Skip to content

History / BMD Compute Advanced Options

Revisions

  • docs: remove redundant page-title H1s Remove the leading H1 (and the blank line after it) from the 16 pages where it only repeats the GitHub Wiki filename-derived page title, including decorative emoji. Pages whose H1 adds information (Home, VASP, band structure, high-throughput, USPEX, XtalOpt, SQS, Reading & Editing, PyMatGen, Tutorial-Y/YY, Hebrew pages) are unchanged. No other content, code blocks, links or anchors changed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 29, 2026
  • docs: correct attribution of Materials Project U values The MP U values were fitted by the Materials Project to experimental binary formation enthalpies following the approach of Wang, Maxisch and Ceder (2006); they were not fitted by Wang et al. to oxidation enthalpies. Wang 2006 remains the methodological reference. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 29, 2026
  • docs: document BMD Compute automatic DFT+U v1 in Advanced Options - Advanced Options: rewrite the DFT+U section to describe the adopted BMD Compute automatic DFT+U behaviour (bmd_compute c241a64): the Materials Project/pymatgen oxide/fluoride rule and U values (Dudarev, J = 0), the conservative BMD d0 screen with fallback to the MP rule when no oxidation-state guess exists, representative examples, NiO worked example, stage placement per Desired Output (PBE stages only, never HSE06), independence from spin/SOC/dispersion, Custom workflow control, and why BMD energies are not strictly MP-compatible. Cite Anisimov 1991, Dudarev 1998, Wang 2006, Jain 2011 and MP docs. - Intro, INCARs, VASP and Workflows: stop describing DFT+U as purely optional; point to the automatic behaviour. SOC section, reference INCARs and POTCAR path unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 29, 2026
  • docs: use complete BMD Compute workflow INCARs as references - INCARs: replace the isolated references with the complete, correctly numbered stage INCARs of the four standard workflows (Static Energy; Relaxed Structure; Density of States; Band Structure), reproduced exactly. Add HSE06 Static Energy. Comparisons now: PBE GO vs PBE Static; PBE vs HSE06 Static; HSE06 Static vs DOS vs Band (TIME operative only with ALGO = Damped). Reference structure marked "to be confirmed". - Workflows: rename Geometry Optimisation to Relaxed Structure (two identical PBE stages, Stage 2 from Stage 1 structure); separate "BMD workflow facts" from general physics; Pulay stress presented as one reason for repeated relaxation, not BMD's stated rationale; remove "HSE06@PBE"; file hand-offs and KPOINTS left unspecified. - Advanced Options: drop Si-specific wording for the base references. Supplied INCAR settings, SOC status and POTCAR path unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 26, 2026
  • docs: add BMD Compute Workflows page and four-page structure - VASP: expand into DFT fundamentals (Kohn-Sham SCF, exchange- correlation and hybrids, plane-wave truncation and PAW, Brillouin- zone sampling, occupations/smearing, forces and relaxation); FCC-Si teaching example unchanged. - BMD-Compute-Workflows (new): why calculations are staged, what can pass between stages, relax->static rationale incl. Pulay stress, repeated relaxation, PBE non-SCF vs HSE06 zero-weight approach; BMD Compute-specific sequences and hand-offs marked "to be provided". - BMD-Compute-INCARs: keep stage-level explanations only; move Pulay and sequence reasoning to Workflows. Reference INCAR blocks unchanged. - Advanced Options, Home: navigation for the four-page sequence. POTCAR path and SOC status unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 26, 2026
  • docs: add BMD Compute INCAR teaching pages and fix Agent CLI - VASP: rework into a fundamentals page (inputs, SCF/ionic loops, ENCUT, k-points, smearing, EDIFF/EDIFFG, relaxation, restarts, parallelisation); FCC-Si example kept and labelled as teaching. - BMD-Compute-INCARs (new): the four supplied BMD Compute Si reference INCARs reproduced exactly, with settings grouped by role and Relax->Static, PBE->HSE06 and HSE06 DOS->Band comparisons. - BMD-Compute-Advanced-Options (new): concept pages for spin, DFT+U, dispersion, Gamma-only and ions-only; reference deltas pending; SOC marked under validation. - BMD-Agent: correct bmd-agent / JOB_ID / PATH semantics, document --verbose, keep expert subcommands secondary; adopt agreed authority wording. - Manual HSE band-structure and USPEX pages: keep as teaching/manual workflows and link to the reference page; Home links new pages. POTCAR path unchanged pending verification. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DHgs1a4r3pZ6JVJ5dXQvoR

    @leeaburton leeaburton committed Sep 26, 2026