Examples

This page indexes every input file in examples/. There are 84 TOML files. Three Python scripts that build methods from AoIntegrals blocks are in examples/python/ (Python bindings). Run a TOML file from the repository root:

OPENBLAS_NUM_THREADS=1 cargo run --release --bin ferric -- examples/water-rhf.toml

What CI checks

CI parses every shipped example and runs the three listed below. The test all_shipped_examples_parse (in crates/ferric-cli/src/config.rs) loads every examples/*.toml through the strict parser. It also checks that each file's [molecule] xyz path exists. It does not execute the calculation or compare any energy.

Three examples are also executed by integration tests in crates/ferric-cli/tests/:

Example	Test	What the test asserts
`water-rhf.toml`	`epistemic_warning.rs`, `verbose_trace.rs`, `qmmm_reports_the_pqr_as_its_geometry.rs`	The run completes and prints no `[warning]` grade line.
`water-tdhf-static-alpha.toml`	`epistemic_warning.rs`	The Smoke-grade warning goes to stderr and not to stdout.
`water-qmmm.toml`	`qmmm_reports_the_pqr_as_its_geometry.rs`	The embedded energy `-74.9653197421` appears in the output.

Every other number in the "Reference in header" column below is copied verbatim from that file's comment block, and CI does not re-check it. Where the header cites a Rust test as the reference, that test is the check.

Examples that need extra setup

water-mp2v.toml, water-scs-mp2-2terfc.toml, water-attmp2-terfc.toml and any rs-mp2-rpa run with attenuator = "terf" need the tempered-erfc interpolation tables: point FERRIC_TERF_TABLE_DIR at them. Without it these runs stop with an error.
benzene-dfb3lyp-mpi.toml is an ordinary input; its header gives the --features mpi build and the mpirun launch line.

SCF and DFT

See SCF and DFT.

