Devito: Symbolic Finite Difference Computation
Devito is a Python package to implement optimized stencil computation (e.g., finite differences, image processing, machine learning) from high-level symbolic problem definitions. Devito builds on SymPy and employs automated code generation and just-in-time compilation to execute optimized computational kernels on several computer platforms, including CPUs, GPUs, and clusters thereof.
Symbolic computation is a powerful tool that allows users to:
- Build complex solvers from only a few lines of high-level code
- Use automated performance optimisation for generated code
- Adjust stencil discretisation at runtime as required
- (Re-)development of solver code in hours/days rather than months
Seismic Inversion using Devito
Devito is primarily designed to create wave propagation kernels for
use in seismic inversion problems. A tutorial for the generation of a
modelling operator using an acoustic wave equation can be found
here and a paper
outlining the verification procedures of the acoustic operator can be
found here.
Key Features
- A functional language to express finite difference operators.
- Straightforward mechanisms to adjust the discretization.
- Constructs to express sparse operators (e.g., interpolation), classic linear operators (e.g., convolutions), and tensor contractions.
- Seamless support for boundary conditions and adjoint operators.
- A flexible API to define custom stencils, sub-domains, sub-sampling, and staggered grids.
- Generation of highly optimized parallel code (SIMD vectorization, CPU and GPU parallelism via OpenMP and OpenACC, multi-node parallelism via MPI, blocking, aggressive symbolic transformations for FLOP reduction, etc.).
- Distributed NumPy arrays over multi-node (MPI) domain decompositions.
- Inspection and customization of the generated code.
- Autotuning framework to ease performance tuning.
- Smooth integration with popular Python packages such as NumPy, SymPy, Dask, and SciPy, as well as machine learning frameworks such as TensorFlow and PyTorch.