How ordern/conquest-release works
Project type and purpose
• CONQUEST is a high-performance, large-scale DFT (density functional theory) simulation code written in Fortran.
• It is a computational engine (no GUI or web frontend), designed to read atomic configurations & pseudopotentials, perform SCF, diagonalisation, MD, structural relaxations, etc., and output energies, forces, densities.
• It integrates with external libraries (MPI, ScaLAPACK, FFT libraries, LibXC) and tools (ONCVPSP pseudopotential generator, ASE).
File structure
Top-level components:
• src/: the core Fortran modules and main.f90 (the simulation engine).
• pseudo-and-pao/: repository of pseudopotentials and PAO definitions (data store).
• tools/: standalone Fortran utilities (BasisGeneration, PostProcessing).
• testsuite/: regression tests driving input files → CONQUEST binary → compare outputs.
• benchmarks/: sample inputs & outputs for performance testing.
• docs/: user manual and API docs (RST sources).
• .github/, CI workflows: build & test automation.
• Makefiles in src/ and system/: cluster-specific build configurations.
README for additional insights
• Describes capabilities (static, MD, ensembles, linear-scaling).
• External integrations: LibXC, ONCVPSP, ASE.
• Scalability targets (up to millions of atoms, hundreds of thousands of cores).
Elements
• Clear labels: “Build System,” “Core Engine,” “Pseudopotential DB,” “MPI/ScaLAPACK,” “CLI/ASE,” “Tests & Benchmarks,” “Utilities,” “Docs”
• Rectangles for code components (src/, tools/)
• Clouds for external dependencies (MPI, LibXC, FFT libs)
• Parallelograms for I/O (Conquest_input, output files)