Re: Hitting 60 Hz in Unreal Engine 3 (was: bothering with C# for 3D)
Richard <[email protected]>
| Newsgroups | gmane.games.devel.sweng |
|---|---|
| Organization | multi-cellular, biological |
| Message-ID | <[email protected]> |
In article <[email protected]>, Tony Albrecht <[email protected]> writes: > Thanks - I'm glad you liked it. Yes, the title is intentionally contentious > but it is a pattern that results from OO design where you focus on the code > over the data, and as such, is something I see far too often. Hopefully it > will convince some coders to think more about their data layout when > designing high performance systems. This comes back to why I was asking about continuous integration and testing with respect to performance goals. The non-performant code wasn't written all at once, it was written incrementally. Over time there should have been a visible downward trend in a CI performance graph. You could even fail a build if metrics indicate non-performant code (i.e. you blew the budget). Maybe you coded naively in the beginning because it was easier to think about and quicker to write. If you never bust your performance budget, then you've done just enough engineering and no more. However, if you bust the budget you find out as soon as possible -- and if you're blowing out the cache, it seems like you're going to find out sooner rather than later. Hmm... is it get the number of cache misses on a profiled run? Then your performance budget could be specified directly in terms of the performance you get from the cache instead of inferring it from a time budget. -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download <http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/> Legalize Adulthood! <http://legalizeadulthood.wordpress.com> _______________________________________________ Sweng-Gamedev mailing list [email protected] http://lists.midnightryder.com/listinfo.cgi/sweng-gamedev-midnightryder.com