llvm-6502/test/Transforms/Reassociate
David Majnemer 48e2671cb6 [Reassociate] Canonicalizing 'x [+-] (-Constant * y)' isn't always a win
Canonicalizing 'x [+-] (-Constant * y)' is not a win if we don't *know*
we will open up CSE opportunities.

If the multiply was 'nsw', then negating 'y' requires us to clear the
'nsw' flag.  If this is actually worth pursuing, it is probably more
appropriate to do so in GVN or EarlyCSE.

This fixes PR23675.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@238397 91177308-0d34-0410-b5e6-96231b3b80d8
2015-05-28 06:16:39 +00:00
..
2002-05-15-AgressiveSubMove.ll
2002-05-15-MissedTree.ll
2002-05-15-SubReassociate.ll
2002-07-09-DominanceProblem.ll
2003-08-12-InfiniteLoop.ll
2005-09-01-ArrayOutOfBounds.ll
2006-04-27-ReassociateVector.ll
2011-01-26-UseAfterFree.ll
2012-05-08-UndefLeak.ll
2012-06-08-InfiniteLoop.ll
absorption.ll
basictest.ll
canonicalize-neg-const.ll
commute.ll
crash2.ll
crash.ll
fast-AgressiveSubMove.ll
fast-ArrayOutOfBounds.ll
fast-basictest.ll
fast-fp-commute.ll
fast-mightymul.ll
fast-MissedTree.ll
fast-multistep.ll
fast-ReassociateVector.ll
fast-SubReassociate.ll
fp-commute.ll
inverses.ll
looptest.ll
mightymul.ll
min_int.ll
mixed-fast-nonfast-fp.ll
mulfactor.ll
multistep.ll
negation1.ll
negation.ll
no-op.ll
optional-flags.ll
otherops.ll
pr12245.ll
pr21205.ll
repeats.ll
secondary.ll
shift-factor.ll
shifttest.ll
subtest.ll
wrap-flags.ll
xor_reassoc.ll