>Score: 130<p>>binarytrees 46.051 seconds<p>Without any frame of reference, those values are meaningless.<p>However there is one I can evaluate:<p>>1.58 MFLOPS<p>Given I have an i7-2600 and its flops on benchmarks[1] (FPU Flops score: 8500.09 MFlops), that tells me something I could have guessed: performance is 3 orders of magnitude worse than native high performing code.<p>Still, in order for this comparison to be complete, I'll have to<p>- try this benchmark in Firefox with OdinMonkey, the new and very promising JS engine[2]<p>- run a similar benchmark with "native" LUA<p><i>edit</i>: results obtained on W7 SP1 with 16GO RAM and Chrome 27.<p>1: <a href="http://www.maxxpi.net/results/show.php?ID=c3x2c5w6d7b7" rel="nofollow">http://www.maxxpi.net/results/show.php?ID=c3x2c5w6d7b7</a><p>2: did you know the logo was awesome? <a href="https://wiki.mozilla.org/File:Odinmonkey800.jpg" rel="nofollow">https://wiki.mozilla.org/File:Odinmonkey800.jpg</a>
Would be interesting to compare performance to the other ways of getting Lua to run in the browser, which include writing a Lua bytecode interpreter in JS, converting Lua to JS, and code generation from bytecode.<p><a href="https://github.com/creationix/brozula" rel="nofollow">https://github.com/creationix/brozula</a>
<a href="https://github.com/mherkender/lua.js" rel="nofollow">https://github.com/mherkender/lua.js</a>
(and others, I had a list somewhere)<p>Quite a few companies are using Lua->JS as a way of doing write once run cross platform games eg we had this talk at London Lua <a href="http://www.londonlua.org/luajavascript/index.html" rel="nofollow">http://www.londonlua.org/luajavascript/index.html</a>