OSLO — Unified ODE Solver Toolbox

OSLO (ODE SoLver toolbOx) is a high-performance, Fortran-based toolbox for solving systems of ordinary differential equations (ODEs). It provides a unified and modular interface to multiple state-of-the-art ODE solvers, allowing users to switch between stiff and non-stiff integrators with minimal code changes.

OSLO is designed for scientific computing and HPC environments, bundling well-established solvers libraries under a single, consistent API.

---

Key Features

• Unified Solver Interface Switch between different ODE solvers using a common Fortran API.

• Broad Solver Coverage

  • Hairer & Wanner solvers for stiff and non-stiff problems
  • FATODE with automatic differentiation
  • Intel ODE (Linux only, Intel compiler)
  • SUNDIALS (CVODE, IDA, etc.)

• Parallel Execution

  • Shared-memory parallelism via OpenMP
  • Distributed-memory parallelism via MPI

• Compiler Flexibility

  • GNU Fortran (gfortran)
  • Intel Fortran (ifort / ifx)

• Robust Linear Algebra

  • LAPACK integration for dense and sparse systems

• Modular Architecture

  • Easy to extend or embed in existing Fortran codes

---

Requirements

Mandatory

• CMake ≥ 3.23
• Fortran compiler: GNU or Intel
• LAPACK
• C compiler (required for SUNDIALS)

Optional

• OpenMP — shared-memory parallelism
• MPI — distributed-memory parallelism

---

Platform Support

OSLO has been tested on the following configurations:

OS	Compiler	Status
macOS	GNU (gfortran)	✅
Ubuntu Linux	GNU	✅
Ubuntu Linux	Intel	✅
OpenSUSE Linux	GNU	✅
OpenSUSE Linux	Intel	✅

The toolbox is expected to work on most Unix-like systems with a compatible Fortran compiler.

---

Installation

Quick Start

git clone https://github.com/MarcoGrossi92/OSLO.git
cd OSLO
./install.sh build --compiler=gnu

Build Variants

./install.sh [GLOBAL_OPTIONS] build [OPTIONS]

Global options

• -v, --verbose — verbose output

Build options

• --compiler=gnu|intel
• --use-openmp
• --use-mpi
• --use-sundials (enabled by default)

Examples

# GNU compiler + OpenMP
./install.sh build --compiler=gnu --use-openmp

# Intel compiler + MPI
./install.sh build --compiler=intel --use-mpi

---

Manual CMake Build

mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=Release \
      -DUSE_OPENMP=ON \
      -DUSE_MPI=OFF \
      -DUSE_SUNDIALS=ON \
      ..
make
make install

---

Usage

OSLO builds a Fortran library that can be linked into your own applications.

Minimal Example

use OSLO
implicit none

integer, parameter :: neq = 3
real(8) :: y(neq), t, tout
real(8) :: rtol(neq), atol(neq)
integer :: err

call setup_odesolver(N=neq, solver='dvodef90', RT=rtol, AT=atol)
call run_odesolver(neq, t, tout, y, Fgeneral, err)

Linking Against OSLO

gfortran -c your_program.f90 -Ibuild/modules
gfortran your_program.o -Lbuild/lib -lOSLO -o your_program

---

Third-Party Software

Bundled

• Hairer & Wanner Solvers
  © University of Geneva — BSD-style license
  https://www.unige.ch/~hairer/software.html

• FATODE
  Fast and Accurate ODE solver with AD — GNUv1.3 license
  https://people.cs.vt.edu/~asandu/Software/FATODE/

Optional

• SUNDIALS
  © Lawrence Livermore National Laboratory — BSD license
  https://computing.llnl.gov/projects/sundials

• Intel ODE
  Intel-optimized ODE solvers (Linux only)

---

Project Structure

OSLO/
├── src/            # Core library and solvers
├── lib/            # Third-party solver sources
├── bin/            # Executables and tests
├── cmake/          # CMake modules
├── scripts/        # Helper scripts
├── install.sh      # Build helper
└── CMakeLists.txt

---

Contributing

Contributions are welcome — bug fixes, solver additions, documentation, or performance improvements.

Typical workflow:

1. Fork the repo
2. Create a feature branch
3. Commit changes
4. Open a pull request

---

Support

Please use the GitHub issue tracker for:

• bug reports
• feature requests
• solver questions

👉 https://github.com/MarcoGrossi92/OSLO/issues

---