Beeg float library, a Rust port of Fabrice Bellard's libbf

A tiny math tool drops in Rust, and the comments instantly turn into a nerd food fight

TLDR: libbeef is a new Rust-based precision math library that promises serious number-crunching without heavy setup or outside dependencies. Commenters split fast between impressed math nerds and skeptics mocking it as an AI-assisted port, turning a niche launch into a full-on credibility debate

A new Rust project called libbeef just strutted onto the internet with a very online promise: big-number floating-point math, tiny footprint, no extra baggage, and a name that basically dares the comment section to behave. It’s a Rust remake of Fabrice Bellard’s respected libbf library, meant to handle everything from weird edge cases like signed zero to heavy-duty functions like sine, logarithms, and powers. Translation for normal people: it’s a compact math engine for programmers who want very precise calculations without dragging in a pile of system-level extras.

But the real action? The community reaction was half applause, half side-eye. One camp was impressed that the library appears to do careful, standards-correct rounding, with one commenter essentially going from skeptic to “oh wow, nice, learned something today,” while even dropping floating-point reading homework. The other camp immediately reached for the flamethrower: the snarkiest hot take accused the whole thing of being little more than “point AI at old code, ask for Rust, collect praise.” Ouch.

Then the author jumped in with a surprisingly chill vibe, basically saying: yes, it’s my library, no, I didn’t even think it was that big a deal, and then calmly explained the real motivation was seeing whether Rust could keep up on speed. That combo of modesty and confidence only added fuel to the thread’s favorite pastime: debating whether this is a genuinely useful portable math tool or another shiny “Rust port of existing thing” moment. Either way, the comments made one thing clear: when you name it libbeef, people are absolutely showing up hungry

Key Points

  • libbeef is a pure-Rust port of Fabrice Bellard’s libbf for arbitrary-precision floating-point arithmetic with full IEEE 754 semantics.
  • The library includes transcendental functions, decimal floating-point support, no_std compatibility with alloc, and zero external dependencies.
  • Its implementation follows libbf’s algorithmic design, including NTT-based multiplication, Newton iteration for division and square root, and AGM/binary-splitting for transcendental functions.
  • Benchmark data in the article shows libbeef near libbf performance, slower than GMP/MPFR in some cases, and significantly faster than num-bigint for large multiplication workloads.
  • The article highlights portability, smaller binary size than some alternatives, verification against libbf’s test suite, and MIT licensing as reasons to use libbeef.

Hottest takes

"I pointed Claude at another existing codebase and told it to port it to Rust" — i2talics
"faithful rounding of operations. Nice. TIL." — RossBencina
"I didn’t think this is that HN-worthy. Ask me anything anyway." — lifthrasiir
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.