This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/11832] Optimization of common code in switch statements
- From: "aldot at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 30 Dec 2007 10:48:00 -0000
- Subject: [Bug rtl-optimization/11832] Optimization of common code in switch statements
- References: <bug-11832-5666@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #5 from aldot at gcc dot gnu dot org 2007-12-30 10:47 -------
Created an attachment (id=14842)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14842&action=view)
rough, half-finished, disfunctional thought; not a patch
The attached thoughts do not yet insert/remove stmt candidates but just record
possible candidates in the corresponding bitmaps.
Unfortunately this prooves that the current SSA-SCCN-VN does not work for the
testcase in comment #1 since antic_in is empty (but it should contain ++i).
Nevertheless this will potentially help for PR23286 and PR5738 since for those
testcases the hoisting-candidates are found properly.
I intend to think about finishing this draft as described by stevenb soonish
but still, i will have to finish up the inefficient variant that only cares
about switch stmts (perhaps defaulting to on at -Os, just for my personal use)
until SSA-SCCN-VN recognizes the VN for the testcase of #1
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11832