Commit

124156f

If your application spends most of its time in dense loops over thousands of entities, with elephc you can work on exactly that hot path using buffer<T>: typed, contiguous storage instead of a PHP array. On the position updates benchmark, 4000 particles over 40 steps: 60 ms with a PHP array 5.4 ms with buffer<T> About 11x faster. The math doesn't change: what changes is how the data sits in memory, with less boxing and less indirection. Not a magic wand for the whole application, but when the bottleneck is a dense loop it can make a real difference. Measured on CI, macos-15 arm64. If you have a similar case, tell me where your hot paths are. And if elephc interests you, a ⭐ on the repo would be much appreciated: https://github.com/illegalstudio/elephc
GitHubGitHub - illegalstudio/elephc: A PHP-to-native compiler. Takes a subset of PHP - including eval() - and compiles it directly to native assembly, producing standalone binaries for macOS ARM64, Linux ARM64, and Linux x86_64. No VM, no external runtime, no dependencies: dynamic eval falls back to an o…A PHP-to-native compiler. Takes a subset of PHP - including eval() - and compiles it directly to native assembly, producing standalone binaries for macOS ARM64, Linux ARM64, and Linux x86_64. No VM...

commit 124156f