Re: Lua versus Ravi versus Luajit - another perf test

[Dibyendu Majumdar <mobile@...>]

On 7 March 2015 at 21:16, Dirk Laurie <dirk.laurie@gmail.com> wrote:
> Some strongly-typed languages have identifiers that signal
> type. Fortran with the IJKLMN => integer convention and
> IMPLICIT statement. Perl with those ugly suffixes.
>
> Would efficiency and/ore user-friendliness be improved with
> that kind of thing?

Hi Dirk,

Thanks - this is very relevant.

I am already using optional typing to make the code more efficient.
This is helped by the fact that Lua 5.3 distinguishes between integer
and floating points.
The optional typing helps the compiler to generate floating point or
integer operations without having to check this at runtime. So for
example in standard Lua 5.3 an addition always requires a type check.
The first test is to see if the values are integers, then if they are
floats, and finally if user defined operation is available. Whereas by
using type annotation I can avoid all that and directly perform
integer or floating point addition.

I like the idea that if you don't specify types then it defaults to
standard Lua behaviour.

I think more optimization is possible though. For instance - could I
avoid setting the types if the compiler knows they are
integer/floating? I thought I could but I found an issue.
Types can be preset for locals so those are fine.
But when expressions are evaluated using temporary registers then if
the type is not set what happens is that any operation that uses the
temporary value gets the wrong type. Perhaps the compiler could be
clever enough to set the type at the end of the expression evaluation
rather than at every step. I will look into this more at some point.

Thanks and Regards
Dibyendu