#compiler

AI Hacker News Mar 4, 2026 2 min read

Show HN: Timber Compiles Classical ML Models into Tiny C Binaries for Microsecond Inference

A Show HN project called Timber claims it can compile tree-based ML models into dependency-free C99 artifacts, with reported ~2 microsecond latency and up to 336x speedup over Python baselines.

#classical-ml #xgboost #inference

LLM Hacker News Feb 18, 2026 1 min read

BarraCUDA Draws HN Attention: A C99 CUDA Compiler That Emits AMD GFX11 Binaries Without LLVM

A high-scoring Hacker News post highlighted BarraCUDA, an open-source C99 compiler that translates CUDA `.cu` code directly into AMD GFX11 `.hsaco` binaries with no LLVM dependency.

#cuda #amd-gpu #compiler