On Tue, Nov 18, 2014 at 06:38:49AM -0600, Segher Boessenkool wrote: > > Does it help pr52714 where we'd like to rip apart a PARALLEL with two > > sets, one of which is dead. If so, it might allow us to optimize that > > code better. > > It does not seem to fix the testcase. I'll investigate why not. > You're talking about the > > (parallel [(set (pc) (pc)) > (set (X) (sp))]) > > right? I guess the "set pc pc" is not marked as unused...
The very first thing that is checked for is !(added_sets_2 && i1 == 0) which matches in this case. I wonder what that condition is supposed to accomplish -- "optimisation" I suppose? Hacking that out, it still won't match because the "set pc pc" is not considered dead. It's a no-op, not the same thing. I'll try to widen the condition... Segher