[ruby-core:32640] Re: Proposal for Optional Static Typing for Ruby

On Wed, Sep 29, 2010 at 3:10 AM, Joshua Ballanco <[e-mail]> wrote=
:
> On Sep 29, 2010, at 12:33 AM, Bill Kelly wrote:
>
>> Is everyone familiar with the approach Charles Nutter has
>> taken with his Mirah language (formerly Duby)?
>
> It's been a while since I've taken a close look at Mirah/Duby, but my und=
erstanding is that the performance gains that Charles has achieved have les=
s to do with static vs dynamic typing and much more to do with Mirah's type=
 system being a closer match to Java's type system (for which the JVM is al=
ready optimized).

Incorrect. The performance gains come from being able to specify
direct invocations or raw numeric operations statically, avoiding the
type checks and numeric boxes that normally are required for Ruby
code.

JRuby can probably get within 2-3x Java's invocation performance, and
perhaps closer to 1x Java's performance with invokedynamic, but in all
cases dynamic languages must pepper their optimizations with type
guards, modification checks, and so on, just in case someone changes
something or an unexpected type floats through the system. Mirah (and
other statically-typed languages) avoids this VM-confounding behavior
by making compile-time guarantees about what subset of types the code
will see, allowing the VM to be more aggressive in its optimizations.

- Charlie