Difficulty: Hard
Category: Networking & Systems
Practice quant interview questions from top firms including Jane Street, Citadel, Two Sigma, DE Shaw, and other leading quantitative finance companies.
Topics: NUMA, memory-allocation, systems-programming, low-latency, trading-systems
You are developing a high-frequency trading application on a NUMA (Non-Uniform Memory Access) system. On this system, accessing memory on the same NUMA node as the processing core is significantly faster than accessing memory on a different node. Specifically, accessing remote memory costs 1.5 to 3 times more than accessing local memory. The trading application processes market data and executes trades, requiring frequent memory access. To minimize latency and maximize performance, how should t
Practice this hard trader interview question on MyntBit - the all-in-one quant learning platform with 200+ quant interview questions for Jane Street, Citadel, Two Sigma, and other top quantitative finance firms.