This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH]: Fix PR tree-optimization/19038


Daniel Berlin wrote:

loop-im would have hoisted it if it was profitable, and then we'd have
copy-proped that version in dom3.

OK, so it's what I thought. We create a dependency on code hoisting. Not a big problem, I guess, but the comment is misleading.

Rather say something like: Do not propagate copies, if the propagated value is at a deeper loop depth. This may move loop variant variables outside of their loops and prevent coalescing opportunities. If the value was loop invariant, it will be hoisted by LICM and exposed for copy propagation.


Diego.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]