🚀 go-pugleaf

In article <2gvtju$9fl@harbinger.cc.monash.edu.au>, 
ins557g@aurora.cc.monash.edu.au (Mark Brophy) writes (answering to a flame
stream I deleted) :

[ ...Amiga is better  grrr...but PC has chunky, textures impossible on Amiga
..grrr.....use c2p routine....] :)

>: BullShit.  Chunky and Planar both have good and bad points.  Chunky is
[..]
>: move gfx, w/o the need for a custom gfx co-processor.  Which (Oh My)
>: the Amiga line of computers have...
>
>Once again this is all very well but when it comes to rotated graphics, it is a
>LOT easier to move them in chunky mode. Sure you can use the blitter for
>some things but with pc graphics you don't need to.
>Once again I must stress that we ARE talking about wolf3d type graphics
>Had we been talking about an 8 way scrolling screen, then the amiga would
>utterly kick butt!

A 286 also could do 8 way scrolling (there are registers in the VGA). But 
nobody seems to have noticed that :-)  The real Amiga performance is shown 
on a wildy splitted dualplayfield with hundreds of colors (copper), many 
bobs and a large number of VSprites. Ok, back to the topic..... :)

I'd now like to talk about ISA and really slow VGA card's I saw (the 
card needed several frames to simply fill the screen :), but, let's 
cancel this....

[....]

>: gfx than another.  Below you make a PROCESSOR comparision (68020 -vs- 80386)
>: based on gfx.  

Good! more about that later...

>And again we ARE talking about wolf 3d stuff so while i'm sure there are a few
>blitter tricks(like resizing in the y dirrection with the modulos)
>you can use but the chunky to planar overhead is just too much for the
>blitter to compensate.
(Well, modulo y resizing: you only can unzoom by 2,3,4.., right ?
Using the copper you can perform realtime-y-zooming, stretching and...)

If the blitter does c2p paralell to the cpu, the overhead is little (A1200
might need FastRam). So if the time for chunky operations is more than for
c2p (what's likely), it doesn't disturb. I think we can forget the problems
of plane operations and say: if gfx is too slow, it's not programed correctly
(talking about AGA machines).

[....]

:>Yes true. The A1200 even has the power to do wolf 3D, but a 16Mhz 386 (eqiv to
:>the 16Mhz 020) would have to do less in order to do this that the 020. This

16Mhz 386 equiv to 16Mhz 020 ? I don't belive. 
A 386 is as fast as an 286 (sometimes even a cycle slower in real mode,
I heard). Of course you could now say 386 has 32 bit, but we are talking
about byte-oriented gfx, so it's really like a 286.

So you say a 286 has more performance than a 020 ???
I think we don't have to argue about that, everybody knows it's not true.

We could argue if a 14Mhz 68000 can beat a 16Mhz 386. Running a routine 
that needs more than 8 registers :-D  I think it could sometimes.

Juergen.