GAMESS US |
|||
Summary | |||
Contact | http://www.msg.ameslab.gov/gamess/ https://groups.google.com/forum/#!newtopic/gamess |
||
Domain | Biomedical research domain | ||
Description | Quantum Mechanical Fragment Molecular Orbital (FMO) based Fragment Expansion. FMO based Virtual Screen. | ||
Documentation | |||
User’s guide | |||
Scientific articles | |||
Requirements | |||
Computational requirements | Programming language: Fortran Dependencies: MPICH Memory requirements: 3GB per core Disk requirements: Small for most calculations ~150KB. However will require the access to the EvoSource Small molecule library. Is 7Million compounds in SDF format. As single conformers will be used should only require 177GB of data storage space for input. Will potentially be running 21 million calculation using the DFTB functional. Keeping all this data could lead to ~3 terabytes of data if all results are kept. Energy + Geometries could be extracted from results and other from FMO run deleted. Complementary tools: |
||
Type of software licence | Open | ||
Input requirements | Format: Text file GAMESS.inp (including pdb and ligand) Coming from: Generated by KNIME at evotec Disk use: FMO Fragment Screen = 3000 input jobs containing geometries. FMO DFTB virtual screen = 21 million input jobs containing geometries |
||
Output requirements | Format: Text GAMESS.out (including pdb and ligand) Used by: Potential post processing by cclib to extract energies / geometries. Will need to be sorted to show best docked ligands / fragments Disk use: 3000 + ~21 million |
||
Performance and HPC profile | |||
Parallelism | Type of parallelism: MPI sockets or Infiniband Scalability: FMO standard = 64 cores in groups of ~16 or 80 cores in Groups of 20. FMO DFTB = 1 core but potentially Millions of Jobs |
||
Mode of operation | Standard FMO = ~64 cores in groups of ~16 or 80 cores in Groups of 20. FMO DFTB = 1 core but potentially millions of Jobs | ||
Deployed on | Frequently installed centrally on HPC resources (for example SuperMUC, Cartesius, Archer) | ||
Getting started | |||
Benchmarks and examples | |||
Instructions for running on dedicated platforms |