NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/programmer/1959

Text File | 1996-08-05 | 3.0 KB | 82 lines

Path: informatik.tu-muenchen.de!fischerj
From: [e-mail] (Juergen "Rally" Fischer)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: PPC compilers
Date: 25 Jan 1996 15:36:54 GMT
Organization: Technische Universitaet Muenchen, Germany
Distribution: world
Message-ID: <4e882m$[e-mail]>
References: <4d42gg$[e-mail]> <4dov8s$[e-mail]> <[e-mail]> <[e-mail]>
NNTP-Posting-Host: hphalle5h.informatik.tu-muenchen.de
Originator: [e-mail]


In article <[e-mail]>, [e-mail] (Ludvig Pedersen) writes:
|> Organization: Dept. of Informatics, University of Oslo, Norway
|> Lines: 39
|> Message-ID: <[e-mail]>
|> References: <4d42gg$[e-mail]> <4dov8s$[e-mail]> <[e-mail]>
|> NNTP-Posting-Host: gymir.ifi.uio.no
|> X-Newsreader: THOR 2.22 (Amiga;TCP/IP)
|>
|>
|> >> To do fast 16:16 bit fixed point maths (ie for bitmap scaling). You have
|> >> the fraction in the  high word and the integer part in the low word.
|> >>áBecause of the addx, every time the fraction wraps around, the X bit is set
|> >>áand next add, the integer bit is incremented.
|> >Addx is really great .. Hmm. We could think of all the
|> >situations where addx could possibly go wrong. (I mean
|> >fixed point interpolations)
|> >At least this has to be done before the loop:
|> >        moveq   #0,d0
|> >        add.l   d1,d0
|>
|> >        ...
|> >.loop   addx.l  d1,d2
|> >        dbf     d7,.loop
|>
|> >But what about if there is two addx? Theis decimal parts
|> >have to be switched, which is done nicely with eor
|> >before swapping, but what about that before loop
|> >correction? Of course this perfect accuracy
|> >is not required on 16.16 fixed point but if
|> >there is a smaller amount of fraction, the
|> >errors can be quite big..
|>
|> Using
|>         sub.w   dx,dx
|>
|> instead of
|>
|>         moveq   #0,d0
|>         add.l   d1,d0

what about cmp.x dx,dx
won't trash dx ;) the aim was to get X flag 0 right ?

mhm in my routine I had to do a previous add in the outer loop, it did
same addition like the inner, BUT only on lo fraction.

So I cause the X that actually will be needed by the inner loop addx
(Without previous add, the addx will cause the X it would have needed
before!).

Clearing X maybe gives less noise, but not correct :)
hope I'm talking about same things like you... ;)

|>
|> is faster by-the-way. And since we are only using the lower word of the
|> register we don't need to use an extra dataregister.

sub.b dx,dx would even have trashed only the lower _byte_ ;)
but all instructions tend to need lower byte ;)

|>
|> <sb>Ludde - Amiga Demo Coder
|> <sb>Virtual Reality & Official Be developer
|> <sb>[e-mail]
|>
------------------------------------------------------------------------
   [e-mail] (Juergen "Rally" Fischer)   =:)