WannierTools2.7.0 WireBand_calc tutorial. Four small finite cross-sections, own anisotropic cubic 2-orbital model; no DFT/MD/GPU.
Existing native wire output energy is Hartree even though generated gnuplot labels eV. All 10128 printed energies checked exactly against independent sine-chain energies rounded to native 7 decimal Hartree output; converted eV precision bound 1.36056925e-6. Raw-as-eV interpretation failed and stays unqualified.
Third column encodes 255 minus normalized square root of boundary sum times255; corners counted twice, global normalization. Not probability or topological invariant.
Native negative half is reflected positive half; valid here because this real-hopping model has even dispersion. Not valid evidence for nonreciprocal models.
Source compiler oneAPI2023.2/ifort2021.10,IntelMPI2021.10,MKLsequential,ARPACK-NG3.9.1.
View:bash show-reference.sh;plot:bash plot-results.sh;run:sbatch job.slurm inside available CPU partition, or bash run.sh within allocated compatible CPU resources. Never overwrite existing work.
