POV-Ray : Newsgroups : povray.off-topic : Trolltech doesn't know how amortized time works : Re: Trolltech doesn't know how amortized time works Server Time
5 Sep 2024 03:19:11 EDT (-0400)
  Re: Trolltech doesn't know how amortized time works  
From: Darren New
Date: 3 Nov 2009 11:47:40
Message: <4af05eac@news.povray.org>
Darren New wrote:
> geometric progression of expansions that converges, rather than bumping 
> up to (say) the next prime number or adding a constant or something.

In other words, "bump the size by 4096" doesn't give you O(1) behavior 
unless you decide that remapping memory blocks is "free". If you assume 
remapping memory blocks takes a fixed amount of time per remap (even if it's 
a fixed amount of time regardless of how many blocks you remap) I think 
you're back to O(1).

Of course, by the time you're talking about that sort of thing, order 
statistics fall apart, because you have a limited amount of address space in 
C and C++ (as in, sizeof(void*) is fixed and finite), so you're going to run 
out of memory before it really makes a difference.

-- 
   Darren New, San Diego CA, USA (PST)
   I ordered stamps from Zazzle that read "Place Stamp Here".


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.