These sources work near the metal. They read file formats, follow memory and tune code to the processor. Most offer measurements or a walk through real code.
A catalog of side effects
A tour of how several optimizing compilers record what each instruction reads and writes.
A tail-call interpreter in (nightly) Rust
A Hacker News thread about a blog post on a Rust tail-call interpreter.
Everyone Says Assembly Is Untyped—Everyone Is Wrong
A design rationale for the typed inline assembly system in the Odin programming language.
How To Make a Fast Dynamic Language Interpreter
A write-up of 21 optimizations that sped up a simple interpreter for the Zef language.
I rewrote PostHog's SQL parser, 70x faster, while barely looking at the code
A Hacker News thread about a PostHog blog post on an AI-written, faster SQL parser.
Strategies for very fast Lexers
A tutorial on techniques used to make the lexer of one hobby language fast.
Two studies in compiler optimisations
A walkthrough of how LLVM optimises two small C++ functions, with source excerpts and exercises.
Bloom Filters
An interactive tutorial that explains how Bloom filters work and how to size them.
DuckDB Internals: Why is DuckDB Fast? (Part 1)
A vendor blog post explaining how DuckDB turns SQL into a plan and stores data.
SQLite performance tuning
A short gist listing SQLite settings for large databases with many readers.
Surprisingly Fast AI-Generated Kernels We Didn’t Mean to Publish (Yet)
A Stanford CRFM post reporting fast GPU kernels written by language models.
8× faster binary search: from compiled code to mechanical sympathy
A post showing how to speed up binary search by working with the CPU.
A Friendly Tour of Process Memory on Linux
A guided tour of how Linux gives a program memory.
C Is Not a Low-level Language
An essay arguing that C no longer matches how modern processors work.
Dataflow Architectures
A 1986 survey of how dataflow computers store tokens and data structures.
Exploring Mach-O, Part 1
Part one of a series that builds a tiny ARM64 program and reads its header.
Girls Just Wanna Have Fast MPMC Queues with bounded waiting
A design write-up of a bounded multi-producer multi-consumer queue in Rust, with benchmarks.
Pure Silicon Demo Coding: No CPU, No Memory, Just 4k Gates
A write-up of two chip demos built for the Tiny Tapeout 8 competition.
The Nitty Gritty of “Hello World” on macOS
A walkthrough of the Mach-O file format using a compiled Hello World program.
The Return of Rigorous Full-System Timing Simulation
A blog post arguing that computer architects should measure full systems with statistically sound sampling.
ZX Spectrum: Experimenting With 1-Bit Sound
A blog post recording attempts to play samples and chords through the ZX Spectrum beeper.