Jeff Hammond

You can validate quadrupole and octupole with finite field but it's a huge pain. I don't know if MRCC can do it but that's the other CCSDT and CCSDTQ response...

Dalton can do all the properties analytically. I have input files from 2006 somewhere. CFOUR should be able to do multipole response but I haven't used it for that.

you can do without rpath if you set LD_LIBRARY_PATH correctly. i just prefer the former.

Yes, the above patch is all that is required to compile the whole thing. I didn't test though. I'm building Neko.

https://raw.githubusercontent.com/jeffhammond/HPCInfo/master/buildscripts/llvm-git.sh might be useful for CI. I am trying to figure out how to get nightly builds packaged somewhere.

i recall that one was fixed a while ago but you need 19 for it. https://apt.llvm.org/ has nightly builds that include 19.

My script installs to /opt/llvm/latest and expects you to add paths manually. You can edit the script to install to /usr/local if you prefer that.

Sounds like a good idea although I'm not really worried about a closed source derivative of Vapaa.

Sounds fine to me. I just don't know how all this works because I'm not a Fortran module guru. I do not have unit tests for any internal procedures. I...

https://github.com/pmodels/mpich/issues/7341#issuecomment-2726386689 Why 0 and not MPI_UNDEFINED?