updates on gw embedding and ctseg solver interface #85
Conversation
…force diagonal Hloc_0 for ctseg; 3. minor updates on doc and
… 2. fix missing conflicts with the latest unstable branch; 3. check adapol installation
|
There are several changes that are not very localized to GW embedding. Feel free to let me know any comments and suggestions to refactor them! |
There was a problem hiding this comment.
Hey @cnyeh ,
thanks for the PR! This looks great. I just had a minor question for the changes in the main routines but other than that we can merge this.
| @@ -96,9 +96,22 @@ def run_dmft(params, config_file_name=None): | |||
| elif general_params['gw_embedding']: | |||
| from solid_dmft.gw_embedding.gw_flow import embedding_driver | |||
| if mpi.is_master_node(): | |||
| # Checks for h5 file | |||
| if not os.path.exists(general_params['seedname']+'.h5'): | |||
| raise FileNotFoundError('Input h5 file not found') | |||
There was a problem hiding this comment.
Why do we need to ensure that the h5 file is already there? I thought all quantities are obtained from the AIMBES chkpt file and we can start from an empty h5 file? Maybe I missed something, let me know.
There was a problem hiding this comment.
Since this PR replaces dummy_sumk by SumkDFT in the current GW embedding workflow (for automatic block_structure detection), general_params['seedname']+'.h5' is necessary for the constructor of SumkDFT. Please correct me if I am wrong!
There was a problem hiding this comment.
ah yes of course! you are absolutely right. That makes sense. I just rerun all tests (including the GW ones, since the PR CI is not running them) and everything looks good. Can I merge it then?
There was a problem hiding this comment.
Yes, feel free to merge it. Thank you!
The PR contains several updates regarding gw embedding and the interface of ctseg solver.
On the
ct-seginterface:D0_taufor ct-seg whenenforce_off_diag = falseSigma_Hartreefor ctseg w/ retarded interactionsimproved_estimatorandperform_tail_fitat the same timeOn gw embedding:
Hloc_0for a given fermions Weiss fielddummy_sumkbySumkDFTfor proper block_structure detection and orbital symmetrization