11.5 KB · ⇊ 7 parts · ⤓ Micron 11.5 KB · ⤓ HTML 33.8 KB

top

LuaJIT is a tracing just-in-time compiler and interpreter for the Lua programming language.

• History

• Notable users

• Performance

• Tracing

• Internal representation

• Extensions

• DynASM

• References

The LuaJIT project was started in 2005 by developer Mike Pall, released under the MIT open source license.cite-ref-4[4]

The second major release of the compiler, 2.0.0, featured major performance increases.cite-ref-5[5]

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.cite-ref-6[6]

Mike Pall resigned in 2015 making only occasional patching to the future 2.1 version since then.cite-ref-7[7]

• CERN, for their Methodical Accelerator Design 'next-generation' software for describing and simulating particle acceleratorscite-ref-8[8]

• OpenResty, a fork of nginx with Lua scriptingcite-ref-9[9]

• Neovim, a text editor based on vim that allows the use of Lua for plugins and configurationcite-ref-10[10]

• Kong, a web API gatewaycite-ref-11[11]

• Cloudflare, who use LuaJIT in their web application firewall servicecite-ref-12[12]

LuaJIT is often the fastest Lua runtime.cite-ref-13[13] LuaJIT has also been named the fastest implementation of a dynamic programming language.cite-ref-14[14]cite-ref-cern1-15-0[15]

LuaJIT includes a Foreign Function Interface compatible with C data structures. Its use is encouraged for numerical computation.cite-ref-numcomp-16-0[16]

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-assignment intermediate representation. LuaJIT's trace compiler is often capable of inlining and removing dispatches from object orientation, operators, and type modifications.cite-ref-17[17]

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.cite-ref-cern1-15-1[15]

LuaJIT adds several extensions to its base implementation, Lua 5.1, most of which do not break compatibility.cite-ref-18[18]

• "BitOp" for binary operations on unsigned 32-bit integers (these operations are also compiled by the just-in-time compiler)cite-ref-19[19]

• "CoCo", which allows the VM to be fully resumable across all contextscite-ref-20[20]

• A foreign function interfacecite-ref-21[21]

• Portable bytecode (regardless of architecture, word size, or endianness, not version)cite-ref-22[22]

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.

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.

cite-note-11. LuaJIT tags

cite-note-22. "LuaJIT". LuaJIT. Retrieved 25 February 2022.

cite-note-github1-33. "LuaJIT/COPYRIGHT at v2.1 · LuaJIT/LuaJIT". GitHub. 7 January 2022.

cite-note-44. ↑ "The LuaJIT Project". luajit.org. Retrieved 2023-06-17.

cite-note-55. ↑ citerefpallPall, Mike. "Re: [ANN] llvm-lua 1.0". lua-users.org. Retrieved 25 February 2022.

cite-note-66. ↑ "Project status - Issue #665 - LuaJIT/LuaJIT". GitHub. Retrieved 3 February 2023.

cite-note-77. ↑ "[ANN] Looking for new LuaJIT maintainers - luajit - FreeLists". www.freelists.org. Retrieved 2023-03-29.

cite-note-88. ↑ citerefdeniauDeniau, Laurent. "Lua(Jit) for computing accelerator beam physics". CERN Document Server. CERN. Retrieved 25 February 2022.

cite-note-99. ↑ "OpenResty® - Official Site". openresty.org.

cite-note-1010. ↑ "Lua - Neovim docs". neovim.io. Retrieved 2024-05-07.

cite-note-1111. ↑ "Kong/kong". GitHub. Kong. 25 February 2022. Retrieved 25 February 2022.

cite-note-1212. ↑ "Helping to make Luajit faster". blog.cloudflare.com. 19 October 2017. Retrieved 25 February 2022.

cite-note-1313. ↑ "LuaJIT Performance".

cite-note-1414. ↑ "Laurence Tratt: The Impact of Meta-Tracing on VM Design and Implementation". tratt.net. Retrieved 2 March 2022.

cite-note-cern1-1515. ↑ citerefd-andrea2019d'Andrea, Laurent (2019). Behavioural Analysis of Tracing JIT Compiler Embedded in the Methodical Accelerator Design Software (Thesis). CERN. Retrieved 31 July 2022.

cite-note-numcomp-1616. ↑ citerefpallPall, Mike. "Tuning numerical computations for LuaJIT (was Re: [ANN] Sci-1.0-beta1) - luajit - FreeLists". www.freelists.org.

cite-note-1717. ↑ Rottenkolber, Max. "Later Binding: Just-in-Time Compilation of a Younger Dynamic Programming Language." ELS. 2020

cite-note-1818. ↑ "Extensions". LuaJIT. Retrieved 25 February 2022.

cite-note-1919. ↑ "BitOp Semantics". LuaJIT. Retrieved 25 February 2022.

cite-note-2020. ↑ "Coco - True C Coroutines". LuaJIT. Retrieved 25 February 2022.

cite-note-2121. ↑ "FFI Library". LuaJIT. Retrieved 25 February 2022.

cite-note-2222. ↑ "Extensions". luajit.org. Retrieved 2022-08-25.

cite-note-234. "DynASM Features". DynASM. Retrieved 25 February 2022.