home *** CD-ROM | disk | FTP | other *** search
- Path: ifi.uio.no!usenet
- From: ludvigp@ifi.uio.no (Ludvig Pedersen)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: CHIP RAM speed test results
- Date: 2 Mar 1996 16:36:24 GMT
- Organization: Dept. of Informatics, University of Oslo, Norway
- Message-ID: <1383.6635T1034T2205@ifi.uio.no>
- References: <1067.6635T1081T1567@direktor.voima.jkl.fi>
- NNTP-Posting-Host: gymir.ifi.uio.no
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP)
-
- >Because there's so much talk about the chip ram speed and chunky
- >conversions I decided to time some values myself.
- >Machine used: A1200/030/28MHz, all caches & bursts on. Both tests took
- >over the system and disabled all interrupts, fmode was %11.
- >Test 1: 320x256x8 screen
- >------------------------
- >Copy $14000 bytes (screenful) of data from fast to chip using the following
- >routine:
- >loop:
- > rept 8
- > move.l (a0)+,(a1)+
- > endr
- > dbf d0,loop
-
- >Time: 336 rasterlines = 21.5 ms (50Hz=20ms=one frame=312 rasterlines, I
- >think)
- >Test 2: 320x256x5 screen
- >------------------------
- >Copy $c800 bytes, otherwise the same as above.
- >Time: 196 rasterlines = 12.6 ms
- >So if all Amigas have about the same chip ram speed it's impossible to do a
- >50Hz 320x256x8 chunky to planar screen, even if you had a 060/80MHz.
-
- You are using a *slow* loop, my 2 pass c2p is actually faster than your copy.
- :-)
-
- Use this:
-
- .loop:
- move.l (a0)+,d0
- move.l d0,(a1)+
- dbra d7,.loop
-
- and see if the speed improves! (Was faster on my 030/50mhz)
-
- BTW:
-
- It is possible to do 1x1 fullscreen c2p on a 040 with a scrambled buffer
- 50 fps, but then you don't have any time left for rendering! :-/
-
- The only reason why c2p is for free is because chipram is so damn slow! :((
-
-
- <sb>Ludde - Amiga Demo Coder
- <sb>Virtual Reality & Official Be developer
- <sb>ludvigp@ifi.uio.no
-
-