1. Introduction
2. Features
3. Installation
4. Prerequisites
5. Usage
6. Benchmarking
7. Acknowledgments

Welcome to the Low-Latency Programming Repository, a project aimed to serve both academics and industry practitioners in the realm of High-Frequency Trading (HFT). With a focus on optimizing latency-critical code, this repository provides comprehensive insights, techniques, design patterns, and best practices that are statistically benchmarked to mitigate latency in HFT systems.

• Comprehensive guide on low-latency programming techniques such as Cache Warming, Constexpr, Loop Unrolling, Lock-Free Programming, and Short-circuiting.
• Market-neutral statistical arbitrage pairs trading strategy optimized for low-latency.
• Implementation of the Disruptor pattern in C++ optimized for speed and scalability.

Clone this repository to your local machine to get started.

git clone https://github.com/0burak/imperial_hft.git
cd imperial_hft

• C++ Compiler with C++11 support or higher
• CMake version 3.10 or higher
• Google Benchmark

After Google Benchmark has been installed and the benchmark folder if in the same directory as the mybenchmark.cc file, the user can compile the my-benchmark.cc file using the makefile provided in the repository.

g++ mybenchmark.cc -std=c++20 -isystem benchmark/include -Lbenchmark/build/src -lbenchmark -lpthread -o mybenchmark

All techniques and strategies are rigorously benchmarked for latency reduction, cache efficiency, and overall performance. The benchmark/ directory contains all benchmarking scripts.

• A special thanks to Dr Paul A. Bilokon for his evaluation and invaluable feedback.