The Hardware is Talking — Are You Listening?

Most engineers treat hardware as a black box. In high-frequency trading, we can't. When your entire latency budget is measured in nanoseconds, understanding the machine isn't optional.

This session is a guided tour through mechanical sympathy: the discipline of writing software that works with the hardware, not against it. I'll walk through CPU caches and memory layout, branch prediction (and when going branchless actually matters), speculative execution, and the profiling tools that make the invisible visible like perf, VTune, and hardware performance counters.

You'll leave with a mental model for reasoning about performance that transfers to any language and any domain where latency or throughput matters.