home *** CD-ROM | disk | FTP | other *** search
- Path: Norway.EU.net!usenet
- From: patrick.hanevold@login.eunet.no (Patrick Hanevold)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: doubling pixels horizontally
- Date: 15 Feb 1996 19:11:36 GMT
- Organization: EUnet Norway
- Message-ID: <2182.6619T1129T1416@login.eunet.no>
- References: <4f4ibc$gl9@news.cs.tu-berlin.de> <591.6610T1165T2102@login.eunet.no><1045.6611T753T2256@vip.cybercity.dk><4faoe1$47@sunsystem5.informatik.tu-muenchen.de><2991.6612T1034T625@vip.cybercity.dk>
- <576.6613T1070T1730@login.eunet.no> <1257.6614T57T
- <4fnq7t$rkv@brachio.zrz.TU-Berlin.DE> <656.6617T1074T2417@login.eunet.no> <4ftcqf$2bm@sunsystem5.informatik.tu-muenchen.de>
- NNTP-Posting-Host: pc2.asker-pm2-1.eunet.no
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP)
-
- >|> >>>There are a few demos using this kind of blitterscreen with the
- >|> >>>spritemask.
- >|> >>It was me and Ludvig Pedersen who wrote those c2p routines, so I know.
- >|> >We don't forget that the sprite mask idea to save a pass is by Juergen
- >|> >Fischer, do we? Changing the first two blitterpasses into CPU passes
- >Yep I hope they don't =:) I guess he referes to cpu pass
- >routines he wrote on his own.
- We didnt steal your code. Ludvig pedersen wrote the cpu passes, and I wrote
- the blitterpasses.
-
- >|> >isn't a great deed.
- >|> He came first with it, so I guess he deserves the credits for that.
- >|> No new idea though, and we designed our from scratch, and our is the
- >|> fastest around. :) And the new one is even 40% faster. :) (Soon to be
- >|> released.)
- >So what is faster than copying data to chipmem ? nothing ;)
- We do 2 cpu passes for free now, and it is non-scrambled. If I remeber
- corectly your was scrambled??
-
- >On 030 blitter limits the framerate, so cpu passes increase
- >timings (but the chunky dump to chipmem is still not faster
- >than 7mb/sec) that's what you refer to ;)
- The blitter isnt slower on a 030 is it? :)
-
- >So the "40% faster" is quite a random number, I guess
- >you refer to the time needed to c2p, which is in some
- >cases (slow rendering, "doom") unimportant when done by
- >blitter. In some other cases (very fast rendering, or
- >fast cpu) your number gets important though.
- Again. The cpu passes of the new version is about %40 faster, stil doing two
- passes.
-
- >The number of passes to be done by cpu depends on
- >the cpu AND the kind of effect (execept 040+ where
- >cpu only is best for any effect).
- Now 2x2 is as fast as it can get, so you certantly dont want to do more passes
- with the CPU do you? I guess your thinking of 1x1/1x2/2x1.
-
- >I would make a file where user can config the
- >number of passes, individually for each effect
- >(or each new kind of effect). And I would make
- >it possible to select writepixelarray8 for maximum
- >c00l showoff on gfx-cards and maybe AGA+ BTW.
- Nice. I would use the rtg.library on gfx-cards thoug.
-
- <sb>Patrick Hanevold - Virtual Reality developer
- <sb>patrick.hanevold@login.eunet.no
- <sb>Amiga and official Be developer
-
-