GENESIS

Official Resources

• Homepage: https://www.r-ccs.riken.jp/labs/cbrt/
• Documentation: https://www.r-ccs.riken.jp/labs/cbrt/genesis/
• Source Repository: https://github.com/genesis-release-r-ccs/genesis
• License: LGPL

Overview

GENESIS (GENeralized-Ensemble SImulation System) is a high-performance molecular dynamics software developed at RIKEN, Japan. It is designed for large-scale biomolecular simulations on supercomputers, featuring excellent parallel scaling and support for enhanced sampling methods.

Scientific domain: Large-scale biomolecular simulations, supercomputing
Target user community: Researchers running large-scale MD on supercomputers

Theoretical Methods

• Classical molecular dynamics
• Replica exchange methods (REMD, gREST)
• Gaussian accelerated MD (GaMD)
• String method
• Multiple force fields (AMBER, CHARMM, GROMOS)
• Coarse-grained models

Capabilities (CRITICAL)

• Massively parallel MD
• Replica exchange MD
• Gaussian accelerated MD
• String method for pathways
• Coarse-grained simulations
• Cellular-scale simulations
• GPU acceleration

Key Strengths

Parallel Performance:

• Excellent weak scaling
• Fugaku supercomputer optimized
• Millions of atoms
• Hybrid MPI/OpenMP

Enhanced Sampling:

• Multiple REMD variants
• GaMD implementation
• String method
• gREST method

Inputs & Outputs

• Input formats:

  • PDB structures
  • AMBER prmtop
  • CHARMM PSF
  • GROMACS top

• Output data types:

  • DCD trajectories
  • Restart files
  • Energy logs

Interfaces & Ecosystem

• CHARMM-GUI: System setup
• AMBER/CHARMM: Force fields
• Fugaku: Optimized for

Advanced Features

• gREST: Generalized REST
• GaMD: Gaussian accelerated MD
• REUS: Replica exchange umbrella sampling
• String method: Reaction pathways
• Cellular MD: Large-scale simulations
• Hybrid parallel: MPI + OpenMP + GPU

Performance Characteristics

• Excellent parallel scaling
• Optimized for ARM (Fugaku)
• GPU support
• Efficient for large systems

Computational Cost

• Scales to 100,000+ cores
• Efficient for large biomolecules
• GPU acceleration available
• Overall: Excellent for supercomputers

Best Practices

• Use hybrid parallelization
• Choose appropriate REMD method
• Validate with smaller systems first
• Use CHARMM-GUI for setup

Limitations & Known Constraints

• Supercomputer focus
• Less desktop-friendly
• Steeper learning curve
• Japanese documentation primary

Application Areas

• Large biomolecular systems
• Virus capsids
• Membrane proteins
• Cellular-scale simulations
• Drug discovery

Comparison with Other Codes

• vs GROMACS/AMBER: GENESIS optimized for supercomputers, especially Fugaku
• vs NAMD: Both scale well, GENESIS better ARM/Fugaku optimization
• vs OpenMM: GENESIS multi-node focus, OpenMM single-node GPU
• Unique strength: Fugaku optimization, gREST/GaMD enhanced sampling, cellular-scale simulations

Community and Support

• RIKEN development team
• Documentation
• Workshops (Japan)
• GitHub issues

Verification & Sources

Primary sources:

1. Website: https://www.r-ccs.riken.jp/labs/cbrt/
2. J. Jung et al., J. Phys. Chem. B 128, 5028 (2024)
3. J. Jung et al., WIREs Comput. Mol. Sci. 5, 310 (2015)

Secondary sources:

1. GENESIS tutorials and workshops
2. RIKEN documentation
3. Fugaku benchmark publications

Confidence: VERIFIED

Verification status: ✅ VERIFIED

• Source code: OPEN (GitHub, LGPL)
• Academic citations: >500
• Active development: RIKEN team
• HPC validated: Fugaku supercomputer