Official Resources
- Homepage: https://github.com/ghb24/NECI_STABLE
- Documentation: GitHub repository and manual
- Source Repository: https://github.com/ghb24/NECI_STABLE
- License: GNU General Public License v3.0
Overview
NECI is a state-of-the-art implementation of Full Configuration Interaction Quantum Monte Carlo (FCIQMC), a stochastic method for solving the electronic Schrödinger equation in a systematically improvable way. Developed primarily at the University of Cambridge, NECI provides numerically exact solutions to the many-electron problem by stochastically sampling the full CI space. The method bridges quantum chemistry and QMC, offering chemical accuracy for strongly correlated systems.
Scientific domain: FCIQMC, quantum chemistry, strongly correlated electrons
Target user community: Quantum chemists, strongly correlated systems, method developers
Theoretical Methods
- Full Configuration Interaction QMC (FCIQMC)
- Initiator approximation
- Spawning algorithm
- Stochastic CI sampling
- Coupled Cluster Monte Carlo (CCMC)
- Density Matrix QMC (DMQMC)
- Finite-temperature extensions
- Exact (in CI space limit)
Capabilities (CRITICAL)
Category: Open-source FCIQMC code
- FCIQMC method
- Initiator FCIQMC
- CCMC variants
- DMQMC (finite-T)
- Molecules and lattices
- Strongly correlated systems
- Exact in complete space
- Systematically improvable
- Chemical accuracy
- Ground and excited states
- MPI parallelization
- Production quality
Sources: GitHub repository, Cambridge group publications
Key Strengths
FCIQMC Method:
- Stochastic full CI
- Numerically exact (converged)
- Strong correlation capable
- Systematically improvable
- Sign coherence
Initiator Approximation:
- Controlled approximation
- Larger systems
- Sign problem mitigation
- Systematic convergence
- Production viable
Versatility:
- FCIQMC, CCMC, DMQMC variants
- Molecules and lattices
- Multiple methods
- Research flexibility
- Method development
Cambridge Development:
- Leading research group
- Method innovation
- Active development
- Strong theoretical foundation
- Publication quality
Inputs & Outputs
-
Input formats:
- NECI input files
- Integrals from quantum chemistry codes
- FCIDUMP format
- Model Hamiltonians
-
Output data types:
- Ground state energies
- Excited states
- Correlation energies
- Observables
- Sampling statistics
Interfaces & Ecosystem
Quantum Chemistry:
- FCIDUMP integrals
- Molpro
- PySCF
- GAMESS
- Standard formats
Lattice Models:
- Hubbard model
- Custom Hamiltonians
- Research applications
Workflow and Usage
Installation:
# Clone repository
git clone https://github.com/ghb24/NECI_STABLE.git
cd NECI_STABLE
mkdir build && cd build
cmake ..
make -j8
Input File (FCIQMC.inp):
FCIQMC
METHODS
method vertex
ENDINIT
CALC
electrons 10
spin-restrict
totalwalkers 1e6
startsinglepart 100
ENDCALC
LOGGING
popsfile-format HDF5
ENDLOG
END
Run FCIQMC:
# MPI parallel
mpirun -n 16 neci FCIQMC.inp > output.out
Analysis:
# Extract energy
grep "Shift" output.out | tail -100
Advanced Features
Initiator Approximation:
- i-FCIQMC
- Controlled bias
- Sign coherence
- Larger systems
- Systematic convergence
CCMC:
- Coupled Cluster Monte Carlo
- Truncated CC
- Alternative approach
- Specific advantages
DMQMC:
- Density Matrix QMC
- Finite temperature
- Thermal properties
- Imaginary-time evolution
Excited States:
- Multiple states
- State-specific
- Systematic approach
- Excitation energies
Performance Characteristics
- Speed: MPI-parallel
- Accuracy: Numerically exact (converged)
- System size: Moderate (CI space limited)
- Purpose: Strongly correlated, benchmarks
- Typical: HPC calculations
Computational Cost
- Walker-number dependent
- CI space scaling
- Expensive for large systems
- Exact results justify cost
- HPC recommended
Limitations & Known Constraints
- CI space: Limited by basis
- Computational cost: Expensive
- Sign problem: Fermion sign issue
- System size: Moderate
- Learning curve: FCIQMC expertise
- HPC required: Production calculations
Comparison with Other Methods
- vs Traditional FCI: NECI stochastic, tractable for larger
- vs DMRG: NECI different approach, complementary
- vs Coupled Cluster: NECI exact, CC approximate
- Unique strength: FCIQMC method, systematically exact, strong correlation, Cambridge development
Application Areas
Strongly Correlated:
- Molecules
- Transition metals
- f-electron systems
- Strong correlation
- Multi-reference
Benchmark Calculations:
- Exact results
- Method validation
- Chemical accuracy
- Reference energies
- Correlation energies
Quantum Chemistry:
- Ground states
- Excited states
- Reaction energies
- Spectroscopy
- Electronic structure
Method Development:
- FCIQMC research
- Algorithm development
- Stochastic methods
- QMC innovations
Best Practices
Walker Number:
- Sufficient population
- Convergence testing
- Statistical analysis
- Plateau region
- Error estimation
Initiator:
- Appropriate threshold
- Systematic convergence
- Balance accuracy/cost
- Production settings
Basis Sets:
- Quality basis
- Convergence studies
- CI space considerations
- System-appropriate
Community and Support
- Open-source (GPL v3)
- University of Cambridge
- GitHub repository
- Active development
- Research community
- Publications
- Method innovation
Educational Resources
- GitHub documentation
- FCIQMC papers
- User manual
- Example inputs
- Cambridge group publications
- QMC schools
Development
- University of Cambridge
- Alavi group
- Active research
- Method development
- Community contributions
- Regular updates
Research Impact
NECI and FCIQMC have revolutionized exact solutions for strongly correlated molecules, enabling benchmark-quality calculations that were previously impossible, with hundreds of publications and major impact in quantum chemistry.
Verification & Sources
Primary sources:
- GitHub: https://github.com/ghb24/NECI_STABLE
- Cambridge group
- Publications: J. Chem. Phys. 132, 041103 (2010)
Secondary sources:
- FCIQMC literature
- Quantum chemistry papers
- User publications
Confidence: VERIFIED - Leading FCIQMC code
Verification status: ✅ VERIFIED
- GitHub: ACCESSIBLE
- License: GPL v3 (open-source)
- Category: Open-source FCIQMC code
- Status: Actively developed
- Institution: University of Cambridge
- Specialized strength: Full Configuration Interaction Quantum Monte Carlo, stochastic CI sampling, numerically exact for strongly correlated systems, initiator approximation, CCMC/DMQMC variants, benchmark quality, Cambridge development, chemical accuracy, systematic improvability, production quality