[molpro-user] SOAOTRA error in UCCSD-F12

[Gerald Knizia]

Gregory Magoon wrote:
> I ran with: -n 2 -m 1250 M.
> Below is the input file...thanks:  [..]

Your geometry minimally breaks Cs symmetry, and it seems like different
parts of the program do not handle these symmetrization thresholds 100% 
consistently.

Your input works if you replace the geometry with

bohr
geometry={
    C       0.000000409   -0.000000001    2.447938006
    C       0.000000409   -0.000000001   -2.447938006
    C       2.735035908   -0.010444413    1.469006954
    C       2.735035908   -0.010444413   -1.469006954
    C      -1.376563408   -2.363388608    1.469006954
    C      -1.376563408   -2.363388608   -1.469006954
    C      -1.358473152    2.373833024    1.469006954
    C      -1.358473152    2.373833024   -1.469006954
    H       0.000000409    0.000000003    4.511556736
    H       0.000000409   -0.000000003   -4.511556736
    H       3.725485528   -1.667456202    2.197963140
    H       3.725485528   -1.667456202   -2.197963140
    H       3.738111878    1.638954604    2.197963140
    H       3.738111878    1.638954604   -2.197963140
    H      -3.306802528   -2.392636891    2.197963140
    H      -3.306802528   -2.392636891   -2.197963140
    H      -0.449678042   -4.056777040    2.197963140
    H      -0.449678042   -4.056777040   -2.197963140
    H      -0.418681760    4.060093087    2.197963140
    H      -0.418681760    4.060093087   -2.197963140
    H      -3.288432595    2.417822433    2.197963140
    H      -3.288432595    2.417822433   -2.197963140
}

instead of the z-matrix. That is the (symmetrized) geometry Molpro 
outputs at
the beginning of the output. If you want to run the (non-symmetric) original
z-matrix geometry, you can do so by putting gthresh,symtol=1e-9 at the
start of the output, which will suppress symmetrization in this case and do
the calculation without symmetry (it will take much longer then of course).
-- 
Gerald Knizia