Subject: [linux-audio-dev] runtime / compiletime processing proof of concept
From: Maarten de Boer (mdeboer_AT_iua.upf.es)
Date: Fri Jul 26 2002 - 14:21:46 EEST
Hello,
Juan Linietsky (reduz on #lad) talked on #lad about an idea of
Benno to have processing graphs generate source code, compile
it on the fly, thus taking advantage of all compiled optimizations,
as an alternative to runtime processing. Especially for sample
based processing instead of block based processing this can be
very useful and a lot faster.
It's an idea that I've had before, so it got me thinking. Wouldn't
it be nice to have a both-worlds solution? Do the prototyping in
runtime, and generate the source code and compile it in a shared
object once you have it running.
The attached .tgz (it's very small so I thought attaching it would
be okay) contains a proof of concept that does simple FM. There are
two version main_educ.cxx, and main.cxx. The former generates slightly
more 'educational' source, the second more optimized. Run it and look
at the generated source to see what I mean. The Makefile does it
all. Try also make time and make play.
Mind you, it's a proof of concept, and an extremely quick hack :-)
Maarten
This archive was generated by hypermail 2b28 : Fri Jul 26 2002 - 14:31:43 EEST