I've been tinkering lately with NeoLua and C#. NeoLua is an implementation of Lua based on the DotNet Dynamic Runtime. I've had good success with it so far.
Then just yesterday, Mr. Vinicius Jarina updated his KeraLua DotNet/Lua bindings to 5.3 (Happy New Year!). I was playing with that last night and the few things I tried worked quite well too. That started me wondering about how NeoLua performs compared to standard Lua. (The NLua C# API - also by Mr. Jarina - needs to be ported to the new KeraLua, then I could do a real "apples to apples" comparison.)
With all that said, I was hoping is that someone might have a battery of Lua scripts to test Lua performance? I'll write some of my own for simple things, but if someone has a test suite that targets performance sensitive aspects of Lua, I'd be interested in using those as well.
I'm getting some weird results so far with my own very limited tests. If I run KeraLua (Lua) first, then NeoLua (dynamic runtime) is faster by about 12% on average. If I run NeoLua first, KeraLua is faster by 30%.