home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!usenet.ins.cwru.edu!agate!sprite.berkeley.edu!shirriff
- From: shirriff@sprite.berkeley.edu (Ken Shirriff)
- Newsgroups: sci.fractals
- Subject: Re: Fractal FAQ (assembly in Xfractint?)
- Date: 20 Jan 1993 23:19:46 GMT
- Organization: University of California, Berkeley
- Lines: 17
- Message-ID: <1jkmmi$rah@agate.berkeley.edu>
- References: <fractal-faq_727397981@sprite.Berkeley.EDU> <AV.93Jan20131744@kielo.uta.fi>
- NNTP-Posting-Host: hijack.berkeley.edu
-
- In article <AV.93Jan20131744@kielo.uta.fi> av@uta.fi (Arto V. Viitanen) writes:
- >Does X/Fractint use assebler code on Intel based Unix machines ? If not, has
- >anybody tried to do this ?
-
- Xfractint is entirely C code. I suppose you could try using assembler code.
- However, I would recommend doing timings before putting much effort into
- it. On my machine, a float multiply takes about twice as long as an int
- multiply. So, after you factor in the shifting and so forth required to
- multiply two ints with an implicit scale factor, you don't really win.
- (Of course, if you don't have a floating point unit, then ints will win.)
- Also, the main cause for slowness in Xfractint is drawing pixels through X,
- and the assembly code won't help there.
-
- If anyone tries using assembler code for xfractint, please let me know how
- it turns out.
-
- Ken Shirriff shirriff@sprite.Berkeley.EDU
-