home *** CD-ROM | disk | FTP | other *** search
- Path: nntp.teleport.com!sschaem
- From: sschaem@teleport.com (Stephan Schaem)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Wolfenstein Texture mapping problem
- Date: 28 Feb 1996 19:39:43 GMT
- Organization: Teleport - Portland's Public Access (503) 220-1016
- Message-ID: <4h2b1v$cn0@maureen.teleport.com>
- References: <4g5mv0$sns@news01.uni-trier.de> <4gimj3$4lp@news01.uni-trier.de> <4glh4j$rgg@maureen.teleport.com> <4guqo5$4l9@news01.uni-trier.de>
- NNTP-Posting-Host: kelly.teleport.com
- X-Newsreader: TIN [version 1.2 PL2]
-
- Hans-Joerg Frieden (hfrieden@fix.uni-trier.de) wrote:
- : Stephan Schaem (sschaem@teleport.com) wrote:
- : : Why the table, and the z buffer? what are those 3 instructions?
- : I do fill a zbuffer with the information from the raycaster, so that
- : while I do the raycasting, the picasso blitter can clear the buffers I am
- : writing to... The three instructions (from memory, dunno if I am correct
- : here) are
-
- You you so a setrast or rectfill on the picaso the function return
- immediatly?!?!
-
- And you fill a zbuffer so you can use the picaso blitter? this is unclear.
-
- : loop:
- : move.b (a0,d0.w),(a1)+
- : addq #2,a0
- : dbra d1,loop
-
- what is this suposed to do.... I suposed its just part of the loop,
- otherwise (a0,d0.w) is superfluous. And why +2? what data are you
- skyping?
-
- : : How is the picaso update done? you write the the buffer directly?
- : : you use an OS function? How long does it take the OS to update
- : : a the fatsmem buffer? I'm asking cause this might be a major cause
- : : for your slowdown.
- : I do write into a fastram buffer and then copy it to the card ram with
- : word writes. The Picasso is ZII, so word writes are better than byte
- : writes. I can directly access the picasso memory, as a linear chunky
- : stream, I don't need OS functions for that. I am using the rtgmaster
- : library at the moment, with my rtgPICA sublib, so this will essentially
- : also work on a CyberGfx or an EGS board
-
- Is just copiying the buffer in byte VS copying in work + rendering
- in byte in fast slower?
-
- You might want to make your code 16bit and write directly to the card
- and take big advantge of bus memory access overlap. Doing might nake
- you stuff run faster then the 8bit version :)
-
- if you do this you will run at picaso max frame rate:
-
- REPEAT 16
- move.w d1,d4
- move.b d0,d4
- movea.l d4,a0
- movea.w (a0),(a1)+
- addx.l d3,d0
- addx.l d2,d1
- ENDR
-
- This should run at picaso speed even on a 16mhz 030, so at
- the speed to clear picaso memory. (My guess, but you should try)
-
- Stephan
-