Molcas Forum

andrewshyichuk

Member

Registered: 2020-02-13

Posts: 86

Cannot change contraction in Stuttgart ECP

Dear Users,
I get a peculiar problem.
If I set a basis to default Stuttgart, e.g. Lu.Stuttgart.Dolg.7s6p5d.5s4p3d.ECP.11el - it works.
But if I change the contraction - I get an "invalid pointer" error, with e.g. Lu.Stuttgart.Dolg.7s6p5d.3s1p1d.ECP.11el

Happens also to Stuttgart_NL ecp.
Does not happen to CG-AIMP, did not happen before with regular bases, on the same build.

EDIT: The smallest one that works is 4s3p2d. Which is odd to me. The ECP is 11e, which should mean 5s2 5p6 6s2 5d1. I thus should be able to get away with 2s2p1d contraction for chemistry, or with 1s1p0d for idealized cation - those do not work. 

Input:

&GATEWAY
Title
    test Lu 0g

BSShow
ECPShow
Expert



Basis Set
*Lu.CG-AIMP.Seijo.14s10p9d8f3g.2s1p1d1f0g.ECP.25el.
*Lu.CG-AIMP.Seijo.14s10p9d8f3g.7s6p5d3f0g.ECP.25el.
*Lu.Stuttgart.Dolg.7s6p5d.3s1p1d.ECP.11el.
*Lu.Stuttgart_NL.Dolg.7s6p5d.5s4p3d.ECP.11el.
Lu.Stuttgart_NL.Dolg.7s6p5d.3s1p1d.ECP.11el.
Lu 0. 0. 0. Angstrom
End of Basis

RICD
End Of Input

&SEWARD
Title
    test Lu 0g

RELA
   R02O02

CHOLesky
THRCholesky
1d-3

Error:

free(): invalid pointer

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:
#0  0x7f2c8829e171 in ???
#1  0x7f2c8829d313 in ???
#2  0x7f2c865c387f in ???
#3  0x7f2c865c37ff in ???
#4  0x7f2c865adc34 in ???
#5  0x7f2c86606986 in ???
#6  0x7f2c8660dd8b in ???
#7  0x7f2c8660f68b in ???
#8  0x447fec in dmma_free_2d_
        at /opt/OpenMolcas-ff0b5d1c/src/Include/mma_allo_template.fh:348
#9  0x488958 in getbs_
        at /opt/OpenMolcas-ff0b5d1c/src/gateway_util/getbs.F90:407
#10  0x41b986 in processbasis
        at /opt/OpenMolcas-ff0b5d1c/src/gateway_util/rdctl_seward.F90:3804
#11  0x41b986 in rdctl_seward_
        at /opt/OpenMolcas-ff0b5d1c/src/gateway_util/rdctl_seward.F90:2968
#12  0x404af4 in gateway_
        at /opt/OpenMolcas-ff0b5d1c/src/gateway/gateway.F90:123
#13  0x4048a0 in MAIN__
        at /opt/OpenMolcas-ff0b5d1c/src/gateway/main.F90:28
#14  0x4048a0 in main
        at /opt/OpenMolcas-ff0b5d1c/src/gateway/main.F90:18

Last edited by andrewshyichuk (2025-01-23 15:15:59)

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Cannot change contraction in Stuttgart ECP

From what the code is doing, I think it's only supposed to work if you increase the number of basis functions. You'll then get something like this:

Example:  library:.6s.3s.  ->  input:.6s.5s.
          x x 0                x x 0 0 0
          x x 0                x x 0 0 0
          x x 0                x x 0 0 0
          x x 0                x x 1 0 0
          x x 0                x x 0 1 0
          0 0 1                0 0 0 0 1

i.e., it adds some decontracted functions for the higher exponents.

Reducing the number does not work, even if it looks like it does. For instance, for the 7s.5s  to 7s.4s reduction what you get is

          x 0 0 0 0            0 0 0 0
          x 0 0 0 0            0 0 0 0
          x 0 0 0 0            0 0 0 0
          0 1 0 0 0            0 0 0 0
          0 0 1 0 0            0 0 0 0
          0 0 0 1 0            0 0 0 0
          0 0 0 0 1            1 1 1 1

which is clearly not what you want (you lose the first function, and the ones you get are all identical).

andrewshyichuk

Member

Registered: 2020-02-13

Posts: 86

Re: Cannot change contraction in Stuttgart ECP

I am sorry, but I cannot agree. This simple test shows that the contraction orbitals are removed correctly, as requested:

>> foreach contraction in ( 6s4p3d1f, 6s4p3d, 6s4p2d, 6s3p3d, 6s2p3d)


&GATEWAY
Title 
    test H ANO contraction

BSShow 
ECPShow 
Expert 

Basis Set
H.ANO-rcc.Widmark.8s4p3d1f.$contraction.
Al 0. 0. 0. Angstrom
End of Basis

End Of Input

>> enddo

With vim and search for "  Number of primitives                 ", a quite fine animation is achieved, the outermost primitives disappear as instructed.
But, for whatever reason, not for Stuttgart sets.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Cannot change contraction in Stuttgart ECP

That's for ANO-type basis set, those allow specifying arbitratry (within limits) contractions. I meant for the particular case where the functions added/removed are the "uncontracted" functions typical for polarization/diffuse.

andrewshyichuk

Member

Registered: 2020-02-13

Posts: 86

Re: Cannot change contraction in Stuttgart ECP

I see. I find it strange that those are not treated on the same footing. I mean, as long as I'm already messing around with contractions - I'm on my own.
But, I guess I can always just create a separate basis set file with whatever contractions I want.