Scientific Computing in Rust Monthly #1
Welcome to the very first issue of Scientific Computing in Rust Monthly!
We're very excited to be starting this newsletter, and hope that read it will become a great way to keep up to date with the latest advances in the use of Rust for scientific computing.
Scientific Computing in Rust Monthly can be viewed online at scientificcomputing.rs/monthly, you can sign up to receive an email copy at mailinglist.scientificcomputing.rs, or you can subscribe using the RSS feed at scientificcomputing.rs/monthly/rss.xml.
Items can be added to the next issue of the newsletter by opening a pull request to the Scientific Computing in Rust Monthly repository on GitHub. You can find more information about adding items here.
Crate of the month: faer
faer is a dense linear algebra library that has implementations of loads of dense linear algebra methods written in pure Rust. If you've not checked it out yet, we recommend watching the recording of Sarah El Kazdadi's talk about faer at Scientific Computing in Rust 2023 and taking a look at the examples in the GitHub repository.
New releases
kifmm-rs v1.0.1
With this release we now offer a highly competitive single node kiFMM, the notable features of this library being:
- Competitive single-node performance, especially in single precision, enabled by the optimisation of BLAS based multipole to local (M2L) field translation, based entirely on level 3 operations with high arithmetic intensity that are well suited to modern hardware architectures that prioritise minimal memory movement per flop.
- The ability to process multiple sets of source densities corresponding to the same point distribution, a common application in the boundary element method.
- A C API, using Rust's C ABI compatibility allowing for the construction of bindings into other languages, with full Python bindings for non-specialist users.
- A highly portable software with support for both x86 and Arm architectures.
Events
Scientific Computing in Rust, July 2024
The second annual Scientific Computing in Rust workshop took place around a month ago. Videos of all the talks are available on the Scientific Computing in Rust YouTube channel.
Publications
M. Bitar, "Rust and Julia for Scientific Computing" in Computing in Science & Engineering, vol. 26
Rust and Julia are two modern languages aiming to advance computing in their unique way. Julia is a dynamic language striving to bridge the performance gap between dynamic languages like Python and statically typed ones like C/C++ while improving the ergonomics for scientific computing. Rust, on the other hand, is a statically typed language with a friendly compiler guiding developers to write reliable, correct, and concurrent code while ensuring memory safety and unlocking the highest performance. Both languages seem to be appealing to the scientific domain, especially because of their focus on performance. But choosing one depends on the type of your project.
S. Kailasa, "kifmm-rs: A Kernel-Independent Fast Multipole Framework in Rust", submitted to Journal of Open Source Software
kifmm-rs is a framework for developing kernel independent fast multipole methods (kiFMM) in Rust. This crucial algorithm accelerates potential evaluation problems to linear complexity in the best case, and finds application across science and engineering. This software demonstrates the capability of Rust as a tool for high-performance scientific computing, while retaining the portability productivity benefits offered by Rust as a platform.
