[SOLVED] ANO-RCC-VTZP and memory issues

Max.M.K. · 2022-10-08 18:01:09

Hi all, it is my first time posting here so I apologise for any issues with my post, please bear with me and let me know what else I can provide to help you with helping me.

I'm running CASSCF calculations and I'm using SEWARD to compute the integrals. The problem is, is that I can successfully run the computations with minimal basis sets (ANO-RCC-MB) and even larger basis sets (ANO-RCC-VDZP), but as soon as I use VTZP basis sets, (even if for only one atom) SEWARD crashes after 2 minutes and returns this error:

 ?mma_allo_?D: error: out-of-memory
 label: ALPHA
  available (kB):            0
  required  (kB):             0

It also specifies that I have 0kb required and 0kb available, which makes me think that Molcas is somehow not understanding itself, even though it would logically follow that this is due to a lack of memory, right?

No matter how much RAM i supply the calculation with >> export MOLCAS_MEM (even 45GB was not enough) I still return this error whenever I use the VTZP basis sets. I have adjusted the Cholesky decomposition parameter (1.0d-7 to 1.0d-5) but even then the only thing that determines whether SEWARD will crash after approx 2 minutes is whether or not I used VTZP.

I will include my input file for SEWARD, but I will remove the xfield coordinates, since there are five thousand lines, and it is not necessary to see them all.

>>export MOLCAS_MEM = 45000

&SEWARD &END
Title
ANGM
7.2141422574   4.1650869743   6.5879365326
AMFI
EXPERT

Basis Set
Sm.ANO-RCC-VTZP
 Sm1	3.81755995368244	2.2040692668959307	3.486186130885215 Angstrom
End of Basis Set

Basis Set
I.ANO-RCC-MB
 I11	1.2711202113038766	2.20649374529601	1.7051770232609482 Angstrom
 I12	2.5422404226077537	4.4081385337918615	5.251890032327323 Angstrom
 I13	5.088680245155076	-0.002424617256442829	1.7051770232609482 Angstrom
 I14	5.092879572561006	4.408138275915757	1.7051770232609482 Angstrom
 I15	2.546439669844924	-0.002424617256442829	5.251890032327323 Angstrom
 I16	6.3639997113312425	2.20649374529601	5.251890032327323 Angstrom
End of Basis Set

Basis Set
Sm.ECP.Seijo.0s.0s.0e-AIMP-Sm3Ga5O12-Sm
 Sm2	0.          	    0.	                    3.4708806325062493 Angstrom
 Sm3	7.635119915	        0.	                    3.4708806325062493 Angstrom
 Sm4	3.8175599575	    6.6122078073            3.4708806325062493 Angstrom
 Sm5	-3.817560045860091e-9	-4.408138540404069	3.486186130885215 Angstrom
 Sm6	-3.81755996131756	2.2040692668959307	    3.486186130885215 Angstrom
 Sm7	7.635119911182439	-4.408138540404069	    3.486186130885215 Angstrom
 Sm8	-3.817560045860091e-9	8.81627707419593	3.486186130885215 Angstrom
 Sm9	11.45267986868244	2.2040692668959307	    3.486186130885215 Angstrom
 Sm10	7.635119911182439	8.81627707419593	    3.486186130885215 Angstrom
End of Basis Set