File	System / basis	`kind` / `task`	Reference in header	Notes
`water-rhf.toml`	H2O / STO-3G	`rhf` / energy	—	`water-qmmm.toml`'s header gives this geometry's energy as −74.9631468000. The file is commented to demonstrate `screening`.
`benzene-rhf.toml`	benzene / cc-pVDZ	`rhf`	—	Exact 4-index J/K.
`benzene-rhf-dfj.toml`	benzene / cc-pVDZ	`rhf`	—	RI-J only (`cc-pvdz-ri`).
`benzene-rhf-rijk.toml`	benzene / cc-pVDZ	`rhf`	—	RI-JK (`def2-universal-jkfit`).
`benzene-rhf-def2.toml`	benzene / def2-SVP	`rhf`	—
`benzene-rhf-def2-rijk.toml`	benzene / def2-SVP	`rhf`	—	RI-JK.
`decane-rhf.toml`	decane / STO-3G	`rhf`	—	`k_builder = "link"`.
`water-rhf-cosx.toml`	H2O / cc-pVDZ	`rhf`	"E(COSX) − E(direct) is reported in crates/ferric-scf/tests/cosx_scf.rs"	COSX exchange.
`water-rhf-symmetry.toml`	H2O / STO-3G	`rhf` + `[scf] symmetry`	—	Point-group-blocked RHF (C2v); `irrep_nelec` shown commented.
`h_uhf.toml`	H atom / STO-3G, doublet	`uhf`	—
`h2_opt.toml`	stretched H2 / STO-3G	`rhf` / optimize	—
`water-frequencies.toml`	H2O / STO-3G	`rhf` / frequencies	—	Finite-difference Hessian (`hessian = "fd"`). Check `Hessian asymmetry`.
`nh2-rohf-frequencies.toml`	NH2 doublet / cc-pVDZ	`rohf` / frequencies	`crates/ferric-scf/tests/validation_rohf_hessian.rs`	Analytic ROHF Hessian.
`benzene-dfb3lyp.toml`	benzene / def2-SVP	`ksdft` B3LYP	—	RI-JK is on automatically for `ksdft`.
`benzene-dfb3lyp-mpi.toml`	benzene / cc-pVDZ	`ksdft` B3LYP	—	Run under `mpirun` with the MPI build (see its header).
`water-wb97xv.toml`	H2O / cc-pVDZ	`ksdft` wB97X-V	—
`water-pbe-d3bj.toml`	H2O / cc-pVDZ	`ksdft` PBE + D3(BJ)	—	Prints E(KS-DFT) and E(D3BJ) separately.
`water-pbe-mbd.toml`	H2O / cc-pVDZ	`ksdft` PBE + MBD@rsSCS	—	Prints E(KS-DFT) and E(MBD@rsSCS) separately.
`water-pbe-pruned-grid.toml`	H2O / cc-pVDZ	`ksdft` PBE / energy	"removes ~23% of the grid points" (at 75×110)	Validated in `crates/ferric-dft/tests/grid_prune_live_scf.rs`.
`h2-lda-opt.toml`	H2 / STO-3G	`ksdft` LDA / optimize	—
`water-qmmm.toml`	H2O + Na⁺ (PQR) / STO-3G	`rhf` + `[qmmm]`	"vacuum −74.9629466809, embedded −74.9653197421, i.e. −1.489 kcal/mol from the ion at 4 A. Verified against `ferric.run_rhf(point_charges=...)` to all 10 digits."	The vacuum number is at the PQR geometry, not the xyz. The embedded number is asserted by a test.
`water-pcm.toml`	H2O / STO-3G	`rhf` + `[pcm]`	—	IEF-PCM water (`solvent = "water"`, ε = 78.4).
`water-rhf-smeared-charge.toml`	H2O / STO-3G	`rhf` + `[external_potential]`	—	One Gaussian-smeared charge (`width`, Bohr) and one point charge.
`o2-uhf-stability-descent.toml`	O2 triplet / STO-3G	`uhf`	"the descent follows the downhill eigenvector to the UHF minimum (-147.63530 Ha)"	`[scf] stability_descent = true`.
`h2-cell-rhf.toml`	H2 / STO-3G, 4 Bohr cubic cell	`rhf` + `[cell]`, Gamma point	—	Periodic, dense J/K, `exxdiv = "ewald"`.
`h2-cell-kpts-rhf.toml`	H2 / STO-3G, 4 Bohr cubic cell	`rhf` + `[cell]`, 2×2×2 k-mesh	—	Periodic k-point RHF, dense J/K.
`h2-cell-kpts-ksdft.toml`	H2 / STO-3G, 4 Bohr cubic cell	`ksdft` + `[cell]`, 1×1×2 k-mesh	PBE0	Periodic k-point closed-shell KS-DFT, dense J/K.
`h2-cell-kpts-rhf-stress.toml`	H2 / STO-3G, 4 Bohr cubic cell	`rhf` + `[cell]`, 1×1×2 shifted (`centring = "mp"`) k-mesh, `stress = true`	—	Analytic k-point stress (Hartree/Bohr³ and GPa, also in the run log's `run_end`), dense J/K.
`h2-cell-kpts-rhf-opt.toml`	stretched H2 / STO-3G, 4 Å cubic cell	`rhf` + `[cell]` 1×1×3 k-mesh / optimize	—	k-point atom relaxation at a fixed lattice, analytic k-point force, dense J/K.
`h2-cell-kpts-rks-opt.toml`	stretched H2 / STO-3G, 4 Å cubic cell	`ksdft` + `[cell]` 1×1×3 k-mesh / optimize	PBE0	k-point atom relaxation at a fixed lattice, analytic k-point Kohn-Sham force (global hybrid), dense J/K.
`h-cell-kpts-uks.toml`	H atom doublet / STO-3G, 4 Bohr cubic cell	`uhf` + `[dft]` + `[cell]`, 1×1×2 k-mesh	PBE	Periodic k-point open-shell Kohn-Sham (UKS), dense J/K; `rohf` gives ROKS/ROHF at a mesh.
`h2-cell-rks-opt.toml`	stretched H2 / STO-3G, 4 Å cubic cell	`ksdft` PBE + `[cell]` / optimize	—	Gamma-point atom relaxation at a fixed lattice, analytic periodic force, dense J/K.

MP2 family

See The MP2 family.

