FourPhonon1.3-fix20260924: TDEP-format harmonic IFCs and CPU THREE-phonon nanowire RTA.
Native build: oneAPI2024.2, IntelMPI2021.13, sequentialMKL, AVX512 CPU.
Activate the pymatgen/2026.9.24 module for Python/NumPy/SciPy/ASE/Matplotlib. Run bash run.sh to read the provided full numerical arrays and draw plots. Results directories must be new.
For a new CPU calculation: bash prepare.sh; sbatch cpu.sh. Use a currently available compatible CPU partition. The included batch file follows the cluster CPU allocation policy; the sample numerical data do not constitute performance validation on a currently offline CPU compute node.
The three native source calculations were completed on compatible CPUs; no GPU result is substituted.
This is the same real-DFT-derived Si16 nearest-neighbor IFC model as the Fourthorder tutorial. TDEP file is a lossless FORMAT conversion, not temperature-dependent fitting. 5-cube q mesh and IFC truncation are not converged material data. Nanowires use a bulk-phonon continuum boundary model, not atomistic wire relaxation. Four-phonon nanowire transport is not supported in the upstream manual and is not claimed here.
inputs/ holds owned force constants and structures; cases/ is completed native numerical output; reference/ holds independent readback. No VASP proprietary files or vendor executable/source code included.
