cppvalleyC++ · Systems · Low Latency · HFT
HFT Core Systems

HFT CORE SYSTEMS

Finding OS noise with rtla, ftrace, and perf

Lesson 27 of 96Phase 03: Linux determinism
Video lesson coming soon

The written objectives and hands-on lab are available now.

cppvalley on YouTube ↗

Learning objectives

What you should understand

  1. Use tracing to localize latency spikes
  2. Correlate scheduler and interrupt events with application time
  3. Choose the least intrusive diagnostic tool

Lab

Hands-on exercise

Capture and explain one synthetic OS-noise event.

Deliverable

Evidence of completion

A trace excerpt with a cause-to-symptom narrative.

Prerequisite
EP 26

Phase 03

Linux determinism

Full curriculum
21Scheduler classes and real-time hazards22Affinity, topology, SMT, and migration23CPU isolation as a system configuration24Interrupts, NAPI, softirqs, and NIC queues25Memory locking and prefaulting26Power management, turbo, C-states, and thermal behavior27Finding OS noise with rtla, ftrace, and perf28System calls and context switches are different things29Shared memory IPC with explicit ownership30cgroup v2, cpusets, memory nodes, and least privilege31Asynchronous logging: page cache, direct I/O, and io_uring32Phase gate: reproducible low-jitter host profile