Understanding HHVM's Compilation Process & Bytecode
For those into compilers, bytecodes, and low-level programming or just wanting to know why Facebook's HHVM project tends to be so much faster for PHP than PHP itself, here's a great article.
Sara Golemon of Facebook wrote on the official HHVM blog about the journey through the HipHop Virtual Machine's compilation process from PHP script to x86 code. Sara covered the lexing process of the simple PHP code, the parsing process, optimizations, compilation to the project's own bytecode, and then optimizations on the HHBBC, stepping into the HipHop intermediate representation, the virtual assembly, and finally the emitted machine code.
Those interested in HHVM as a faster alternative to PHP or for the Hack Language, see Sara Golemon's post entitled The Journey of a Thousand Bytecodes.
Sara Golemon of Facebook wrote on the official HHVM blog about the journey through the HipHop Virtual Machine's compilation process from PHP script to x86 code. Sara covered the lexing process of the simple PHP code, the parsing process, optimizations, compilation to the project's own bytecode, and then optimizations on the HHBBC, stepping into the HipHop intermediate representation, the virtual assembly, and finally the emitted machine code.
Those interested in HHVM as a faster alternative to PHP or for the Hack Language, see Sara Golemon's post entitled The Journey of a Thousand Bytecodes.
1 Comment