seqlock

Adaptive Radix Tree in C++

C++ Concurrency Patterns

A blazingly fast™ single producer multiple consumer broadcast queue

High-performance AutoHotkey v2 library suite for real-time pipelines (Lock-Free IPC, SPSC, EMA, Telemetry, Watchdog, Raw Input...) with selftests and Gold stacks.

Ultra-low-latency SPMC/MPMC pub/sub using stamped ring buffers. Formally sound with atomic-slots feature. no_std, zero-alloc hot path. Apache-2.0.

A C++ implementation of the Sequence Lock

Header-only C++20 single-writer seqlock with consumer-side cacheline prefetch chains and writer-side CLDEMOTE, for sub-microsecond publish/consume paths such as market data quotes. CMake package, TSan-tested.

An allocation-free, ultra-low-latency C++20 concurrent cache for multi-platform User-Mode and the Windows Kernel (WDM/KMDF). Utilizes SeqLocks and NUMA sharding to deliver wait-free, nanosecond reads.

A SeqLock (Sequence Lock) is a Reader-Writer Synchronisation Mechanism Used in Linux Kernel Development and Concurrent Programming

Experimental wait/lock-free primitives tuned for ultra-low-latency systems

Explore the intricate world of Kernel Sequential Locks through this repository, which provides a comprehensive understanding of seqlocks, their utility, working, and applications in Linux Kernel. Dive into the how-tos and subtle nuances with provided examples!

Early C++20 sketch of consumer-side cacheline prefetch and writer-side CLDEMOTE for a low-latency quote seqlock. Headers are not included here; the complete library is Mattbusel/fin-prewarm.

C++20 shared-memory market-data bus: wait-free broadcast ring (seqlock, exact loss accounting), lossless SPSC; 167 ns RTT vs 4.6 us pipe; full NASDAQ ITCH day to 8 processes, bit-exact

Lock seqlock

Lock-free primitives for FPU-less microcontrollers (SPSC/MPSC ring, seqlock, flag)

Typed shared-memory contract research for coherent state across workers, AudioWorklets, and WebAssembly runtimes.

Low-latency limit-order-book engine in Rust. Four book structures benchmarked to the cache line, with the microarchitecture teardown of why the "optimal" one loses on real data.

Sovereign dependent type kernel: Quantitative Type Theory (QTT) + Normalization by Evaluation (NbE) in pure zero-dependency Rust

C++20 threaded OpenGL renderer built around three systems problems: thread-affine GL contexts, a lock-free input handoff, and precise frame pacing, all measured, not asserted. The cube is the demo, not the point.