[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
- Subject: Re[2]: LuaJIT performance
- From: Bulat Ziganshin <bulat.ziganshin@...>
- Date: Mon, 10 Aug 2009 21:21:38 +0400
Hello Mike,
Monday, August 10, 2009, 9:15:02 PM, you wrote:
> This works out just fine for desktop-class CPUs, but of course one
> needs to rethink narrowing for FP-challenged CPUs. In this case a
> dual representation of stored number values with normalization
> steps after FP operations is probably better. But this has
> far-reaching consequences for the whole VM. So far I've been able
> to avoid it.
it's probably obvious, but are you considered distinguishing 2 number
types and dispatching code for them separately? JIT does it for Lua
standard types, i.e. you may generate two invocations of some
functions depending on number/string type of its parameter. how about
separating also the int/double cases?
--
Best regards,
Bulat mailto:Bulat.Ziganshin@gmail.com
- References:
- LuaJIT performance, John C. Turnbull
- Re: LuaJIT performance, Mike Pall
- Re: LuaJIT performance, Alex Davies
- Re: LuaJIT performance, Michael Bauroth
- Re: LuaJIT performance, RJP Computing
- Re: LuaJIT performance, Mike Pall
- Re: LuaJIT performance, Alexander Gladysh
- Re: LuaJIT performance, Timm S. Mueller
- Re: LuaJIT performance, Mike Pall
- Re: LuaJIT performance, David Given
- Re: LuaJIT performance, Mike Pall