FDBB: Fluid Dynamics Building Blocks
Matthias M\"oller, Andrzej Jaeschke

TL;DR
FDBB is an open-source framework that simplifies development of fluid dynamics applications on heterogeneous high-performance computing systems by providing unified APIs and customizable building blocks.
Contribution
It introduces a unified low-level API and an extendible high-level library for fluid dynamics, facilitating development on diverse hardware architectures.
Findings
Performance is validated through micro-benchmarks.
Mini-applications demonstrate practical effectiveness.
Framework supports various linear algebra back-ends.
Abstract
High-performance computing platforms are becoming more and more heterogeneous, which makes it very difficult for researchers and scientific software developers to keep up with the rapid changes on the hardware market. In this paper, the open-source project FDBB (Fluid Dynamics Building Blocks) is presented, which eases the development of fluid dynamics applications for heterogeneous systems. It consists of a low-level API that provides a unified interface to many different linear algebra back-ends and a lightweight and extendible high-level expression template library, which provides largely customizable fluid dynamics building blocks, like transformations between primary and secondary variables as well as expressions for Riemann invariants, equations of state, inviscid fluxes and their flux-Jacobians. The performance of the developed approach is assessed both for synthetic…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsModeling and Simulation Systems · Parallel Computing and Optimization Techniques · Computational Physics and Python Applications
