BPF LRU Per-CPU Hash Maps
Explains the structure, lookup and update behavior, eviction policy, concurrency model, and use cases of BPF LRU per-CPU hash maps.
// TAG
#Performance
Posts in this collection, ordered newest first.
The TLB-Hit Model
Generalizes the TLB fast-path idea into a cache model for repeated policy decisions and analyzes its performance trade-offs.
Busy Waiting
Explains how busy waiting consumes CPU while waiting for a lock, along with its benefits, costs, and appropriate use cases.
Context Switching
Explains how processes and threads save and restore execution state during context switches and where the overhead comes from.