Re: The STL and performance

Subject: Re: The STL and performance
From: Dan Herman <email@hidden>
Date: Sat, 22 Jun 2002 15:14:18 -0700

On Saturday, June 22, 2002, at 01:41 PM, Tim Conkling wrote:

Anyway, my question right now deals with the STL as a viable solution to
data access in speed-critical components of my code.

My current engine uses STL extensively throughout, as do several past engines I've written, and the performance is excellent. I wrote a particle system using STL for the rain in A Bug's Life... it performed complex operations (mass deletions, insertions, sorting, filtering, n-way interactions) on 100K's of particles very cleanly and the end result was way fast. And I never had to worry, "is my container code working correctly?"

It's probably true that in some circumstances hand-rolled containers can out-perform those in STL, but I'd give the same advice that the yacc authors gave regarding hand-rolled parsers: it may be theoretically possible to write a faster [container,parser] than [STL,yacc] can produce, but it's a heck of a lot of work to do it and do it correctly, and there are far better minds than mine (and probably yours ;-) who have put a lot of time into making it optimal. Life's too short to spend reinventing [yacc,STL,wheels].

Dan

---
Dan Herman
DigitalFish Films
http://www.digitalfish.com
_______________________________________________
mac-games-dev mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/mac-games-dev
Do not post admin requests to the list. They will be ignored.

References:
	>The STL and performance (From: Tim Conkling <email@hidden>)

Prev by Date: Re: The STL and performance
Next by Date: Re: The STL and performance
Previous by thread: Re: The STL and performance
Next by thread: Re: The STL and performance
Index(es):
- Date
- Thread

Home