MRCC

Official Resources

• Homepage: https://www.mrcc.hu/
• Documentation: https://www.mrcc.hu/manual/
• Source Repository: Available to licensees
• License: Academic and commercial licenses available

Overview

MRCC (Multi-Reference Coupled Cluster) is a specialized quantum chemistry program suite featuring arbitrary-order coupled cluster and configuration interaction methods. Developed by Mihály Kállay and collaborators at Budapest University of Technology and Economics, it is renowned for implementing the highest-order correlation methods available, including fully automated arbitrary-order CC and CI implementations generated via string-based equations.

Scientific domain: High-order correlation methods, multireference systems, benchmark calculations
Target user community: Researchers requiring very high accuracy or exploring high-order correlation methods

Theoretical Methods

• Hartree-Fock (RHF, UHF, ROHF)
• Density Functional Theory (DFT with Libxc)
• Møller-Plesset perturbation theory (MP2 through arbitrary order)
• Coupled Cluster up to arbitrary order:
  • CCSD, CCSD(T), CC(T), CCSDT
  • CCSDTQ, CCSDTQP, CCSDTQPH
  • CC(n), CC[n] series
• Multireference CC (Mk-MRCC, BW-MRCC)
• Configuration Interaction up to full CI
• Linear Response CC for excited states
• Equation-of-Motion CC (EOM-CCSD, EOM-CCSDT)
• Symmetry-Adapted Perturbation Theory (SAPT)
• Local Natural Orbital methods (LNO-CCSD(T))
• Explicitly correlated F12 methods
• Analytical gradients (CCSD, CCSD(T), LNO-CCSD(T))
• Relativistic methods (DKH, X2C)
• Automated generation of CC equations

Capabilities (CRITICAL)

• Ground-state electronic structure
• Very high-order coupled cluster (up to CCSDTQPH)
• Automated coupled cluster equation generation
• String-based many-body theory
• Geometry optimization with analytic gradients
• Excited states via EOM-CC at various orders
• Local correlation for large molecules (LNO-CCSD(T))
• Explicitly correlated F12 methods
• Multi-reference CC calculations
• Molecular properties
• Benchmark-quality calculations
• Interface to external programs for integrals
• Efficient parallelization (OpenMP + MPI)
• Automated focal-point analysis
• Composite thermochemistry protocols

Sources: Official MRCC documentation, cited in 6/7 source lists

Key Strengths

Arbitrary-Order CC:

• Automated equation generation
• String-based many-body theory
• CC up to CCSDTQPH (sextuple excitations)
• CI up to full CI
• Systematic hierarchy

Local Correlation (LNO):

• Local Natural Orbital framework
• LNO-CCSD(T) for large molecules
• Near-linear scaling
• Analytical gradients
• Production quality

Benchmark Accuracy:

• Highest-order correlation methods
• Full CI limit approachable
• Thermochemical protocols
• Reference calculations
• Method validation

Automated Methods:

• Focal-point analysis
• Composite methods
• Basis set extrapolation
• Automated workflows
• Error estimation

Inputs & Outputs

• Input formats:

  • MRCC input file (MINP)
  • XYZ coordinate files
  • Interface inputs from CFOUR, Molpro, ORCA

• Output data types:

  • Detailed output files
  • Energies, gradients
  • Correlation energies by order
  • Property calculations
  • Wavefunction analysis

Interfaces & Ecosystem

• Integral interfaces:

  • Built-in integral code (default)
  • CFOUR for advanced integrals
  • Molpro interface
  • ORCA interface (>v5.0)
  • Dirac interface for relativistic integrals
  • PSI4 interface

• Standalone capabilities:

  • Complete standalone operation
  • Built-in SCF, DFT
  • Built-in basis sets

• Utilities:

  • dmrcc - main driver
  • Automated protocol execution
  • Focal-point automation

Workflow and Usage

Input Format (MINP):

MRCC uses a keyword-based MINP input file.

calc=CCSD
basis=cc-pVDZ
mem=1000MB

geom=xyz
3

O 0.0 0.0 0.0
H 0.0 0.7 0.0
H 0.7 0.0 0.0

Running MRCC:

dmrcc > minp.out

Common Tasks:

• Single Point: calc=CCSD(T)
• Optimization: geom=opt
• LNO Method: localcc=on
• F12: densfit=on and F12 keywords

Advanced Features

Arbitrary-Order CC:

• calc=CC(n) allows easy access to high orders (e.g., calc=CCSDT)
• Automated derivation of equations
• Validated against FCI for small systems
• Up to sextuple excitations (or more)

LNO-CCSD(T):

• Linear Scaling Local Natural Orbital CCSD(T)
• "Black-box" accuracy control (Tight/Normal/Loose)
• Applicable to systems with 100+ atoms
• Analytical gradients for geometry optimization

