Scheduling complete loop unrolling early and unconditional?

[Richard Guenther]

Hi!

I seem to get significant better code by exchanging code like

 for (int i=0; i<4; ++i)
   a[i] = b[i] + c[i];

with a template meta-program.  This seems to be because SRA
cannot scalarize the arrays if they are accessed in a loop.

Therefore, is it possible to schedule either complete unrolling
of loops very early (before sra), or to schedule another sra
run after loop optimizations (that's what I'll try at the moment,
but possibly mess everything up)?

Again, can we have complete loop unrolling enabled unconditionally?

Thanks,
Richard.

--
Richard Guenther <richard dot guenther at uni-tuebingen dot de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/