[Molpro-user] Molpro executable hang on single node multi-proc run

[Peter Knowles]

This job runs just fine for me. It needs quite a bit of memory to do  
the integral transformation in FCI. Note that if you really want to do  
Li2 as a 2-electron problem without core-valence correlation, you can  
get exactly the same results by using mrci instead of fci!
Peter

On 15 Sep 2008, at 23:47, Jason Byrd wrote:

> Dear molpro users,
>
> I am running the evaluation version of molpro 2008.5 for testing  
> before
> my research group buys the full version.  Currently  I am doing a ECP
> FCI run with a reasonably large basis set for Li2, and the processes
> freezes after x (appears arbitrary) time.  What I mean by freeze is  
> that
> the process threads will stop counting and more often than not remain
> running but at 0% cpu and memory.  It is running on an emt64t intel 2
> proc quad core with 8 gb ram.  Are there any specific debugging or
> system configuration settings I should look into? or even any thing  
> else
> to fix this.
>
> Jason
>
>
> ***,FCI for li2; jason cvtz experimental
>
>  geometry={
>  ang;
>  2;
>  Li, 0.0,R,0.00;
>  Li, 0.0,-R,0.00;}
>
> basis={
> ecp,li,SBKJC;
> !bla
> s,Li,0.66010000E+04,0.98970000E+03,0.22570000E+03,0.64290000E 
> +02,0.21180000E+02,0.77240000E+01,0.30030000E+01,0.12120000E 
> +01,0.49300000E+00,0.95150000E-01,0.47910000E-01,0.22200000E-01
> c, 
> 1.12,0.11700000E 
> -03,0.91100000E 
> -03,0.47280000E-02,0.19197000E-01,0.63047000E-01,0.16320800E 
> +00,0.31482700E+00,0.39393600E+00,0.19691800E 
> +00,0.99970000E-02,-0.54020000E-02,0.17040000E-02
> c, 
> 1.12 
> ,-0.18000000E 
> -04 
> ,-0.14200000E 
> -03 
> ,-0.74100000E 
> -03 
> ,-0.30200000E 
> -02 
> ,-0.10123000E 
> -01,-0.27094000E-01,-0.57359000E-01,-0.93895000E-01,-0.12109100E 
> +00,0.27660800E+00,0.54954800E+00,0.27738500E+00
> c,10.10,0.10000000E+01
> c,11.11,0.10000000E+01
> c,12.12,0.10000000E+01
> p,Li,0.62500000E+01,0.13700000E+01,0.36720000E+00,0.11920000E 
> +00,0.44740000E-01,0.17950000E-01
> c,1.6,0.33880000E-02,0.19316000E-01,0.79104000E-01,0.27409500E 
> +00,0.51948800E+00,0.28442300E+00
> c,4.4,0.10000000E+01
> c,5.5,0.10000000E+01
> c,6.6,0.10000000E+01
> d,Li,0.34400000E+00,0.15300000E+00,0.68000000E-01
> c,1.1,0.10000000E+01
> c,2.2,0.10000000E+01
> c,3.3,0.10000000E+01
> f,Li,0.24600000E+00,0.12920000E+00
> c,1.1,0.10000000E+01
> c,2.2,0.10000000E+01
> g,Li,0.23800000E+00
> c,1.1,0.10000000E+01
> !core-corr functions
> s,li,10.5312774432
> s,li,4.0529638411
> s,li,1.3348062293
> p,li,4.9725550370
> p,li,16.6288182114
> p,li,1.6991632061
> d,li,10.5551766814
> d,li,3.0582755118
> f,li,6.6306837249
> spdf,li,even,nprim=1;
> }
>
> pesgrid 
> = 
> [2.6800,2.7000,2.7200,2.7400,2.7600,2.7800,2.8000,2.8200,2.8400,2.8600,2.8800,2.9000,2.9200,2.9400,2.9600,2.9800,3.0000,3.02,3.04,3.1 
> ]
>
> !R=100.0
> !rhf
> !{FCI;
> !wf,2,1,0}  ! if we are using the PP, should only care about 2  
> electrons
> !lili=energy ! asymptotic energy
> lili=-0.39269404
>
> DO I=1,#pesgrid
>  R=pesgrid(I)/2
>  geometry={
>  ang;
>  2;
>  Li, 0.0,R,0.00;
>  Li, 0.0,-R,0.00;}
>
>  rhf
>  {FCI;
>  wf,2,1,0}  ! if we are using the PP, should only care about 2  
> electrons
>  li2(I)=energy  ! equilibrium bond energy
>  final(I)=li2(I)-li_li  ! the D_e is the difference
> ENDDO !I
>  table,pesgrid,final
>
>
> -- 
> Jason Byrd
> Graduate Student
> Department of Physics
> University of Connecticut U-3046
> 2152 Hillside Road
> Storrs, CT 06269-3046
>
> Office: P209 Tel: 860-486-3484
> Email: byrd at phys.uconn.edu
>
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user

--
Prof. Peter J. Knowles
School of Chemistry, Cardiff University, Main Building, Park Place,  
Cardiff CF10 3AT, UK
Telephone: +44 29208 79182 Fax: +44 2920874030 Email KnowlesPJ at Cardiff.ac.uk
WWW http://www.cardiff.ac.uk/chemy/contactsandpeople/academicstaff/knowles.html