C++ Electronic Trading for Cpp Programmers
summary
Explains how to build C++ low‑latency trading: maintain an order book with preallocated, integer prices (no floating rounding), use intrusive containers and open‑address hash tables to avoid heap allocations, measure latency on the wire with packet capture and TSC/RDPMC, and bypass the kernel via dpdk and io_uring or push to FPGA for extreme speed — software reaches microseconds (100 µs down to ~3 µs), FPGA claims ~45 ns FPGA; includes replayable matching‑engine simulation and concrete networking/serialization choices (SBE, fixed‑point).
AI-written summary · about
chapters (35)
- 0:00Introduction
- 0:25Why this talk
- 1:07C jobs in trading
- 2:57Who am I
- 7:04Can you trade
- 8:26What do you trade
- 9:21Why is it electronic
- 10:17Types of people trading
- 11:46Types of businesses
- 13:02Vocabulary
- 14:23Continuous Matching
- 15:19Order Group
- 20:46Code
- 22:48Conclusions
- 23:58Market Making
- 25:12Asset Management
- 30:04Network
- 33:01Multicast
- 34:02Data
- 34:47Over complexity
- 38:00Decimal numbers
- 40:54Scale Integers
- 41:24Price
- 41:45Edge
- 43:48Price Prediction
- 46:16Fair Prices
- 47:45Momentum
- 48:28Statistical Correlation
- 49:07Derivatives
- 51:01Options
- 55:21Execution
- 56:14Fairness
- 57:20Randomness
- 58:16Pickoff
- 59:26Priority
related talks
25mHow to Simulate a Low Latency Exchange in C++
01h 19mBuilding a C++/JS Browser for the Financial Markets
05mMaximising Throughput & Minimising Latency in High-Performance C++ Code
55mAbusing Your Memory Model for Fun and Profit
01h 11mC++ in the World of Embedded Systems
41mHigh-Performance Message Dispatch in C++
21mHPC Green Computing: Relationships Between Throughput and Latency
50mLeveraging Modern C++ in Quantitative Finance