[molpro-user] Can't optimize C4H10 in Molpro 2009.1

[Hans-Joachim Werner]

This works without problem using option
{optg,3n;method,diis}

ITER.   ENERGY(OLD)    ENERGY(NEW)      DE          GRADMAX      
GRADNORM    GRADRMS     STEPMAX     STEPLEN     STEPRMS
1  -158.34447468  -158.34520181    -0.00072714  0.00935751   
0.01546155  0.00042949  0.21935364  0.26534118  0.07091541
2  -158.34520181  -158.34538832    -0.00018651  0.00583359   
0.00977338  0.00027148  0.05926875  0.07500797  0.02004672
3  -158.34538832  -158.34546050    -0.00007218  0.00092486   
0.00156171  0.00004338  0.00881342  0.01370650  0.00366322
4  -158.34546050  -158.34546108    -0.00000058  0.00039942   
0.00052371  0.00001455  0.00238855  0.00403855  0.00107935
5  -158.34546108  -158.34546111    -0.00000003  0.00017323   
0.00019039  0.00000529  0.00065066  0.00086729  0.00023179

Optimization in z-matrix coordinates is often more difficult, but we  
will look at this case again to see if there is a bug.

Best regards
Joachim Werner


Am 07.05.2010 um 10:50 schrieb Konstantin Tokarev:

> Geometry of C4H10 doesn't converge with any optimization algorithm.  
> I've tried qsd (default), rf, diis with step interpolation (as  
> recommended for floppy molecules) - nothing helps. I've even used  
> pre-calculated hessian (see input below) - nothing helps. Gradient  
> doesn't fall below 1.d-2
>
> The same things occur with other molecules with many degrees of  
> freedom. Level of theory doesn't matter, problem is in optimizer
>
> basis, 6-31G**
>
> r2 =         1.09419 ang
> r3 =         1.09467 ang
> a3 =       108.09176 ang
> r4 =         1.09473 ang
> a4 =       108.09063 ang
> d4 =       117.02306 ang
> r5 =         1.52031 ang
> a5 =       110.25733 ang
> d5 =       238.50846 ang
> r6 =         1.09614 ang
> a6 =       109.16630 ang
> d6 =       301.55424 ang
> r7 =         1.09631 ang
> a7 =       109.16295 ang
> d7 =        58.44923 ang
> r8 =         1.52778 ang
> a8 =       111.49085 ang
> d8 =       180.02562 ang
> r9 =         1.52064 ang
> a9 =       111.48546 ang
> d9 =       179.97438 ang
> r10 =         1.09443 ang
> a10 =       110.98794 ang
> d10 =       299.73523 ang
> r11 =         1.09416 ang
> a11 =       110.24023 ang
> d11 =       179.97438 ang
> r12 =         1.09450 ang
> a12 =       110.98796 ang
> d12 =        60.25483 ang
> r13 =         1.09610 ang
> a13 =       109.86122 ang
> d13 =        58.84686 ang
> r14 =         1.09617 ang
> a14 =       109.85801 ang
> d14 =       301.13536 ang
> symmetry,nosym
> geometry={
> H
> C, 1, r2
> H, 2, r3, 1, a3
> H, 2, r4, 1, a4, 3, d4
> C, 2, r5, 1, a5, 3, d5
> H, 5, r6, 2, a6, 1, d6
> H, 5, r7, 2, a7, 1, d7
> C, 5, r8, 2, a8, 1, d8
> C, 8, r9, 5, a9, 2, d9
> H, 9, r10, 8, a10, 5, d10
> H, 9, r11, 8, a11, 5, d11
> H, 9, r12, 8, a12, 5, d12
> H, 8, r13, 5, a13, 2, d13
> H, 8, r14, 5, a14, 2, d14
> }
>
> {rks,b3lyp
> wf,34,1,0}
>
> {frequencies, forward
> thermo,print,thermo}
>
> {optg,readhess
> method,diis,5,step
> }
>
> {frequencies
> thermo,print,thermo}
>
>
> P.S. In GAMESS C4H10 converges with default settings
>
> -- 
> Regards,
> Konstantin
>
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user