MATILDA.FT: A mesoscale simulation package for inhomogeneous soft matter.

Area of Science:

• Computational Physics
• Materials Science
• Software Engineering

Background:

• Mesoscopic simulations are crucial for understanding complex materials.
• Existing simulation packages often lack the ability to integrate diverse simulation methodologies.
• Efficient, massively parallel computation is needed for large-scale mesoscopic modeling.

Purpose of the Study:

• To introduce MATILDA.FT, a new, massively parallel, GPU-accelerated software package.
• To present a unified simulation package combining coarse-grained particle dynamics and field-theoretic methods.
• To provide an accessible, extendable, and efficient tool for mesoscopic simulations.

Main Methods:

• Development of a CUDA/C++ object-oriented software leveraging Thrust library acceleration.
• Implementation of massively parallel algorithms for efficient computation on GPUs.
• Integration of both coarse-grained particle dynamics (Langevin, Dissipative particle dynamics) and field-theoretic simulations.

Main Results:

• MATILDA.FT is the first software to combine coarse-grained particle and field-theoretic simulations.
• The software demonstrates efficient simulation of various mesoscopic systems, including polymer solutions, nanoparticle-polymer interfaces, peptide models, and liquid crystals.
• The object-oriented design facilitates code understanding and extension.

Conclusions:

• MATILDA.FT offers a powerful and versatile platform for mesoscopic simulations.
• The software enables efficient harnessing of GPU parallelism for complex material modeling.
• Public release of source code, documentation, and examples promotes accessibility and further development.