Chris Lattner b01c24a945 Teach loop rotate to hoist trivially invariant instructions ... in the duplicated block instead of duplicating them. Duplicating them into the end of the loop and the preheader means that we got a phi node in the header of the loop, which prevented LICM from hoisting them. GVN would usually come around later and merge the duplicated instructions so we'd get reasonable output... except that anything dependent on the shoulda-been-hoisted value can't be hoisted. In PR5319 (which this fixes), a memory value didn't get promoted. llvm-svn: 113134		2010-09-06 01:10:22 +00:00
..
2009-01-25-SingleEntryPhi.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
LRCrash-1.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
LRCrash-2.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
LRCrash-3.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
LRCrash-4.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
LRCrash-5.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
PhiRename-1.ll	Delete useless trailing semicolons.	2010-01-05 17:55:26 +00:00
PhiSelfRefernce-1.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
basic.ll	Teach loop rotate to hoist trivially invariant instructions	2010-09-06 01:10:22 +00:00
dg.exp	sabre brings to my attention that the 'tr' suffix is also obsolete	2008-05-20 21:00:03 +00:00
indirectbr.ll	Make Loop::getLoopLatch() work on loops which don't have preheaders, as	2009-11-20 20:51:18 +00:00
phi-duplicate.ll	When rotating loops, put the original header at the bottom of the	2010-08-17 17:39:21 +00:00
pr2639.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00
preserve-scev.ll	Make these tests more interesting by using	2009-10-24 23:23:04 +00:00