[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: Virtual methods?

On Fri, 22 Aug 2003, Jan Ekholm wrote:

> >> 	I was wondering; how much impact on performance do I get if I use a
> >> base class that defines a virutal method "Render" and derive all
> >> "renderable" objects from it? I mean, I'll probably have some loops
> >> calling the "Render" method for each object. Is this too expensive? Any
> >> alternatives?
> >
> >Really. algorithmic optimisation is far more valuable than worrying
> >about a tiny overhead like v-method dispatching.
> So true. Premature optimization is the root bla bla... It probably just
> guarantees that your game is never completed.

I think you are beeing a bit harsh here. Depending on how you approach 
things, some people (not me) will argue that virtual methods have an 
overhead. Personally I would never hesitate, but what if you want to call 
10000 rendering objects pr. frame? Would it be worth the hassle to store a 
pointer to the functions instead of relying on virtual methods?

I wouldn't think so, but I was looking forward to some comments that 
weren't merely a disregard of any objections to virtual methods.

Also, the original poster clearly have not yet implemented anything. While 
you consider eliminating the overhead of a virtual method an optimization 
best delayed until actually proven a problem, unless you know that or have 
a feeling that it will not actually be a major time "bummer", you should 
be entitled to ask and get an usuable answer.

/me puts on his asbestous suit


Mads Bondo Dydensborg.                               madsdyd@challenge.dk
Today, the record companies are saying MP3s are the biggest evil. Tomorrow 
they're going to say they're the greatest thing when they figure
them out.
                               - Gene Kan, Gnutella developer