File	System / basis	`kind`	Reference in header	Notes
`water-rimp2.toml`	H2O / cc-pVDZ	`rimp2`	—	All-electron.
`water-rimp2-frozen-core.toml`	H2O / cc-pVDZ	`rimp2`	Expected log line: "`[ferric] frozen core: 1 orbital(s) frozen from [mp2] frozen_core = "auto"`"
`water-rimp2-rijk.toml`	H2O / cc-pVDZ	`rimp2`	—	RI-JK reference SCF (`[scf] jk = "rijk"`, `jk_aux = "def2-universal-jkfit"`); the MP2 RI aux stays `cc-pvdz-ri`.
`water-rimp2-local.toml`	H2O / 6-31G	`rimp2` + `[local]`	—	Amplitude-threshold local MP2, `eps = 1e-4`. Sets `reference = true`, so it also computes the exact RI-MP2 and prints the local error against it (off by default).
`alkane8-rimp2-local-direct.toml`	octane / 6-31G	`rimp2` + `[local]`	—	Integral-direct local MP2 (`integral_direct = true`), every locality map at its default.
`water-drpa.toml`	H2O / 6-31G	`drpa`	—	Exact dRPA (no `[local]`).
`water-drpa-local.toml`	H2O / 6-31G	`drpa` + `[local]`	—	Amplitude-threshold dRPA, `eps = 1e-4`, `reference = true` (prints the error against the canonical plasmon dRPA). A comment shows the `eps_sweep` form.
`water-mp3.toml`	H2O / cc-pVDZ	`mp3`	—
`water-oo-rimp2.toml`	H2O / cc-pVDZ	`oo-rimp2`	—	Proven (narrow) grade.
`water-attmp2.toml`	H2O / aug-cc-pVDZ	`att-rimp2`	—	ω = 0.420 Å⁻¹.
`water-attmp2-terfc.toml`	H2O / aug-cc-pVDZ	`att-rimp2`	—	`att_operator = "terfc"`, `att_r0` = 1.05 Å. Needs the terf tables.
`alkane8-attmp2-terfc-local-direct.toml`	C8H18 / 6-31G	`att-rimp2`	`[local]`	Integral-direct amplitude-threshold LMP2 with a decoupled terfc seam (`att_omega`, r0·ω = 4). Needs the terf tables.
`water-scs-mp2.toml`	H2O / cc-pVDZ	`scs-mp2`	—	Grimme coefficients (defaults).
`water-scs-mp2-2terfc.toml`	H2O / cc-pVDZ	`scs-mp2-2terfc`	—	Thesis defaults r0 = 0.75/1.05 Å. Needs the terf tables.
`water-laplace-rimp2.toml`	H2O / cc-pVDZ	`laplace-mp2`	—
`water-laplace-sos-mp2.toml`	H2O / cc-pVDZ	`laplace-sos-mp2`	—
`water-mp2v.toml`	H2O / aug-cc-pVDZ	`mp2-v`	—	The header warns that aDZ is outside the fitted basis. Needs the terf tables.
`water-rs-mp2-rpa.toml`	H2O / aug-cc-pVDZ	`rs-mp2-rpa`	—	Smoke grade.

Coupled cluster and double hybrids

See Coupled cluster and RPA and GW § double hybrids.

File	System / basis	`kind`	Reference in header	Notes
`water-ccsd.toml`	H2O / cc-pVDZ	`ccsd`	"PySCF CCSD run on the same density-fitted integrals gives E_corr = -0.2135061893 Ha"	All electrons, aux `cc-pvdz-ri`.
`water-linlccd.toml`	H2O / 6-31G	`linlccd`	—	Exact LinLCCD(hh).
`water-linlccd-local.toml`	H2O / 6-31G	`linlccd` + `[local]`	—	Amplitude-threshold LinLCCD(hh), `eps = 1e-4`.
`water-ccd.toml`	H2O / STO-3G	`ccd`	—
`water-ccsd-t.toml`	H2O / STO-3G	`ccsd(t)`	—	Prints the CCSD correlation energy, the (T) correction and the total.
`water-wb97xlv.toml`	H2O / 6-31G	`wb97x-l-v`	—	λ = 0.6, ω = 0.1 Bohr⁻¹ (published values). Proven (narrow) grade.
`water-b2plyp.toml`	H2O / cc-pVDZ	`b2plyp`	—	Spike grade. Aux `cc-pvdz-rifit` (an alias of `cc-pvdz-ri`).

RPA, C6 and properties

See RPA and GW.

