Author Topic: RIPER display negative root while Cholesky factorization failed  (Read 5358 times)

skundu07

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
RIPER display negative root while Cholesky factorization failed
« on: November 07, 2018, 06:31:14 PM »
Hi, I am doing Single Point Energy Calculation via RIPER. I have 48 metal atoms with an adsorbed moiety on top of that. I have auxiliary basis sets (auxbasis), basis and molecular orbitals (alpha, beta) defined and a coord file where first two layer of metal atoms are fixed. Following is my control file, where I have used RIPER with periodic lattice defined, matrix diagonalization on.

$title
$operating system unix
$symmetry c1
$user-defined bonds    file=coord
$coord    file=coord
$periodic 3
$lattice
  8.38324586     0.00000000     0.00000000
  0.00000000    11.17766115     0.00000000
  0.00000000     0.00000000    20.92784999
$kpoints
  nkpoints 3 3 1
$riper
  thrints    1.0d-12
  lenonly    off
  lchgprj    on
  sigma      0.20
  northol    5
  pqmatdiag  on
  pqsingtol  1.0d-8
$optimize
 internal   off
 redundant  off
 cartesian  on
 global     off
 basis      off
$atoms
c  1                                                                           \
   basis =c def2-TZVP                                                          \
   jbas  =c def2-TZVP
o  2                                                                           \
   basis =o def2-TZVP                                                          \
   jbas  =o def2-TZVP
pd 3-50                                                                        \
   basis =pd ecp-28-mwb-TZVP                                                   \
   ecp   =pd ecp-28-mwb                                                        \
   jbas  =pd ecp-28-mwb-TZVP
$basis    file=basis
$ecp    file=basis
$rundimensions
   dim(fock,dens)=1296396
   natoms=50
   nshell=550
   nbf(CAO)=1608
   dim(trafo[SAO<-->AO/CAO])=1916
   rhfshells=2
   nbf(AO)=1454
$uhfmo_alpha   file=alpha
$uhfmo_beta   file=beta
$uhf
$alpha shells
 a       1-450                                  ( 1 )
$beta shells
 a       1-428                                  ( 1 )
$scfiterlimit       6000
$thize     0.10000000E-04
$thime        5

$scfintunit
 unit=30       size=0        file=twoint
$scfdiis
$maxcor    500 MiB  per_core
$drvopt
   cartesian  on
   basis      off
   global     off
   hessian    on
   dipole     on
   nuclear polarizability
$interconversion  off
   qconv=1.d-7
   maxiter=25
$coordinateupdate
   dqmax=0.3
   interpolate  on
   statistics    5
$forceupdate
   ahlrichs numgeo=0  mingeo=3 maxgeo=4 modus=<g|dq> dynamic fail=0.3
   threig=0.005  reseig=0.005  thrbig=3.0  scale=1.00  damping=0.0
$forceinit on
   diag=default
$energy    file=energy
$grad    file=gradient
$forceapprox    file=forceapprox
$dft
   functional b-p
   gridsize   m4
$scfconv   7
$scfdamp   start=0.700  step=0.050  min=0.050
$scforbitalshift  closedshell=0.3   automatic! 0.3
$ricore      500
$rij
$jbas    file=auxbasis
$last step     define
$end


But in my riper.out output file it is saying riper ended abnormally

              +--------------------------------------------------+
              |      SCREENING OF BASIS FUNCTION PRODUCTS        |
              +--------------------------------------------------+
               
         Threshold for shell products neglect:          0.1E-11
         Tolerance for shell products extents:          0.1E-06
         Number of {mu,nu,L} shell products:            2817642
         Number of primitive basis function products:   3867397
       
               
              +--------------------------------------------------+
              |             INITIAL ORBITALS/BANDS               |
              +--------------------------------------------------+
               

 reading orbital data $uhfmo_alpha  from file alpha
 orbital characterization : expanded

 reading orbital data $uhfmo_beta  from file beta
 orbital characterization : expanded

========================
 internal module stack:
------------------------
    riper
========================

 Cholesky factorization failed in <ztritrn>... negative root
 riper ended abnormally


Could anyone suggest me how to fix that?

Thanks
Subrata

mas

  • Guest
Re: RIPER display negative root while Cholesky factorization failed
« Reply #1 on: November 08, 2018, 04:42:07 PM »
What Turbomole version do you use?

skundu07

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: RIPER display negative root while Cholesky factorization failed
« Reply #2 on: November 09, 2018, 09:13:48 PM »
Thanks Dr. Sierka for your reply. I am using turbomole 7.2.1 for my RIPER calculation.

Currently my jobs are running. I have made some modification is my control file $riper section-

$periodic 3
$lattice   angs
  8.38324586     0.00000000     0.00000000
  0.00000000    11.17766115     0.00000000
  0.00000000     0.00000000    20.92784999
$kpoints
  nkpoints 4 4 1
$ricore      1000
$rij
$riper
 lmaxmom    20
 epsbext 1.0d-9
 sigma 0.20


Thanks

mas

  • Guest
Re: RIPER display negative root while Cholesky factorization failed
« Reply #3 on: November 09, 2018, 09:42:26 PM »
The reason for your troubles are diffuse basis functions. Turbomole version 7.3 includes automatic elimination of basis functions with small exponents. Please use the version 7.3 if you have access to it.

skundu07

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: RIPER display negative root while Cholesky factorization failed
« Reply #4 on: November 13, 2018, 06:11:15 PM »
Thanks Dr. Sierka for your reply.

I am using a structure which has 48 metal atoms and an adsorbed moiety (CO) on top of that. Following is my control file $riper section command-line for Single Point Energy Calculation via Turbomole.7.2.1.

$periodic 3
$lattice   angs
  8.38324586     0.00000000     0.00000000
  0.00000000    11.17766115     0.00000000
  0.00000000     0.00000000    20.92784999
$kpoints
  nkpoints 4 4 1
$ricore      1000
$rij
$riper
 lmaxmom    20
 epsbext 1.0d-9
 sigma 0.20

I am also using :

$scfiterlimit       6000   
$scfconv   7

At first I gave run for 48 hours. But I noticed that only 46 scf iterations had completed within 48 hours and energy had not converged. I used default auxbasis, basis and mos file generated from "define" for my calculation and I did not perform any low spin calculation. Is there any way to make the calculation fast?


Thanks