Develop to master - Githubissues

issp-center-dev / mVMC

A numerical solver package for a wide range of quantum lattice models based on many-variable Variational Monte Carlo method

http://www.pasums.issp.u-tokyo.ac.jp/mvmc/en/

GNU General Public License v3.0

59 stars 19 forks source link

Develop to master #27

Closed tmisawa closed 2 years ago

tmisawa commented 2 years ago

bug fix

specify ip in calham.c and calham_real.c as shared #3
fix bug in calculating 2-body Green functions in Lanczos with complex variational parameters
bug fix in fourier2k [density-density correlations]
bug fix for interall in UHF calculations

new functions

adding ctest
adding GHActions
adding blocked pfaffian calculations

tmisawa commented 2 years ago

@k-yoshimi @xrq-phys @yomichi @k-ido Now, PFAFFIAN_BLOCKED is "ON" in intel.cmake, aocc.cmake, and fugaku.cmake. PFAFFIAN_BLOCKED is an advanced option and it is better to set PFAFFIAN_BLOCKED=OFF in the default setting. How do you think?

In addition, in obcx and our clusters, I found that -DCMAKE_CXX_FLAGS="-std=c++11" or -std=c++11 is necessary for compiling the blocked pfaffian submodule. I think this is just because the GCC in the clusters is old. However, for usability, it is better to add the above options in the config files.

xrq-phys commented 2 years ago

@tmisawa PFAFFIAN_BLOCKED has a default value OFF. Forcefully setting it to be OFF in config file would result in commands like cmake .. -DPFAFFIAN_BLOCKED=ON being ineffective.

If keeping default OFF is what should be expected, we should delete those lines instead of setting it OFF. Otherwise, persons who wants it be ON shall suffer from inconvenience.

xrq-phys commented 2 years ago

@tmisawa Thanks for confirming.

tmisawa commented 2 years ago

@xrq-phys Thank you for your comments! I commented out PFAFFIAN_BLOCKED OFF.