On Fri, 26 Feb 2016 09:18:20 +0100
Anselm R Garbe <garbeam_AT_gmail.com> wrote:
> On 26 February 2016 at 09:01, Mattias Andrée
> <maandree_AT_kth.se> wrote:
> > Performance is not really something suckless
> > concerns itself about. They favour solutions
> > that are simpler to implement and maintain
> > but asymptotically slower. But in the case of
>
> I have to object here. It is correct that performance is
> not the primary concern of suckless, but that does not
> mean that suckless doesn't care for performance
> considerations or would accept slow software algorithms.
>
> Actually the truth is often the opposite, simple
> solutions often prove to be faster(!) than their complex
> counterparts, because intensely smart but complex
> implementations often just catch up the computational
> waste of the complexity itself in the first place.
>
> Looking at dwm vs gnome is a good starting point. The
> latter one has fewer window management capabilities, but
> consists of a software stack that is probably by a factor
> of 1000 more complex in terms of LOC and hence a lot
> slower (just talking about the window management of
> Gnome).
>
> Of course the discussion about numeric algorithms is a
> bit different, though still a simple implementation
> should not imply bad performance. If you believe so, then
> someone must have misguided you quite a bit.
>
> -Anselm
>
Yes, complex and asymptotically faster algorithms usually
have an overhead that makes it them subpar for small input.
What you can do is estimate which algorithm would be
fastest.
Received on Fri Feb 26 2016 - 09:33:08 CET