GitHub

C++20 Synchronization Benchmarks

Celeris is a multicore event-driven simulation engine built in C++20. This page runs the actual compiled binary on the server and shows real profiler data — std::barrier, std::jthread, std::latch, per-signal atomics — comparing three synchronization strategies across four experiments.

Worker threads
Seed events
Sim end: t=5000 · Design: 8 signals, 6 processes, 3 regions
Waiting — configure parameters and click Run Engine

Ready to Run

Click Run Engine to execute the compiled C++20 simulation binary on the server. Results show real profiler counters from std::atomic<uint64_t> across all worker threads.

Performance Workshop

The exact profiling and optimization sequence used during development — from a mutex-saturated baseline to a 3.89× throughput gain. Paste your own multithreaded C++ code or use the built-in Celeris example. Steps that trigger benchmarks run the actual compiled binary on this server and show real numbers.