home *** CD-ROM | disk | FTP | other *** search
- Path: oreig.uji.es!ii202
- From: ii202@rossegat.uji.es (Jorge Acereda Macia)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Demo/game to OS frien
- Date: 23 Feb 1996 17:52:13 GMT
- Organization: Universitat Jaume I. Castell≤ de la Plana. Spain
- Distribution: world
- Message-ID: <4gkusd$r6s@oreig.uji.es>
- References: <38232498@kone.fipnet.fi> <4gctlm$bcn@oreig.uji.es> <38232604@kone.fipnet.fi> <4gj3ve$f9c@sunsystem5.informatik.tu-muenchen.de>
- NNTP-Posting-Host: @rossegat.uji.es
- X-Newsreader: TIN [version 1.2 PL2]
-
- Juergen "Rally" Fischer (fischerj@Informatik.TU-Muenchen.DE) wrote:
- > Jyrki Saarinen (jsaarinen@kone.fipnet.fi) wrote:
-
- > : > c2p8.s is too slow for 25 fps effects (due to blitter speed).
- > : > For this 3 CPU passes and one blitter pass is better.
-
- > : Yep. I still believe 030/50 could do free c2p by doing
- > : three passes with the CPU and the last one with the blitter.
- > : This one would not limit the frame rate too much either.
-
- Well, after some headaches I don't believe it's possible to
- do 3 passes for free on 030@50 :-( If someone tells me I'm
- wrong I'll be really happy...
-
- > : Hmm. I calculated that those 2 blitter passes limit to 25fps
- > : in c2p8.s in 320x128x8, but you are talking about 320x256x256
- > : I guess?
-
- I'm experimenting with 256x256. The blitter can't do 2 passes at
- 25 fps. So, for fast effects a 3cpu+1blitter routine is the solution.
-
- > one pass (without cpu doing chipacess!) 320x256x8planes needs
- > almost exactly 0.04s, i.e 25fps. copying to chipmem by 030
- > will need 0.014s, when assuming 6mb/sec. Well, doing 3 passes
- > maybe 030-50 will drop down a bit.
-
- > REQUEST: Those who said 030-50 does 2 passes for free, does
- > this mean at 7mb/sec if no plane dma ? Ludde said 4-5ms, but thats
- > rather 4.4mb/sec! Anyone knos if a 040/25 can do 7mb/sec c2p if
- > sitting in A1200 (the real Ami :-> ) ?
- > END REQ.
-
- > I guess blitter doing 1 pass will need 0.05 sec including cpu interference ?
- > would mean 20fps.
-
- > : > Peter uses PC relative addressing in his c2p8.s to get the
- > : > current task and signal bits, but AFAIK I can't do this in E...
-
- > mhm, why not just use global variables.
-
- > Why you can't do this in E ?
-
- My problem was in my qblit cleanup routine. In a qblit.function you
- get A1 pointing to the bltnode structure. But in the cleanup this
- is not true (I think). So I needed some way to access the bltnode
- structure. I've solved it assembling with A68k using pc relative
- addressing :-) Thanks anyway.
-
- cu,
- --
- ---------------------------- --------------------------------------------
- | Jorge Acereda | Dream the same thing everynight |
- | ii202@rossegat.uji.es | I see our freedom in my sight |
- | Intel Outside | No locked doors, no windows barred |
- | Amiga Rules | No things to make my brain seem scarred |
- ---------------------------- --------------------------------------------
-