Cramming a 400B Model into 48GB: The Magic Behind LLM in a Flash
An Apple paper from 2023 made it possible to run a 400 billion parameter model on an ordinary MacBook. The …
#Apple Silicon
90 Seconds of Waiting, Gone: How oMLX Buries Ollama on Mac
oMLX is built for Apple Silicon, using the MLX framework, SSD-backed KV cache, and continuous batching to cut …
Rust 1.89: Intel Macs demoted to Tier 2, clearer lifetimes, and new x86 intrinsics
Rust 1.89 demotes x86_64-apple-darwin (Intel Mac) from Tier 1 to Tier 2, adds the mismatched_lifetime_syntaxes …
A Survival Guide to Zig Performance Profiling on Apple Silicon (M1/M2)
Performance profiling for Zig on Apple Silicon (M1/M2) can be challenging as common tools like perf and …