Page 1 of 1

@ Reason: PPST: TOO MANY START VECTORS

Posted: 16 Apr 2020, 03:12
by andanian
Hi, Dear Dalton Community

I am trying to calculate the Spin-Spin property with the option .SOSOCC (without .SOSOCC, the calculate is done) for some molecule, for example, CH4 but when I used a set bases too big, I obtain the following error (attach .out).

** PPST ** NUMBER OF START VECTORS 405
EXCEED THE MAXIMUM ALLOWED NUMBER 400

--- SEVERE ERROR, PROGRAM WILL BE ABORTED ---
Date and time (Linux) : Wed Apr 15 22:47:46 2020
Host name : parca

@ MPI MASTER, node no.: 0
@ Reason: PPST: TOO MANY START VECTORS

In the following subroutine

rsp/rspc82.F:3211: SUBROUTINE PPST(IBTYP,XINDX,FCAC,H2AC,WRK,NSIM,BINMEM,LWRK)

it is defined the MAXIMUM ALLOWED of START VECTOR, which is of 400. Then, How can I make calculates of Spin-Spin with the option .SOSOCC and big set bases?

Thanks!

Re: @ Reason: PPST: TOO MANY START VECTORS

Posted: 16 Apr 2020, 16:31
by hjaaj
With .SOSOCC Dalton calculates all p-h excitation energies (#occupied * #virtual), which quickly becomes too big if you try it for bigger molecules. However, for CH4 you can edit .../DALTON/rsp/rspc82.F and change the 400 in the code line

Code: Select all

     PARAMETER (MAXST=400)
to something bigger than what you need.

-- Hans Jørgen.

Re: @ Reason: PPST: TOO MANY START VECTORS

Posted: 17 Apr 2020, 03:21
by andanian
Hi, Hans Jørgen

I tried you advice but it wasn't enough, also it is need to edit abacus/abamag.F at line 1548 and abacus/abalrs.F at line 192 (MXRM) with that my calculates now they are finish without error.

I have another problem, :roll:

I calculate .SOSOCC with occupied and virtual localized orbitals but the information about FC, PSO and SD(+FC) in "Contributions to the indirect spin-spin coupling constant !", They didn't print, but when I used only occupied localized orbitals, the information if it's print.

I Attach the followinf files

spin_localization_occandvir_CH4.out (occupied and virtual localized orbitals)
spin_localization_occ_CH4.out (occupied localized orbitals)

What happened?

Thanks!

Re: @ Reason: PPST: TOO MANY START VECTORS

Posted: 18 Apr 2020, 12:12
by sauer
Hola Andy,

el problema es, que no pediste el análisis con orbitales occupados y virtuales. En ambos inputs tiene el keyword .SOSOCC y en realidad con el comentario " !activa impresión de contribución solo occupados". Solamente pedir la transformación de los orbitales virtuales à la Foster y Boys con el keyword .FBOVIR no significa que obtienes también el análysis del acoplamiento en orbitales virtuales. Para eso necesitas cambiar también el keyword .SOSOCC en .SOS.

Si tienes mas preguntas, podrías preguntar también a Patricio, que conoce este parte del programa del fondo, por él ha escrito grand parte de esto y tiene su oficina en el otro lado del pasillo, no?

Salu2
Stephan

Re: @ Reason: PPST: TOO MANY START VECTORS

Posted: 20 Apr 2020, 17:35
by andanian
Hola, S. Sauer

Antes que nada, muchas gracias por la ayuda y la aclaración, con la cual solucione el inconveniente. Segundo, me soprende como se encuentra de bien informado.

Feliz día.

Re: @ Reason: PPST: TOO MANY START VECTORS

Posted: 20 Apr 2020, 18:35
by sauer
Hola Andy,

de nada. Y tal vez vos no sabés que estoy cada año 2 o 3 semanas en Corrientes trabajando con Patricio. Y espero que podría volver también este año en el diciembre.

Salu2 y tal vez nos vemos en diciembre
Stephan