I made a Laravel Octane benchmark comparing Swoole, OpenSwoole, RoadRunner, FrankenPHP, and PHP-FPM.
Report: terrylinooo.github.io/laravel-octane-benchmark
The benchmark compares p99 latency, memory, CPU overhead, DB workload, I/O workload, and different concurrency levels.
My main takeaway: Octane is not a simple “turn it on and everything is faster” solution. It depends a lot on workload, concurrency, and operational tradeoffs.
I’d appreciate feedback from people running Octane in production, especially around methodology or missing test cases.
Yes! Great benchmark comparison! It’s interesting to see how Swoole, OpenSwoole, RoadRunner, and FrankenPHP perform under different workloads. Testing with your own real application and traffic patterns is probably the best way to choose the right setup instead of relying only on synthetic benchmarks.
The focus on p99 latency and different workloads adds a lot of value
Please or to participate in this conversation.