Svitla Smart Talk. Demystifying CPU vs. IO Bottlenecks: Real-World Examples Workshop

CPU-bound and IO-bound are two fundamentally different workflows, and they have to be considered during design and later optimization. This workshop aims to build a well-grounded understanding of key differences between the two with a bunch of examples using `perf` or much more modern and fancy (yet powerful!) `eBPF`.

Attendees are supposed to be familiar (basic level is fine) with concepts like “CPU”, “network”, “file system”, “async/await” and related. Prior Linux and console tools like `perf` are not required, however will be demonstrated a lot. 

In this session, you’ll learn:

✅ The key differences between CPU-bound and I/O-bound processes and how they originate at the hardware level, how runtimes reflect that, and what specific problems recently appeared features like `async/await` do solve.

✅ Which core tools and metrics might help you understand the system’s behavior live and be confident in your next move: theory must agree with practice.

✅ As a pleasant side-effect: how to design & optimize based on that.

✅ Most importantly: A LOT of mysteries and lies around async/await will get DEMISTIFIED.

Whether you're a FE/BE engineer, DevOps, DBA, or tech enthusiast, this workshop will equip you with actionable insights to improve your systems' performance (especially on a modern Linux system, but definitely not limited to it). Don't miss out!

Speaker

Serhii Hrechko - Senior Software Engineer  at Svitla Systems.