A more detailed overview of WASM: “In this article, I start off by reviewing the WebAssembly developments during 2023 around Garbage Collection, Tail Calls, fixed-width SIMD, multiple memories, improvements in .NET, and work happening with the WebAssembly System Interface (WASI) and the Component Model. Then I try to predict where I think things will go in 2024.”
@KC1PYT@Perl@shiar I don’t have any direct #Wasm experience so I can't confirm or deny without doing the same research as anyone else. Maybe somebody receiving these group messages can opine.
The optimization work @katei is doing on WasmKit is nothing short of amazing 🤩 AFAIU this #Wasm interpreter written in Swift is ~15x faster in common use cases, when compared to the state it was before Yuta started working on it, and now it's 100% compliant with the WebAssembly 1.0 spec and a few major accepted proposals.
Teaches a lot about the benefits of knowing underlying hardware well. Sometimes worth taking into account cache locality and register layout after all https://github.com/swiftwasm/WasmKit/pull/70
"Creating a WebAssembly component with WAT" is a blog post that looks at the #Wasm component model and its Canonical ABI from the perspective of hand-written WAT:
WASI Preview 2 is officially out, and it's a big deal. Beyond the APIs that open WASM to a growing number of use cases and environments, the component model allows assembling interoperable modules developed in different languages. #wasm#wasihttps://blog.sunfishcode.online/wasi-preview2/
I made a Dungeon Crawler Roguelike that you can play in your web browser1.
It isn't original though: as part of learning the Rust programming language, I followed the tutorial project in Hands-on Rust by @herberticus . I posted a recap on my blog2.
I'm very much looking forward to Herbert Wolverson's next book!
The WASI Subgroup has now voted, and WASI Preview 2 is now officially launched! A lot of people have contributed to making this possible.
I wrote up a blog post that looks at what this means in the present, looks back at some of the things that shaped this moment, and look forward to what's coming:
(tl;dr: wasmtime is fast, our kernel uses wasm time’s linker, & as a result, we can transfer about 1.5GiB/s across the host/guest boundary — at about 4.7ns per 8 byte-function call on my M2 mac.)
I'd like to show my team the insane power of WASM. Many of them are Python users. Is there an equivalent of this webR demo that hits just as hard for a Python user? https://webr.r-wasm.org/latest/
📣 New blog post: Six not-so-basic base R functions 🔥
There are so many great functions in base R. Let's explore six lesser-known ones in our latest post. Run them in your browser using the magic of webR and Quarto 💫