rr
A lightweight record-and-replay debugger that captures process and thread execution for replay in gdb with efficient reverse continue, reverse stepping, and hardware data watchpoints.
Why use it
Once an intermittent race, memory corruption, or crash is recorded, engineers can repeatedly trace state backward from failure, improving diagnosis of difficult C++ engine defects.
Where it fits
Use testing, performance, accessibility, and diagnostic tools to remove problems players will notice.
What to check
It requires Linux and supported Intel, AMD Zen, or AArch64 microarchitectures, while virtual machines need performance-counter access; measure recording overhead on syscall-heavy or highly parallel workloads.