Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!emory!gatech!hubcap!ncrcae!ncr-sd!se-sd!cns!dltaylor
From: dltaylor@cns.SanDiego.NCR.COM (Dan Taylor)
Newsgroups: comp.sys.amiga.advocacy
Subject: Re: AMIGA
Message-ID: <950@cns.SanDiego.NCR.COM>
Date: 9 May 91 00:43:43 GMT
References: <1991May6.115535.8982@sugar.hackercorp.com> <1991May6.164821.8807@cunixf.cc.columbia.edu> <937@cns.SanDiego.NCR.COM> <1991May8.062658.14796@cunixf.cc.columbia.edu>
Distribution: usa
Organization: NCR Corp. SE-San Diego
Lines: 32

In <1991May8.062658.14796@cunixf.cc.columbia.edu> es1@cunixb.cc.columbia.edu (Ethan Solomita) writes:

>low-res screen, or an 030 on a high-res screen. My question was
>simply about getting it into memory. Do compressed ANIMs lend
>themselves to off-disk loading?

ANIMs are part of a family of specified file formats called, collectively,
IFF.  The nearest-to-mind analog is COFF, or ELF, on System V.  The file
structure has defined header tags that tell the reader program how long
the following segment is.  Therefore, off-disk loading works well in the
faster Amigas.  Some of the PD demos assume that the Amiga is a "vanilla"
A1000 or A500, so do a disk-to-memory copy first, but it isn't necessary
on a DMA, or fast-processor-assisted disk interface.  The ability to
have 2, 3, or more, full frame buffers, then switch, without processor
intervention, during vertical retrace, really helps.  Other machines
with more power have trouble maintaining the smoothness and sync.

I'm not at my Amiga, now, so I don't have compression ratios handy.  If
I don't see a posting in a day, or so, I'll get some numbers.

>	Maybe try a Wren drive. I've been told that they'll do
>2M/sec, but the question would then be the realistic throughput
>of the controller.

I've seen 5mbs in synchronous, media to memory, but that's why I
qualified it with "through the filesystem".  None of the AmigaDOS
interfaces I've seen run scatter-gather lists, so there's probably
too much setup/completion overhead for block-at-a-time transfers to
sustain 2mbs.  I was just pointing out that the I/O bus can do it,
if the hardware/software is in place.

Dan Taylor