LuaJIT is a tracing just-in-time compiler for the Lua programming language. Mike Pall, a primary maintainer of the project had resigned in 2015, resorting only to occasional patching to the future 2.1 version.[4]
History
The LuaJIT project was started in 2005 by developer Mike Pall, released under the MIT open source license.[5]
The second major release of the compiler, 2.0.0, featured major performance increases.[6]
LuaJIT uses rolling releases. Mike Pall, the creator and maintainer recommends using the tip of the v2.1 branch, and does not believe in releases.[7]
Notable users
CERN, for their Methodical Accelerator Design 'next-generation' software for describing and simulating particle accelerators[8]
LuaJIT is often the fastest Lua runtime.[13] LuaJIT has also been named the fastest implementation of a dynamic programming language.[14][15]
LuaJIT includes a Foreign Function Interface compatible with C data structures. Its use is encouraged for numerical computation.[16]
Tracing
LuaJIT is a tracing just-in-time compiler. LuaJIT chooses loops and function calls as trace anchors to begin recording possible hot paths. Function calls will require twice as many invocations to begin recording as a loop. Once LuaJIT begins recording, all control flow, including jumps and calls, are inlined to form a linear trace. All executed bytecode instructions are stored and incrementally converted into LuaJIT's static single-assignmentintermediate representation. LuaJIT's trace compiler is often capable of inlining and removing dispatches from object orientation, operators, and type modifications.[17]
Internal representation
LuaJIT uses two types of internal representation. A stack-based bytecode is used for the interpreter, and a static single-assignment form is used for the just-in-time compiler. The interpreter bytecode is frequently patched by the JIT compiler, often to begin executing a compiled trace or to mark a segment of bytecode for causing too many trace aborts.[15]
-- Loop with if-statementlocalx=0fori=1,1e4dox=x+11ifi%10==0then-- if-statementx=x+22endx=x+33end
---- TRACE 1 start Ex.lua:5---- TRACE 1 IR0001 int SLOAD #2 CI0002 > num SLOAD #1 T0003 num ADD 0002 +110004 int MOD 0001 +100005 > int NE 0004 +00006 + num ADD 0003 +330007 + int ADD 0001 +10008 > int LE 0007 +100000009 ------ LOOP ------------0010 num ADD 0006 +110011 int MOD 0007 +100012 > int NE 0011 +00013 + num ADD 0010 +330014 + int ADD 0007 +10015 > int LE 0014 +100000016 int PHI 0007 00140017 num PHI 0006 0013---- TRACE 1 stop -> loop---- TRACE 2 start 1/4 Ex.lua:8---- TRACE 2 IR0001 num SLOAD #1 PI0002 int SLOAD #2 PI0003 num ADD 0001 +220004 num ADD 0003 +330005 int ADD 0002 +10006 > int LE 0005 +100000007 num CONV 0005 num.int---- TRACE 2 stop -> 1
Extensions
LuaJIT adds several extensions to its base implementation, Lua 5.1, most of which do not break compatibility.[18]
"BitOp" for binary operations on unsigned 32-bit integers (these operations are also compiled by the just-in-time compiler)[19]
"CoCo", which allows the VM to be fully resumable across all contexts[20]
DynASM is a lightweight preprocessor for C that provides its own flavor of inline assembler, independent of the C compiler. DynASM replaces assembly code in C files with runtime writes to a 'code buffer', such that a developer may generate and then evoke code at runtime from a C program. It was created for LuaJIT 1.0.0 to make developing the just-in-time compiler easier.[citation needed]
DynASM includes a bare-bones C header file which is used at compile time for logic the preprocessor generates. The actual preprocessor is written in Lua.