???
#3  0x7fc0cea4ea9f in ???
#4  0x7fc0cea21e04 in ???
#5  0x7fc0cea91036 in ???
#6  0x7fc0cea9819b in ???
#7  0x7fc0cea989db in ???
#8  0x7fc0cea9b55a in ???
#9  0x7fc0cea9c6cd in ???
#10  0x7fc0cfa1e608 in ???
#11  0x7fc0cfc1b047 in ???
#12  0x7fc0cfc2b03f in ???
#13  0x435762 in ???
#14  0x42146e in ???
#15  0x402460 in ???
#16  0x7fc0cea3acf2 in ???
#17  0x4024ad in ???
#18  0xffffffffffffffff in ???

What could be the reason for this error, and what should I do in order to fix it?

Improve this question
$\endgroup$
3
  • $\begingroup$ +1 but ideally you would provide the entire output file. If it doesn't fit within the maximum number of characters allowed in a post, then you can put it here in a folder called 12546, which is the number in this questions URL. $\endgroup$
    – Nike Dattani - No Free Time
    Commented Mar 10 at 17:13
  • $\begingroup$ I'm not sure if this makes a difference, but why did you have OUTOrbitals= Canonical twice? $\endgroup$
    – Nike Dattani - No Free Time
    Commented Mar 10 at 17:55
  • $\begingroup$ @Ihavenofreetimeanymore, I have added the files to the repo, you can check the pull request, then I will update the question here. About the OUTOrbitals= Canonical I had it from the example provided in the documentation (molcas.gitlab.io/OpenMolcas/sphinx/tutorials/…), I checked the case when it is mentioned only once and I found no difference in the output . $\endgroup$
    – Jaafar Mehrez
    Commented Mar 11 at 6:04

1 Answer 1

Reset to default
6
$\begingroup$

I have run a calculation with that input file, and can see that an error occurred when trying to calculate the (T) part of the CCSD(T) calculation, as you can see that we have ************************ instead of a real number:

       CCSD     =      -186.3936292872034
       T3 corr. =************************
       CCSD + T3=************************
  T3 energy decomposition into spin parts
  Eaaa =  -1320133193.83886
  Eaab =  -851388296422.193
  Eabb =   1505020301064.51
  Ebbb =   24408709436.1437


      Happy Landing!

--- Stop Module: ccsdt at Sun Mar 10 13:19:14 2024 /rc=-1 --

The next hint that I see, is that you have a "return code" of rc=-1. In my case, the .err file says nothing except for:

double free or corruption (!prev)

which is very different from your .err file, so we are probably using slightly different compilers, since the CCSD(T) code in OpenMolcas probably didn't change much between my version, which is the openmolcas/20.10 module on Compute Canada's machines, which have the following settings:

Host name: build-node.computecanada.ca (Linux)
C Compiler ID: Intel
C flags: -O2 -march=core-avx2 -axCore-AVX512 -ftz -fp-speculation=safe -fp-model source -std=gnu99 -qopenmp
Fortran Compiler ID: Intel
Fortran flags: -O2 -march=core-avx2 -axCore-AVX512 -ftz -fp-speculation=safe -fp-model source  -fpp -i8 -r8 -heap-arrays -qopenmp
Definitions: _MOLCAS_;_I8_;_LINUX_;_GA_;_MOLCAS_MPP_;SCALAPACK;_MKL_;_FDE_
Parallel: ON (GA=ON)

Ideally you would have included the entire output file, along with the entire input file and .err file, so that we could see more output data (such as the ************************ that I showed, and the version and compiler information. I have saved your input file as inp.inp, and all of the output that I included in this answer, can be found in this repository in the .log and .err files, which are complete.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .