Skip to content

Latest commit

 

History

History
329 lines (290 loc) · 17.5 KB

OVERVIEW.md

File metadata and controls

329 lines (290 loc) · 17.5 KB

ECCE overview

ECCE, the Extensible Computational Chemistry Environment, is a graphical environment for computational chemistry. A molecule or a periodic system is built or imported, a calculation is set up for one of several codes, the job is submitted to a workstation or a cluster and followed while it runs, and the results are read back and shown: energies, structures, orbitals, densities, spectra and trajectories. Every calculation is kept, with its input, output and parsed properties, in projects in the Organizer.

How ECCE is organised

Two ways to run ECCE. With an ECCE server (the default on one computer, or a central server for a lab or class): the client saves to and loads from the server, which holds the data server and message broker. On its own (optional): the client keeps its data in a local folder with its own broker, and there is no ECCE server. In both, the client's job agent submits jobs over ssh to a compute machine and follows them.

The client (ecce-client) is the windows you work in, plus a job agent that submits and follows jobs. Normally it works with an ECCE server (ecce-server: a data server and a message broker) that stores projects and results. The client can instead run on its own, keeping its data in a folder on your disk (Edit → Preferences → Data folder; off by default on Linux, the default on macOS and Windows); it still starts a private message broker, ECCE's own, but no ECCE server. Neither server nor client runs calculations, except ECCE-QM on the client's own computer: they run on a compute machine, a workstation or an HPC cluster, where the client submits each job over ssh.

There are three deployment modes:

  • Local: client and server on one computer, for one user; each user has a private broker reachable only by their own processes.
  • Central server: one data server and broker for a group, lab or class; students or colleagues connect to it from their own computers as clients (ecce -remote).
  • Shared broker: several users run ECCE on one shared machine (an application server) and use one system broker service.

Getting started describes how each mode is set up.

General features

  • Build molecular models and periodic systems, or import a structure and work from that.
  • Set up calculations through one interface for these codes; each is built, submitted, monitored and parsed back:
    • ECCE-QM, a Hartree–Fock and density functional program built into ECCE, for small molecules; it runs on the computer you work at and needs nothing else installed;
    • ORCA, Gaussian 16, NWChem, MOPAC and xTB (GFN2-xTB, GFN1-xTB and GFN-FF through the xtb program);
    • GROMACS, CP2K and Quantum ESPRESSO (experimental).
  • Check the input before it goes out: Verify inspects the generated input file before it is submitted, and the Calculation Editor lists in a notes panel what the chosen code will and will not use.
  • Choose basis sets graphically, with the code's own built-in sets used where they match.
  • Submit to workstations, clusters and supercomputers, through Slurm, PBS, SGE, LSF, Moab and HTCondor, or by running directly on the machine without a batch system.
  • Follow a job while it runs: energies, geometry steps, convergence and the current stage of the run appear in the Progress box and the Viewer before the job ends.
  • Visualise results in 3-D: molecular orbitals, electron and spin density, electrostatic potential, vibrational modes with animation, dipole moments, optimisation, scan and IRC paths, and MD trajectories.
  • Analyse orbitals with the MO diagram (experimental), the Density Map and the Density Profile.
  • Import output from jobs run outside ECCE, for centres where ECCE cannot submit directly.
  • Teach with ready-made example sets for general chemistry, organic chemistry and spectroscopy, each calculation with a note on what it shows.
  • Run one server for a group, with students or colleagues connecting to it as clients.

Platforms

  • Linux: client and server packages for Debian 13 (tested), and for Ubuntu 24.04, RHEL/Rocky 9 and Fedora (built on those systems, not tested).
  • Windows 10 and 11: client, as an installer or zip.
  • macOS 11 or newer, Apple silicon and Intel: client, experimental.

Capabilities

Building

  • Builder with element and atom-shape choice, Clean (geometry clean-up) and Add Hydrogens; bonds carry no order, the atom shape sets the geometry.
  • Measure toolbar: distances, angles and torsions, with units.
  • Structure Library: amino acids, DNA and RNA bases, ligands, polyoxometalates, simple structures, and a Teaching folder of small molecules at their measured bond lengths.
  • Peptide and DNA builders.
  • Point groups: find, symmetrise, generate symmetry-equivalent atoms.
  • Organometallic complexes: σ ligands and π ligands (cyclopentadienyl, benzene) bonded to a metal.
  • Dummy atoms as reference points.
  • Geometry constraints: fixed bonds, angles and torsions; relaxed-scan definitions (Scan...).
  • Periodic systems: cell from lattice parameters, atoms generated from a space group (listed with number and crystal system), Miller-index planes.
  • Layer Builder: stacks of one layer with a spacing, a sideways shift, and ions and water between the layers (graphite, layered oxides, clays).
  • Structure import: XYZ, PDB, CAR and MVM; multi-frame XYZ and .trj trajectories.

Codes and methods

ECCE-QM ORCA Gaussian 16 NWChem MOPAC xTB
Hartree–Fock, DFT ✓ ✓ ✓ ✓ – –
MP2 – ✓ ✓ ✓ – –
Coupled cluster – ✓ ✓ ✓ – –
CI, QCISD – – ✓ – – –
Semi-empirical – – ✓ – ✓ ✓ (tight binding)
Energy, gradient ✓ ✓ ✓ ✓ ✓ ✓
Geometry optimisation ✓ ✓ ✓ ✓ ✓ ✓
Frequencies, IR ✓ ✓ ✓ ✓ ✓ ✓
Raman activities ✓ ✓ ✓ – – –
TS search from one structure ✓ ✓ ✓ ✓ – –
TS from two structures (NEB-TS, QST2/QST3) – ✓ ✓ – – –
IRC ✓ ✓ ✓ – – –
Relaxed scan ✓ ✓ ✓ – – –
Initial Hessian choice – ✓ ✓ – – –
Implicit solvent ✓ ✓ ✓ ✓ – ✓
Dispersion correction – ✓ ✓ ✓ – ✓ (built in)
TD-DFT excited states, UV-vis – ✓ ✓ – – –
NMR shielding – ✓ ✓ ✓ – –
NBO analysis – ✓ ✓ – – –
Start from a previous calculation's orbitals ✓ ✓ ✓ ✓ – –
Import of an output run elsewhere ✓ ✓ ✓ ✓ ✓ ✓

Per code:

  • ECCE-QM: RHF, UHF, ROHF; SVWN, PBE, B3LYP, PBE0 (restricted, unrestricted, restricted open shell); STO-3G, Pople, Dunning and Karlsruhe basis sets; analytic Hessian for HF and LDA, GGA and global-hybrid DFT (closed shell and unrestricted); Opt-TS by eigenvector following; CPCM solvent (water, methanol, ethanol, acetonitrile, DMSO, chloroform, toluene, hexane or a custom dielectric); improved virtual orbitals (HF); Pipek–Mezey localised orbitals; Mulliken, Löwdin, CHELPG and Merz–Kollman charges; Raman activities with the static polarisability.
  • ORCA: HF, DFT, MP2, CCSD, CCSD(T) and their DLPNO forms; RIJCOSX; D3BJ, D3(0) and D4 dispersion; CPCM and SMD solvent; TD-DFT (number of states, singlets/triplets, Tamm–Dancoff); NEB-TS and OptTS; NMR (Magnetic run type); CHELPG, Hirshfeld and NBO charges.
  • Gaussian 16: HF, DFT, MP2–MP5, CIS, CISD, QCISD, CCD, CCSD, CCSD(T) and semi-empirical methods (AM1, PM3, PM6, PM7, MNDO, MINDO/3, CNDO, INDO); PCM, CPCM, IPCM, SCIPCM and SMD solvent; Grimme D2, D3, D3(BJ) and Petersson–Frisch dispersion; TD-DFT; QST2/QST3; CalcFC, CalcAll, ReadFC and RecalcFC; NMR, polarisability and absorbance run types; Merz–Kollman, CHelpG, Hirshfeld, CM5 and NBO (6 or 7) charges.
  • NWChem: HF, DFT, MP2, CCSD and CCSD(T); COSMO solvent; Grimme D1–D3(BJ) dispersion; ESP-fitted charges; NMR shielding (GIAO), hyperfine and indirect spin–spin coupling; plane-wave PSPW and BAND, and Car–Parrinello dynamics; NWChem MD (Prepare, Energy, Optimize, Dynamics, Solvate).
  • MOPAC: AM1, MNDO, PM3, PM6, PM7 and RM1, restricted or unrestricted; energy, gradient, optimisation and frequencies.
  • xTB: GFN2-xTB, GFN1-xTB and GFN-FF; energy, optimisation (crude to extreme), frequencies and thermochemistry; ALPB or GBSA solvent; orbitals from the Molden file (GFN2, GFN1), atomic charges; runs made outside ECCE can be imported with the files xtb wrote beside the output.
  • Quantum ESPRESSO (experimental): plane waves and pseudopotentials; energy, geometry optimisation and dynamics; cut-off, k-points, smearing and dispersion correction; the cluster's pseudopotential folder is used when the job runs.
  • CP2K (experimental): Quickstep (GPW); energy and geometry optimisation; dispersion and smearing.
  • GROMACS (experimental): an MD study of Prepare, Energy, Optimize, Equilibrate and Dynamics tasks; Prepare from a PDB file (pdb2gmx), a mineral or clay (MINFF or ClayFF, bundled) or the Builder's structure; NVE, NVT and NPT; defines, position restraints, groups, barostat type and extra .mdp lines; pull runs; grompp messages mapped to the editor's fields; import of a finished run as a study.

