How does the PHP 8 JIT work and when does it improve performance?

Summary

The JIT compiles hot opcodes to native machine code at runtime on top of OPcache. It greatly speeds up CPU-bound work (math, image processing), but gives little benefit to typical I/O-bound web apps.

Longer Answer

PHP normally compiles source to opcodes (cached by OPcache) that the Zend VM interprets. The JIT, built on DynASM, adds a further step: it compiles frequently executed code to native machine code. The tracing JIT (opcache.jit=tracing, default mode when enabled) profiles hot loops and functions; the function JIT compiles whole functions.

Most web requests spend their time waiting on databases, network and I/O, so benchmarks like WordPress or Laravel see small gains. Fractals, ML in PHP and long-running workers benefit far more. PHP 8.4 moved to a new IR-based JIT backend and changed the configuration defaults (opcache.jit=disable with a non-zero buffer).

Code Example

; php.ini
opcache.enable=1
opcache.enable_cli=1
opcache.jit_buffer_size=128M
opcache.jit=tracing

Key Points

• Requires OPcache; enable via opcache.jit_buffer_size and opcache.jit
• Tracing JIT targets hot paths
• Large gains for CPU-bound code, small for I/O-bound apps
• Measure with real workloads before enabling in production

What interviewers look for

Nuance: why JIT doesn't matter much for CRUD apps, and that OPcache (not JIT) is the big production win.

Common mistakes

• Enabling JIT without OPcache
• Promising 3x speedups for a typical Laravel app

Follow-up questions

• What does OPcache preloading do?
• How would you benchmark JIT impact?