File	System / basis	`kind`	Reference in header	Notes
`water-pdep-rpa.toml`	H2O / cc-pVDZ	`pdep-rpa`	—	Writes eigenpotential cube files.
`h2o-pdep-rpa-props.toml`	H2O / cc-pVDZ	`pdep-rpa`	—	NPZ export.
`benzene-pdep-rpa.toml`	benzene / def2-SVP	`pdep-rpa`	—	RI-JK SCF.
`benzene-pdep-rpa-export.toml`	benzene / cc-pVDZ	`pdep-rpa`	—	Exports cube files.
`benzene-rijk-pdep-rpa.toml`	benzene / cc-pVDZ	`pdep-rpa`	—	Full rank (`trunc_thresh = 0`).
`water-c6-pdep.toml`	H2O / aug-cc-pVTZ	`pdep-rpa`, `c6_source = "pdep"`	"DOSD molecular reference (Meath/Toulouse): C6(H2O–H2O) = 45.3 a.u."	Compare the printed "molecular C6" line against it, not the sum of the NPZ `c6_iso` entries. Writes to `/tmp`.
`argon-c6-rpa-pbe.toml`	Ar / aug-cc-pVTZ	`pdep-rpa` @PBE	"C6(Ar-Ar) = 56.4 a.u. vs DOSD 64.3 (-12%); the full He/Ne/Ar sweep gives mean \|err\| 8.9% at RPA@PBE — vs 39% at RPA@HF"	Writes to `/tmp`.
`water-tdhf-static-alpha.toml`	H2O / cc-pVDZ	`tdhf-static-polarizability` @PBE	"alpha_iso = 5.20 a.u. … against the DOSD reference 9.64 a.u.: 46% LOW"	Sets `[gw] scissor = 0.36` Ha; at `scissor = 0` the run is refused (negative α diagonal). Static α only. Executed by a test (for the warning only).

GW and BSE

File	System / basis	`kind`	Reference in header	Notes
`water-g0w0-pbe.toml`	H2O / cc-pVDZ	`gw` G0W0@PBE	"HOMO IP should match … PySCF gw_ac reference (11.1714 eV) to <0.1 eV" (`crates/ferric-gw/tests/g0w0_pbe_h2o.rs`)
`oh-ugw.toml`	OH doublet / cc-pVDZ	`gw` U-G0W0@UHF	"alpha-HOMO IP in the ~13-14 eV window, bracketing experiment 13.02 eV" (`crates/ferric-gw/tests/oh_u_g0w0.rs`)
`oh-ugw-rohf.toml`	OH doublet / cc-pVDZ	`gw` U-G0W0@ROHF	—	`[gw] reference = "rohf"`.
`water-bse-tda.toml`	H2O / cc-pVDZ	`bse-tda`	"Measured via this exact TOML (2026-07-18): 8.4572 eV", against a PySCF-integral cross-check of 8.46 eV and a sanity window of [5, 12] eV
`water-augccpvdz-bse-tda.toml`	H2O / aug-cc-pVDZ	`bse-tda`	Literature: "aug-cc-pVDZ CCSDT Delta_Evert = 9.279 eV, f = 0.058; CBS TBE = 7.71+-0.02 eV, f = 0.052+-0.001"	No ferric number is recorded.
`h2co-bse-tda.toml`	H2CO / cc-pVDZ	`bse-tda`	—	Pilot run. The lowest state is dark (f ≈ 0).
`c2h4-bse-tda.toml`	C2H4 / cc-pVDZ	`bse-tda`	—	Pilot run.
`formaldehyde-bse-tda-augdz.toml`	H2CO (QUESTDB geometry) / aug-cc-pVDZ	`bse-tda`	Literature: "lowest singlet 1^1A2 (n->pi*, symmetry-forbidden) at 3.966 eV (QUESTDB TBE…)"	The template for the Thiel-set files below.
`acetaldehyde-bse-tda-augdz.toml`, `butadiene-bse-tda-augdz.toml`, `cyclopropene-bse-tda-augdz.toml`, `ethylene-bse-tda-augdz.toml`, `furan-bse-tda-augdz.toml`, `glyoxal-bse-tda-augdz.toml`, `pyrazine-bse-tda-augdz.toml`, `pyridine-bse-tda-augdz.toml`, `pyrimidine-bse-tda-augdz.toml`	QUESTDB Thiel-set molecules / aug-cc-pVDZ	`bse-tda`	—	Same setup as the formaldehyde file. The reference values are in `testdata/reference/thiel_set_subset.json`.

TDDFT

File	System / basis	`kind`	Reference in header	Notes
`water-tda.toml`	H2O / cc-pVDZ	`tda` (CIS, 5 roots)	—	Default aux `cc-pvdz-rifit` (an alias of `cc-pvdz-ri`).
`water-tddft-pbe.toml`	H2O / cc-pVDZ	`tddft` @PBE (5 roots)	—	Includes the f_xc kernel.

See also Capabilities and validation for the grade of each kind and its evidence, and Input reference for every key.