Im trying to optimize my code, from the C level (specially schedulling the instruction in the C level).
I would like to know what are the key points that I have to keep in mind in order to make my code friendly to the gcc scheduler. I tried to modfy the code manually to increase parallelism, decrease data dependancy, etc....
but since gcc takes my c files and re-schedules everything, my optimizations are not transparent in the assembly file. And often my optimizations makes the program run slower.
Can anybody give a hint on this problem? Is there any documentation in the gcc scheduler?
I know that there are a lots of techniques in improving performance, but I'm interested in the
instrcution schedulling part.
thank you.
_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail
