Re: No realloc

[Enrico Colombini <erix@...>]

Peter Pimley wrote:
> It depends on your definition of 'better' ;) (i.e. whether you're after
> speed or minimal memory usage).

I may have CPU limits, but it's probably much better to save RAM in my 
system (and it won't depend on program execution patterns).

> It's certainly worth trying the memcpy+malloc+free listed earlier.  Then
> measure it and see if it's actually a problem on your particular setup.
> The couple of times I've done this, it has turned out to be negligable.

Thanks for the shared experience. As I'm expecting more significant 
bottlenecks elsewhere, I'll follow this approach.

I now have a working system, but measurements at this point make little 
sense as I'm not working on the actual target hardware.
Any suggestion of 'standard' benchmarks to evaluate (mainly) 
(re-)allocation, table access and general Lua speed (no libs, no fp 
calculations)?

(or is there already some sort of repository with Lua performance on 
varius hardware architectures?)

  Enrico