[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
- Subject: Re: C vs LuaJit
- From: Mike Pall <mikelu-1008@...>
- Date: Mon, 30 Aug 2010 15:53:13 +0200
Vyacheslav Egorov wrote:
> Indeed we need some pretty strong both control-flow and data-flow
> sensitive analysis here. That definetely smells like abstract
> interpretation to me.
Now we're coming full circle. The recording phase of a trace
compiler may be considered to be the complement of abstract
interpretation. It closely follows the runtime activity of an
actual interpreter and outputs matching IR instructions.
The main difference over abstract interpretation is that it has
easy access to all the important runtime information (types,
values, call targets, execution statistics). This helps to reduce
the dynamisms inherent in the language semantics.
> I switched to plain text. Hope that helps.
Much better.
--Mike
- References:
- C vs LuaJit, Patrick McCavery
- Re: C vs LuaJit, Joe Finn
- Re: C vs LuaJit, Jean-Yves F. Barbier
- Re: C vs LuaJit, Alex Stup
- Re: C vs LuaJit, Jean-Yves F. Barbier
- Re: C vs LuaJit, Miles Bader
- Re: C vs LuaJit, Mike Pall
- Re: C vs LuaJit, Vyacheslav Egorov
- Re: C vs LuaJit, Mike Pall
- Re: C vs LuaJit, Vyacheslav Egorov