Overview
Explore the intricacies of low latency C++ programming in this comprehensive conference talk from CppNow 2023. Delve into the distinctions between low latency and high performance, examining use cases across finance, audio processing, and video game industries. Discover common challenges in low latency systems and learn C++ techniques to overcome them. Investigate topics such as CPU throttling, data types, algorithms, allocators, mutexes, concurrent code, ring buffers, atomic operations, SpinLocks, and luxury data structures. Gain insights into deferred reclamation, RCU, and weight-free implementations. Understand the tradeoffs and practical applications of various low latency programming techniques, guided by expert developer Timur Doumler.
Syllabus
Assumption
Cache
CPU throttling
Data types and algorithms
Allocators
Options
Mutexes
concurrent code
luxury queues
blockfree vs weightfree
ring buffer
sharing data
double with atomics
mutex
SpinLock
Krill
Low Latency Library
SpinLock Implementation
SpinLock Problems
SpinOnRight
Unique Pointer
Hide Logic
Tradeoffs
How it works
Deferred Reclamation
Luxury Data Structures
Use Cases
ICU
Atomic SharePoint
Lock 3 Implementation
Has pointers
Weight free
RCU
Deferred Reclamation Object
Taught by
CppNow