F12 Methods:

• Compatible with arbitrary order CC
• Drastically reduces basis set error
• Approach FCI limit with manageable basis sets

Relativistic Effects:

• calc=X2C or calc=DKH2
• Essential for heavy element accuracy
• Compatible with LNO methods

Automated Protocols:

• calc=HEAT for thermochemistry
• calc=W4
• Automated basis set extrapolation
• Composite energy schemes

Performance Characteristics

• Accuracy: The "Reference" code for high-order coupled cluster
• Speed: Generally slower than specialized low-order codes (PSI4/Molpro/ORCA) for standard CCSD(T), but unique for high orders
• Scalability: Hybrid MPI/OpenMP parallelization; LNO methods scale linearly
• Memory: Arbitrary order methods scale factorially in memory/disk
• Disk I/O: Very heavy for high-order methods

Computational Cost

• CCSD(T): O(N^7), standard
• CCSDT: O(N^8)
• CCSDTQ: O(N^10)
• LNO-CCSD(T): Linear scaling, O(N), breaks even ~30-50 atoms
• Arbitrary Order: Grows extremely fast, limited to <10 atoms for very high orders

Comparison with Other Codes

• vs CFOUR: CFOUR better for analytical derivatives/properties; MRCC better for higher Order energies and local correlation.
• vs Molpro: Molpro is "gold standard" for MRCI; MRCC is "gold standard" for high-order CC.
• vs ORCA: ORCA's DLPNO is similar to MRCC's LNO; MRCC offers higher order canonical CC benchmarks.
• Unique strength: Arbitrary-order Coupled Cluster (CCSDT, CCSDTQ, ...), LNO-CCSD(T) for large molecules.

Best Practices

High-Order Calculations:

• Use small basis sets first to test feasibility
• Estimate memory requirements carefully
• Use restart options for long jobs

LNO-CCSD(T):

• Use lc_ortho=tight for benchmark accuracy
• Check domain sizes
• Use density fitting (densfit=on) for speed

Methods:

• Use calc=CCSD(T) for standard chemical accuracy
• Use calc=CCSDT only for benchmarking small systems
• Use F12 basis sets with F12 methods

Community and Support

• Support: Active email support from developers
• Manual: Detailed description of keywords
• Development: Kállay group (Budapest)
• License: Academic (free/low cost) and Commercial

Application Areas

Benchmark Calculations:

• Reference energies
• Method validation
• Convergence studies
• Correlation hierarchy
• FCI extrapolation

High-Accuracy Thermochemistry:

• Atomization energies
• Reaction barriers
• Heats of formation
• Isomerization energies
• Sub-kJ/mol accuracy

Large Molecule Applications:

• LNO-CCSD(T) for 50-100 atoms
• Organic molecules
• Biomolecule fragments
• Drug-like molecules
• Noncovalent interactions

Method Development:

• High-order CC research
• Perturbation theory studies
• Local correlation development
• Multi-reference theory

Limitations & Known Constraints

• Registration required: Free for academics but requires registration
• Molecular focus: Not designed for periodic systems
• System size: High-order CC limited to very small molecules (<10 atoms)
• Memory: Very high-level methods extremely memory-intensive
• Computational cost: High-order CC scales steeply (factorial-like)
• Basis sets: Gaussian-type; large basis required for accuracy
• Learning curve: Steep; requires expert knowledge
• Documentation: Good but assumes high-level theory knowledge
• Parallelization: Efficient OpenMP+MPI hybrid
• Platform: Linux, macOS, Windows

Verification & Sources

Primary sources:

1. Official website: https://www.mrcc.hu/
2. Manual: https://www.mrcc.hu/manual/
3. M. Kállay et al., J. Chem. Phys. 152, 074107 (2020) - MRCC program
4. M. Kállay and P. R. Surján, J. Chem. Phys. 115, 2945 (2001) - Arbitrary-order CC
5. P. R. Nagy, M. Kállay, J. Chem. Phys. 150, 104101 (2019) - LNO-CCSD(T)

Secondary sources:

1. MRCC manual and examples
2. Published benchmark calculations
3. High-accuracy thermochemistry studies
4. Confirmed in 6/7 source lists (claude, g, gr, k, m, q)

Confidence: CONFIRMED - Appears in 6 of 7 independent source lists

Verification status: ✅ VERIFIED

• Official homepage: ACCESSIBLE
• Documentation: ACCESSIBLE
• Software: Free for academics (registration required)
• Community support: Active (email support)
• Academic citations: >1,000
• Unique capability: Automated arbitrary-order CC equations, highest-order correlation, LNO local correlation
• Specialized strength: Arbitrary-order CC (up to CCSDTQPH), string-based many-body theory, local correlation (LNO-CCSD(T)), benchmark thermochemistry, automated protocols