Re: LuaJIT2 performance for number crunching

[Francesco Abbate <francesco.bbt@...>]

2011/2/14 Geoff Leyland <geoff_leyland@fastmail.fm>:
>> By the way if someone want to help with this task he is very welcome.
> Any particular reason to use RK4?  My trusty old copy of Numerical Recipes suggests higher order methods with implicit error estimation are better.

Hi,

I agree with you, other methods are better but the idea is that I want
to test the LuaJIT implementation with a simpler algorithm and if it
works I will implement later a better algorithm.

As you have seen I have discovered already a major problem with the
RK4 implementation that I wrote and I need to adopt a different
approach with dynamic code generation to specialize the code with the
dimension of the system.

-- 
Francesco