[QE-developers] [External] Re: [External] Re: Hybrid meta-GGAs in QE

Thu Feb 2 21:51:18 CET 2023

Dear Dr. Ko,

Running SCAN0 with PWscf also results in a segfault - the EXX calculation is initialized, but the calculation immediately stops afterwards.

For example, using ACE for EXX and setting input_dft = SCA0 (same as XC-000L-000L-000L-000L-264L-267L) prints the following lines in pw.x ' s output:
convergence has been achieved in  42 iterations
     Writing meta-gga kinetic term

     Using ACE for calculation of exact exchange

     EXX grid:    31103 G-vectors     FFT dimensions: (  40,  40,  40)
     ACE projected onto    58 (nbndproj) and applied to    58 (nbnd) bands
After which the segfault is thrown (guessing that at least one variable needed for the EXX calculation isn't allocated for meta-GGAs).

Best,
Aaron
________________________________
From: Hsin-Yu Ko <hsin-yu.ko at cornell.edu>
Sent: Thursday, February 2, 2023 9:23 AM
To: General discussion list for Quantum ESPRESSO developers <developers at lists.quantum-espresso.org>
Cc: Aaron Kaplan <kaplan at temple.edu>; Paolo GIANNOZZI <paolo.giannozzi at uniud.it>
Subject: [External] Re: [QE-developers] [External] Re: Hybrid meta-GGAs in QE

Hi Aaron,

While there is not a systematic implementation of hybrid meta-GGA from libxc, as Paolo pointed out, I think there is an instance of SCAN0 (25% EXX mixing with SCAN meta-GGA) implemented (https://gitlab.com/QEF/q-e/-/blob/develop/XClib/dft_setting_routines.f90#L166<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.com%2FQEF%2Fq-e%2F-%2Fblob%2Fdevelop%2FXClib%2Fdft_setting_routines.f90%23L166&data=05%7C01%7Ckaplan%40temple.edu%7C5082fea1d5d1417dfdfd08db05290c54%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109446082237004%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=B2LtR6Jaqpi8lyQJnr95a8cvwNI6Oq4JlUdDAkk5hHE%3D&reserved=0>). You can probably follow a similar procedure to introduce the needed meta-GGA part from libxc.

Best,
Hsin-Yu

--
Hsin-Yu Ko
Postdoctoral Research Fellow
Department of Chemistry and Chemical Biology
Cornell University

________________________________
From: developers <developers-bounces at lists.quantum-espresso.org> on behalf of Aaron Kaplan <kaplan at temple.edu>
Sent: Thursday, February 2, 2023 8:55 AM
To: Paolo Giannozzi <paolo.giannozzi at uniud.it>; General discussion list for Quantum ESPRESSO developers <developers at lists.quantum-espresso.org>
Subject: Re: [QE-developers] [External] Re: Hybrid meta-GGAs in QE

Thank you, Dr. Giannozzi.

Agreed that meta-GGAs + unscreened-EXX may be unusable in a PW code, but I'm tasked with determining that.

I will try to follow the implementation of hybrid-GGAs and see if it's feasible to add the functionality for hybrid meta-GGAs.

Best,
Aaron
________________________________
From: Paolo Giannozzi <paolo.giannozzi at uniud.it>
Sent: Thursday, February 2, 2023 4:17 AM
To: General discussion list for Quantum ESPRESSO developers <developers at lists.quantum-espresso.org>
Cc: Aaron Kaplan <kaplan at temple.edu>
Subject: [External] Re: [QE-developers] Hybrid meta-GGAs in QE

Hybrid meta-GGA functionals are not currently implemented. I don't think
it is difficult, but: hybrid functionals are computationally heavy,
meta-GGAs are numerically unstable, the combination of the two may turn
out to be unusable

Paolo

On 2/1/23 22:17, Aaron Kaplan wrote:
>
> Hello,
>
> I'm a former student in John Perdew's group, currently working with Mike
> Klein, and one of the r2 SCAN developers.
>
> I'm curious if it's possible to use hybrid meta-GGAs with PW in QE 7.1 -
> I've tried many variations of this calculation but a segfault is always
> thrown when the ACE EXX calculation starts.
>
> I saw in the mail archives that PW in QE 6.* could not perform hybrid
> meta-GGA calculations
> (https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fpipermail%2Fusers%2F2020-February%2F044020.html&data=05%7C01%7Ckaplan%40temple.edu%7C15aba46157ac483cf47308db04fe55cf%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109262658111891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=5TiSFsCKjOl5xoQo%2BZBRqmqiAk1Skbg85nqWXZxb%2Bzo%3D&reserved=0 <https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fpipermail%2Fusers%2F2020-February%2F044020.html&data=05%7C01%7Ckaplan%40temple.edu%7C15aba46157ac483cf47308db04fe55cf%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109262658111891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=5TiSFsCKjOl5xoQo%2BZBRqmqiAk1Skbg85nqWXZxb%2Bzo%3D&reserved=0<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fpipermail%2Fusers%2F2020-February%2F044020.html&data=05%7C01%7Ckaplan%40temple.edu%7C5082fea1d5d1417dfdfd08db05290c54%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109446082237004%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Z8HkFisRFHA6Ei53Y%2BpxkP6001mZyolS0MfYW0jmWj4%3D&reserved=0>>), but am not sure if that still applies to QE 7.
>
> I'll note a few other issues: it seems like LibXC's hybrid GGAs don't
> work either (e.g., the global hybrid of PBEsol, ID = 274), and newer
> versions of LibXC cannot be used because they removed identifiers like
> HYB_GGA_* and HYB_MGGA_* (e.g., in LibXC v. 6 and on, a hybrid GGA is
> identified as a GGA).
>
> Best,
> Aaron Kaplan
>
> ________________________________________________
> The Quantum ESPRESSO community stands by the Ukrainian people
>   and expresses its concerns about the devastating effects that
> the Russian military offensive has on their country and on the
> free and peaceful scientific, cultural, and economic cooperation
> amongst peoples.
> _______________________________________________
> developers mailing list
> developers at lists.quantum-espresso.org
> https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fmailman%2Flistinfo%2Fdevelopers&data=05%7C01%7Ckaplan%40temple.edu%7C15aba46157ac483cf47308db04fe55cf%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109262658111891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=%2B7BSyzGDTsUHs9kykpGCctwmfRhDljGpsjp6e9d%2FYLo%3D&reserved=0<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fmailman%2Flistinfo%2Fdevelopers&data=05%7C01%7Ckaplan%40temple.edu%7C5082fea1d5d1417dfdfd08db05290c54%7C716e81efb52244738e3110bd02ccf6e5%7C0%7C0%7C638109446082237004%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=q0ZWprC6r2np6wT255sXTbLqbw%2BC9EYyIo8285Istm0%3D&reserved=0>

--
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20230202/5919efb3/attachment-0001.html>