Re: [TCLCORE] Performance advantage of compiled primitives

[Andrew P. <at...@pi...>]

On Mon, Mar 03, 2008 at 08:43:05AM +0100, Alexandre Ferrieux wrote:
> On 3/3/08, Donal K. Fellows <don...@ma...> wrote:

> > Why is that so intimately connected? If you want to do native code
> > generation, stop talking and get on with it. :-)
> 
> It's just that it is impossible to do it partially (a mix of decoded
> bytecodes and executed native instructions), so I could use a
> reduction by 2 before starting ;-)

Impossible?  Doesn't the existence of LuaJIT demonstrate otherwise?

http://luajit.org/

-- 
Andrew Piskorski <at...@pi...>
http://www.piskorski.com/