Calculation setup

  • Calculation Editor per code: Theory, Theory Details, Run Type, Runtype Details, charge, spin multiplicity, basis set.
  • Basis Set Tool: basis sets and effective core potentials from ECCE's library, per element.
  • Memory entered in GB for every code.
  • Notes panel above the log: what is unused by the chosen code (constraints, scans, symmetry), transition-state warnings, why a calculation is not ready.
  • Verify: the generated input checked before submission.
  • Warnings before launch with Launch anyway as the default.
  • Duplicate, Duplicate with Final Orbitals, Reset for Rerun.
  • Symmetry switched off for transition-state searches, IRC, constraints and scans, with a note.

Job submission and monitoring

  • Compute machines registered under Tools > Register Machines: codes, modules, run directory and account per user; codes on the PATH registered at first start.
  • Run on this computer, or on another machine over ssh.
  • Queue managers: Slurm, PBS, SGE, LSF, Moab, HTCondor.
  • One ssh login per compute machine per session (Linux and macOS).
  • Launch Selected: several calculations with their stored launch settings.
  • Reconnect Selected: several jobs reconnected to their calculations after a restart or a lost connection.
  • Machine Browser: machine and queue status, allocation (Slurm sshare or a per-machine command).
  • Progress box: SCF and optimisation plots with force convergence, the current stage of the run (including frequency and NEB stages), MD averages.
  • Job states recorded with the calculation, including the reason for a failed submit; Terminate Calculation, View Run Log, tail of the output file, shell in the run directory.

