>>>>> "P" == Patrick De Smet <[EMAIL PROTECTED]> writes:
P> window_subband; why did he/you move away from pointers, aren't
P> arrays less efficient i.e. is win[i+16] not compiled less
P> efficient than *ep0++
P> And when using pointers why not different y[i] loops;
>>> see http://telin.rug.ac.be/~pds/mpeg_audio/xlame/here.html
P> Of course then you would need a lot of pointers, but still ??
Because on my architecture, (intel celeron 464MHz + gcc2.95.1), arrays
are more efficient than lots of pointers. I think most currently
available MPUs are too. I think it is because most of them don't have
post-increament / pre-decreament addressing mode.
Currently, GCC is wise enough. storing y[i] will
---
Takehiro TOMINAGA // may the source be with you!
# latest snapshot available at
# http://www3.plala.or.jp/initialT/lame-beta/lame-1013.tar.bz2
# or http://www.isoternet.org/~tominaga/lame-beta/lame-1013.tar.bz2
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
