Running the same benchmarks as in LuaJIT against both the canonical and LuaJIT interpreters would be interesting to have a ballpark estimation about whether this is also a proper replacement where speed matters.
It may also help guide optimizations if needed.
Running the same benchmarks as in LuaJIT against both the canonical and LuaJIT interpreters would be interesting to have a ballpark estimation about whether this is also a proper replacement where speed matters.
It may also help guide optimizations if needed.