Analysis and visualisation

  • Energies, energy components, run statistics, point group, ⟨S²⟩.
  • Geometry Trace and Geometry Step Plots of an optimisation; playback with delay and loop.
  • Relaxed Scan and IRC Path panels: graph or table, a structure per point, New calculation from this point, CSV export.
  • Molecular orbitals as surfaces, orbital energies and occupations; the MOs panel lists localised orbitals where computed.
  • MO diagram (experimental): orbitals as levels joined to fragment orbitals, chosen by picking an atom or a bond; a ligand-field view.
  • Density Map: density or one orbital as contours on a plane through the molecule.
  • Density Profile: density, ψ² or ψ along a line or radially around an atom, with each orbital's share; orbitals of an atom named 1s, 2s, 2p…
  • Total density, electrostatic potential on the density surface, spin density with Mulliken spin populations.
  • Charges: Mulliken, Löwdin, Hirshfeld, CM5, CHELPG/ESP-fitted, Merz–Kollman, natural (NPA), APT, MBIS, AIM (Bader), as each code provides them.
  • Dipole moment as an arrow from the negative to the positive pole; polarisability as an ellipsoid; quadrupole and higher moments.
  • NMR: isotropic and anisotropic shielding, shielding tensors, spin–spin coupling, Fermi contact.
  • Vibrations: frequencies with symmetry and description, animation and displacement vectors, IR and Raman spectra, frequency scale factor, spectra compared side by side, mode composition by internal coordinates, CSV export.
  • UV-vis spectrum from TD-DFT, transitions listed with their orbital weights.
  • Reaction Energies panel for a transition state built from two calculations.
  • Compare Energies (Organizer): the difference of two final energies in Hartree, kJ/mol, kcal/mol and eV, with notes on differing code, method or basis set. Its Reaction overview button charts dE and dG of several reactions (reactants -> products, one per line), each dG with its spread over the six low-frequency treatments as an error bar; clicking a reaction lists the modes below 100 cm-1 with their share of S and of the G correction. A reaction whose species differ in code, method, basis set or solvent is refused with the reason. Export: CSV and PNG. The same numbers: ecce-thermo --reactions FILE [--format text|json|csv].
  • MD: MD Summary (final values, averages with standard deviations), trajectory playback in the true box (including dodecahedral and truncated-octahedral cells), every logged series against time with a moving average, speed distribution against Maxwell–Boltzmann, mean-square displacement with the diffusion coefficient, pull force.
  • Copy values from any value list, export plots, maps and spectra.

Data management

  • Organizer: projects and calculations in a tree, run-state icons, a Calculation Summary, buttons for the editor, Builder, Basis Set Tool, Launcher and Viewer.
  • Data on an ECCE server (local or central) or in a local data folder.
  • Import Calculation from Output File: ECCE-QM, ORCA, Gaussian 16, NWChem, MOPAC, xTB.
  • Workflows: File > New > Workflow... keeps a chain of calculations together and draws it as a tree of boxes; a new step of any code starts from the final geometry, charge and multiplicity of the box it comes from (e.g. an xTB optimisation refined with ORCA), several steps from one box are branches, and Launch Ready Steps starts every step whose source has finished.
  • Import GROMACS Run.
  • Manage Files, View Input/Output File, reviewed/unreviewed marks.

Teaching

File > Open Example Calculations imports example sets, already read in, sorted by topic, each calculation with a note:

  • Getting Started: small calculations with ECCE-QM, MOPAC, NWChem, ORCA and xTB, imported at first start.
  • Workflows: hydrated Na+: xTB optimises arrangements of Na+(H2O)6 and ORCA refines each result in the gas phase and in water, the water step starting from the Hessian of the gas step.
  • General Chemistry: atoms; ionisation energies and electron affinities; Koopmans' theorem; molecular orbitals of diatomics; polar bonds and dipole moments; VSEPR; Lewis structures and resonance; gases, liquids and solids by molecular dynamics.
  • Organic Chemistry: bonding in hydrocarbons; conformations (relaxed torsion scans); reactions (SN2, Diels–Alder: transition states and IRC).
  • Spectroscopy: sulfate coordination, carbonyl compounds, amides and peptides, sugars, small molecules: IR, Raman and orbitals.
  • Computational: a GROMACS study of the Trp-cage mini-protein, from Prepare to Dynamics.

Opening a set again adds what is missing without changing existing calculations.

Screenshots

Click any image for the full-size version.

The Organizer, with a completed ORCA calculation selected

The Organizer: projects and calculations on the left, each with a run-state icon, and a summary of the selected one in the middle, with buttons for its editor, builder, basis set tool, launcher and viewer — here a completed ORCA calculation on benzene.

The viewer, showing a molecular orbital of benzene

The viewer: a calculation's orbitals and energies beside the structure they were computed for — benzene, from ORCA.

Benzene's highest occupied molecular orbital The electrostatic potential on benzene's surface A vibrational mode of water, drawn as displacement vectors
Benzene's π HOMO (ORCA) Electrostatic potential on the surface (ORCA) A vibrational mode of water, as displacement vectors (Gaussian 16)

A qualitative MO correlation diagram for water

A qualitative MO correlation diagram — water's orbitals, from Gaussian 16, against the oxygen on one side and the hydrogens' symmetry orbitals on the other, with the non-bonding lone pair picked out. The same diagram is drawn from any of the supported codes. Experimental.