A noncollinear functional ensemble based on multi-collinear approach #5542
Conversation
| #include <utility> | ||
|
|
||
| // 2x2 complex matrix | ||
| using Matrix2x2 = std::array<std::array<std::complex<double>, 2>, 2>; |
There was a problem hiding this comment.
Please add namespace for your code
There was a problem hiding this comment.
NCLibxc serves as an independent package and it is not designed to have a namespace. However, it can also be added due to the specific program. If you think that it is necessary, you can conduct this small modification so that it fits your requirements.
| std::ofstream log_file("NCLibxc.log", std::ios::out | std::ios::app); | ||
| if (log_file.is_open()) | ||
| { | ||
| log_file << "You are using the multi-collinear approach implemented by Xiaoyu Zhang. Please cite:\n"; |
There was a problem hiding this comment.
which paper should be site?
There was a problem hiding this comment.
This paper hasn't been well prepared. This log is only for tests. We are going to add citation requests after papers published.
| if (PARAM.inp.nspin == 4 && PARAM.inp.multicolin) | ||
| { // noncollinear case added by Xiaoyu Zhang, Peking University, 2024.10.02. multicollinear method for lda Since NCLibxc needs libxc, this part codes will not be used. | ||
|
|
||
| std::cerr << "Error: Multi-collinear approach does not support running without Libxc." << std::endl; |
There was a problem hiding this comment.
please use WARNING_QUIT function to quit the program.
There was a problem hiding this comment.
I am not familiar with the function in ABACUS. If you think that it is necessary, you can conduct this small modification so that it fits your requirements.
| #ifdef USE_LIBXC | ||
| #include "NCLibxc/NCLibxc.h" | ||
| #include "xc_functional_libxc.h" | ||
| #include <tuple> |
There was a problem hiding this comment.
Why do you add so many "include" but not use them?
There was a problem hiding this comment.
It is for future development of this package and more flexible interfaces. We don't submit those parts to ABACUS. If they are annoying, it is OK to delete them.
There was a problem hiding this comment.
I inspect all of them and did some modification.
There was a problem hiding this comment.
As for the problem you mentioned, I didn't modify xc_functional_vxc from the very beginning.
dyzheng
left a comment
dyzheng
left a comment
There was a problem hiding this comment.
Test should be added for this new method!
mohanchen
added
the
Features Needed
|
I have uploaded a new version of NCXC and implement it into ABACUS. Now it supports lda and gga. |
please refer to benchmark_tests.cpp |
could you be more specific? |
|
I have uploaded a new version of NCXC and implement it into ABACUS. Now it supports lda and gga. Local tests are successful. |
|
Before merge, we need to discuss how to maintain such as large amount of new codes, which are mostly written by machine. |
There was a problem hiding this comment.
how to maintain this file?
There was a problem hiding this comment.
I think that situations are very similar with maintaining libxc. ABACUS provides and maintains interface to NCXC and the owner of NCXC will maintain NCXC on its github page. NCXC serves as an external program like libxc.
There was a problem hiding this comment.
how to maintain this file?
There was a problem hiding this comment.
Please see the above reply.
|
@YuleZhang936 and I decied to move this part outside of ABACUS and set up a new repository. Therefore, I will close this PR. |
6 participants
This PR is to merge a noncollinear functional ensemble NCXC programmed by Xiaoyu Zhang into ABACUS. NCXC is based on the multi-collinear approach published on PHYSICAL REVIEW RESEARCH 5, 013036 (2023). This ensemble can currently support LDA and GGA and can be elaborated to mGGA and hybrid functional in the future.
Xiaoyu Zhang