Official Resources
- Homepage: https://computing.llnl.gov/projects/qball
- Documentation: https://github.com/LLNL/qball/blob/master/README.md
- Source Repository: https://github.com/LLNL/qball
- License: GNU General Public License v3.0 (LLNL-CODE-635376)
Overview
Qb@ll (also written as Qball or qb@ll) is a first-principles molecular dynamics code developed at Lawrence Livermore National Laboratory. It computes electronic structure of atoms, molecules, solids, and liquids using Density Functional Theory with a plane-wave basis. Qb@ll is a fork of the Qbox code by François Gygi, optimized for high-performance computing including Real-Time TDDFT capabilities.
Scientific domain: Ab initio molecular dynamics, electronic structure, real-time electron dynamics
Target user community: HPC users at national labs and institutions needing scalable plane-wave DFT/TDDFT
Theoretical Methods
- Density Functional Theory (DFT)
- Plane-wave basis set
- Norm-conserving pseudopotentials
- DFT-GGA and Hybrid DFT functionals
- Born-Oppenheimer molecular dynamics
- Car-Parrinello molecular dynamics
- Real-Time TDDFT (RT-TDDFT) - developed in Qb@ch branch
- NVT simulations with stochastic thermostats
Capabilities
- Ground-state DFT for all system types
- First-principles molecular dynamics (FPMD)
- Born-Oppenheimer MD with forces from DFT
- Car-Parrinello MD for efficient dynamics
- Real-Time TDDFT (via Qb@ch development)
- Hybrid functionals support
- Large-scale parallel execution
- HPC optimized (designed for supercomputers)
Key Strengths
HPC Performance:
- Designed for leadership-class supercomputers
- Excellent parallel scaling
- Blue Gene/Q optimized configurations
- MPI + OpenMP hybrid parallelization
Qbox Foundation:
- Fork of established Qbox code
- Plane-wave accuracy
- Proven algorithms
- Active LLNL support
RT-TDDFT Development:
- Qb@ch branch (UNC Chapel Hill) focuses on RT-TDDFT
- Electron dynamics capabilities
- Continued development beyond Qbox
Molecular Dynamics:
- Both BO-MD and CP-MD
- Various thermostats
- Production-quality simulations
Inputs & Outputs
-
Input formats:
- Qball input files (.i)
- Coordinate files (.sys)
- Pseudopotential files (.xml)
- Example inputs in examples/ directory
-
Output data types:
- Energies and forces
- MD trajectories
- Electronic structure data
- Wavefunction outputs
Interfaces & Ecosystem
- Build system: GNU Autotools (autoconf, automake)
- Dependencies: BLAS, LAPACK, ScaLAPACK, FFTW
- Parallelization: MPI + OpenMP
- Platforms: Linux HPC clusters, Blue Gene systems
Advanced Features
RT-TDDFT (Qb@ch):
- Real-time propagation
- Strong-field dynamics
- Continued development at UNC
HPC Optimization:
- Vendor-specific optimizations (IBM ESSL)
- Custom parallel strategies
- Designed for 10,000+ cores
Performance Characteristics
- Speed: Highly optimized for large core counts
- Scaling: Excellent to thousands of MPI ranks
- Accuracy: Plane-wave systematic convergence
- Memory: Distributed memory model
Computational Cost
- Ground state: Plane-wave standard scaling
- MD: Efficient for large trajectories
- RT-TDDFT: Time-step dependent
- Parallelization: Near-linear scaling on HPC
Limitations & Known Constraints
- Compilation: Requires HPC environment and libraries
- Learning curve: HPC expertise helpful
- Pseudopotentials: Requires compatible formats
- Documentation: Technical focus, less beginner-friendly
- INQ successor: LLNL now developing GPU-focused INQ code
Comparison with Other Codes
- vs Qbox: Qb@ll fork with LLNL optimizations and Qb@ch RT-TDDFT
- vs VASP: Both plane-wave; Qb@ll open-source, HPC optimized
- vs QE: Qb@ll smaller codebase, HPC focus; QE broader features
- vs SALMON: Both RT-TDDFT capable; Qb@ll also strong in MD
- Unique strength: HPC optimization, combined MD + RT-TDDFT development
Application Areas
- High-pressure physics
- Warm dense matter
- Extreme conditions simulations
- Materials under shock
- Large-scale molecular dynamics
- Ultrafast dynamics (via Qb@ch)
Best Practices
- Use example input files as templates
- Configure for target HPC architecture
- Consult LLNL documentation for optimization
- Contact maintainers for support
Community and Support
- Open-source GPL v3
- LLNL development team
- GitHub repository with 8 contributors
- Contact: Erik Draeger, Xavier Andrade (LLNL)
- Academic development at UNC (Qb@ch)
Verification & Sources
Primary sources:
- GitHub repository: https://github.com/LLNL/qball
- LLNL Computing: https://computing.llnl.gov/projects/qball
- Original Qbox: http://qboxcode.org/
Secondary sources:
- Qb@ch development at UNC Chapel Hill
- LLNL publications using Qb@ll
Confidence: VERIFIED
- Repository: ACCESSIBLE (GitHub, LLNL)
- License: GPL v3 (LLNL-CODE-635376)
- Active: 8 contributors, ongoing development
Verification status: ✅ VERIFIED