From: ellman AT xs4all DOT nl () Newsgroups: comp.os.msdos.djgpp Subject: Re: Allegro & sprite stretching optimization Date: 1 Jun 1997 13:15:41 GMT Organization: XS4ALL Message-ID: <5mrslt$a36$1@news0.xs4all.nl> References: <01bc6d4e$720f1900$ec3e63c3 AT default> NNTP-Posting-Host: xs2.xs4all.nl Lines: 23 To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Precedence: bulk Another problem with Allegro's compiled sprite stretcher is that for sprites with a large width, you risk overflowing the cache with the compiled line stretcher, causing it to be re-loaded into the cache for each line. 486s have the aditional problem of a single code and data cache, which means that when data is written to memory, it sometimes replaces some code that's in the cache. Pentiums have seperate code and data caches, but you must make sure the stretcher can fit in the cache. This routine works best on 386s with no cache, and when the destination size has a small width. However, as you add extra things to the sprite stretcher (such as translucency, Z-buffering, etc. (BTW, I've already added these things to the stretcher compiler myself, and hopefully, Shawn will include them in the next version of Allegro)), the ammount of code per pixel gets larger. AE. -- Andrei Ellman -- URL: http://www.xs4all.nl/~ellman/ae-a -- ae1 AT york DOT ac DOT uk "All I wanna do is have some fun :-) || ae-a AT minster DOT york DOT ac DOT uk I've got the feeling I'm not the only one" || mailto:ellman AT xs4all DOT nl -- Sheryl Crow :-) || It's what you make of it.