Basis Set
Br.ECP.Abdalla.0s.0s.0e-AIMP-Cs2NaYBr6
I17	-6.363999703696123	2.20649374529601	1.7051770232609482 Angstrom
I18	-5.092879492392246	4.4081385337918615	5.251890032327323 Angstrom
I19	-5.088680245155075	-0.002424617256442829	5.251890032327323 Angstrom
I20	-3.81755996131756	2.2040692668959307	3.486186130885215 Angstrom
I21	-2.5464397461961235	-4.40571406200399	1.7051770232609482 Angstrom
I22	-2.546439746196123	8.81870155259601	1.7051770232609482 Angstrom
I23	-2.5464396698449243	-0.002424617256442829	1.7051770232609482 Angstrom
I24	-2.542240342438994	4.408138275915757	1.7051770232609482 Angstrom
I25	-1.2753195348922461	-2.2040692735081384	5.251890032327323 Angstrom
I26	-1.2753195348922461	11.020346341091862	5.251890032327323 Angstrom
I27	-1.271120287655076	6.609783190043557	5.251890032327323 Angstrom
I28	-1.2711202876550756	-6.614632424556443	5.251890032327323 Angstrom
I29	-1.2711202036687568	2.20649374529601	5.251890032327323 Angstrom
I30	-3.817560045860091e-9	-4.408138540404069	3.486186130885215 Angstrom
I31	-3.817560045860091e-9	8.81627707419593	3.486186130885215 Angstrom
I32	1.2711202876550756	-6.614632424556443	1.7051770232609482 Angstrom
I33	1.271120287655076	6.609783190043557	1.7051770232609482 Angstrom
I34	1.2753196150610058	-2.204069531384243	1.7051770232609482 Angstrom
I35	1.2753196150610058	11.020346083215756	1.7051770232609482 Angstrom
I36	2.5464397538312427	8.81870155259601	5.251890032327323 Angstrom
I37	2.546439753831243	-4.40571406200399	5.251890032327323 Angstrom
I38	3.8175599575	    6.6122078073	    3.4708806325062493 Angstrom
I39	5.088680168803876	8.81870155259601	1.7051770232609482 Angstrom
I40	5.088680168803877	-4.40571406200399	1.7051770232609482 Angstrom
I41	6.359800380107753	-2.2040692735081384	5.251890032327323 Angstrom
I42	6.359800380107753	11.020346341091862	5.251890032327323 Angstrom
I43	6.363999627344924	-6.614632424556443	5.251890032327323 Angstrom
I44	6.363999627344924	6.609783190043557	5.251890032327323 Angstrom
I45	7.635119911182439	-4.408138540404069	3.486186130885215 Angstrom
I46	7.635119911182439	8.81627707419593	3.486186130885215 Angstrom
I47	7.635119915	        0.	                3.4708806325062493 Angstrom
I48	8.906240126303876	2.20649374529601	1.7051770232609482 Angstrom
I49	8.906240202655075	-6.614632424556443	1.7051770232609482 Angstrom
I50	8.906240202655075	6.609783190043557	1.7051770232609482 Angstrom
I51	8.910439530061005	11.020346083215756	1.7051770232609482 Angstrom
I52	10.177360337607753	4.4081385337918615	5.251890032327323 Angstrom
I53	10.181559668831243	8.81870155259601	5.251890032327323 Angstrom
I54	11.45267986868244	2.2040692668959307	3.486186130885215 Angstrom
I55	12.727999487561005	4.408138275915757	1.7051770232609482 Angstrom
I56	13.999119626331243	2.20649374529601	5.251890032327323 Angstrom
End of Basis Set

XFIELD
5469 Angstrom
...

CHOL
 CHOInput
 THRCholesky
 1.0d-7
 ENDChoinput

End of Input

Thank you for reading this until the end, I hope we can fix this together!

Last edited by Max.M.K. (2022-10-10 16:29:58)

Ignacio · 2022-10-10 12:51:02

You'll make everyone's life easier, including you own, if you reduce your input to the minimum necessary to show the problem, e.g.:

&SEWARD
  Basis Set = Sm.ANO-RCC-VTZP
    Sm 0.0 0.0 0.0
  End of Basis Set
  Cholesky

It seems some parts of the code are extremely greedy and (needlessly) use up all available memory, leaving nothing for later procedures. There will be a fix available soon.

Max.M.K. · 2022-10-10 14:51:08

Thank you for the answer, just to be sure that I understand:

- There is no inherent issue with the basis sets, and the memory error is valid, even though there is no information about how much memory is required.

- The issue will be fixed in a future release of OpenMolcas. (I suppose reinstalling and recompiling at a later date will be sufficient for updating?)

Thanks for letting me know about reducing the input, I'll keep it in mind for the next time I have a problem I can't figure out. Looking forward to hearing back from you specifying on my points above.

Best,
Max

Ignacio · 2022-10-10 16:03:48

The memory error is valid, at least it agrees with what is happening: the code takes all available memory for some task, and then it complains that there is no memory left for some other task. When I say "takes all available memory" I mean that literally, not that it takes the amount it needs, and it happens to be all there is, but that it just takes it all, whatever it is.

If you wait a couple of days, it the fix will be in the master, and depending on how you installed it, it could be enough to run "git pull" and "make".

Max.M.K. · 2022-10-10 16:27:47

Alright!! I'll keep my eyes firmly planted on the patch notes.

Thank you again so much, have a great afternoon!

Molcas Forum

Announcement

#1 2022-10-08 18:01:09

[SOLVED] ANO-RCC-VTZP and memory issues

#2 2022-10-10 12:51:02

Re: [SOLVED] ANO-RCC-VTZP and memory issues

#3 2022-10-10 14:51:08

Re: [SOLVED] ANO-RCC-VTZP and memory issues

#4 2022-10-10 16:03:48

Re: [SOLVED] ANO-RCC-VTZP and memory issues

#5 2022-10-10 16:27:47

Re: [SOLVED] ANO-RCC-VTZP and memory issues

Board footer