llvm/test/Transforms
Owen Anderson 7a0201c6a6 Remove an InstCombine that transformed patterns like (x * uitofp i1 y) to (select y, x, 0.0) when the multiply has fast math flags set.
While this might seem like an obvious canonicalization, there is one subtle problem with it.  The result of the original expression
is undef when x is NaN (remember, fast math flags), but the result of the select is always defined when x is NaN.  This means that the
new expression is strictly more defined than the original one.  One unfortunate consequence of this is that the transform is not reversible!
It's always legal to make increase the defined-ness of an expression, but it's not legal to reduce it.  Thus, targets that prefer the original
form of the expression cannot reverse the transform to recover it.  Another way to think of it is that the transform has lost source-level
information (the fast math flags), which is undesirable.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@215825 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-17 03:51:29 +00:00
..
ADCE
AddDiscriminators
ArgumentPromotion ArgPromo+DebugInfo: Handle updating debug info over multiple applications of argument promotion. 2014-07-23 22:09:29 +00:00
AtomicExpandLoadLinked/ARM CodeGen: it turns out that NAND is not the same thing as BIC. At all. 2014-07-07 09:06:35 +00:00
BBVectorize Reduce verbiage of lit.local.cfg files 2014-06-09 22:42:55 +00:00
BranchFolding
CodeExtractor
CodeGenPrepare CodeGenPrep: fall back to MVT::Other if instruction's type isn't an EVT. 2014-07-29 10:20:22 +00:00
ConstantHoisting [ConstantHoisting][X86] Improve the cost model for small constants with large types (i64 and above). 2014-06-10 00:32:29 +00:00
ConstantMerge
ConstProp Look through addrspacecast in IsConstantOffsetFromGlobal 2014-07-14 22:39:26 +00:00
CorrelatedValuePropagation
DeadArgElim
DeadStoreElimination Revert "r214897 - Remove dead zero store to calloc initialized memory" 2014-08-06 19:30:38 +00:00
DebugIR
EarlyCSE
FunctionAttrs [optnone] Make the optnone attribute effective at suppressing function 2014-08-13 10:49:33 +00:00
GCOVProfiling Fix coverage for files with global constructors again. Adds a testcase to the commit from r206671, as requested by David Blaikie. 2014-06-05 04:31:43 +00:00
GlobalDCE Use "weak alias" instead of "alias weak" 2014-07-30 22:51:54 +00:00
GlobalOpt GlobalOpt: Optimize in the face of insertvalue/extractvalue 2014-08-08 05:50:43 +00:00
GVN Use "weak alias" instead of "alias weak" 2014-07-30 22:51:54 +00:00
IndVarSimplify Add a description to the test from r211433 explaining why it's written that way. 2014-06-22 12:22:04 +00:00
Inline Copy noalias metadata from call sites to inlined instructions 2014-08-14 21:09:37 +00:00
InstCombine Remove an InstCombine that transformed patterns like (x * uitofp i1 y) to (select y, x, 0.0) when the multiply has fast math flags set. 2014-08-17 03:51:29 +00:00
InstMerge MergedLoadStoreMotion pass 2014-07-18 19:13:09 +00:00
InstSimplify InstSimplify: Simplify (X - (0 - Y)) if the second sub is NUW 2014-07-31 04:49:18 +00:00
Internalize Use "weak alias" instead of "alias weak" 2014-07-30 22:51:54 +00:00
IPConstantProp No need for those tests to go thru llvm-as and/or llvm-dis. 2014-05-27 22:03:28 +00:00
JumpThreading In LVI(Lazy Value Info), originally value on a BB can only be caculated once, 2014-08-11 05:02:04 +00:00
LCSSA
LICM Make Value::isDereferenceablePointer handle offsets to pointer types with dereferenceable attributes 2014-07-19 03:25:16 +00:00
LoadCombine Add LoadCombine pass. 2014-05-29 01:55:07 +00:00
LoopDeletion
LoopIdiom R600: Implement TTI:getPopcntSupport 2014-07-18 06:07:13 +00:00
LoopReroll
LoopRotate
LoopSimplify No need for those tests to go thru llvm-as and/or llvm-dis. 2014-05-27 22:03:28 +00:00
LoopStrengthReduce Reduce verbiage of lit.local.cfg files 2014-06-09 22:42:55 +00:00
LoopUnroll After unrolling a loop with llvm.loop.unroll.count metadata (unroll factor 2014-07-24 22:36:40 +00:00
LoopUnswitch
LoopVectorize [LoopVectorizer] Enable support for floating-point subtraction reductions 2014-08-08 12:41:08 +00:00
LowerAtomic IR: add "cmpxchg weak" variant to support permitted failure. 2014-06-13 14:24:07 +00:00
LowerExpectIntrinsic
LowerInvoke
LowerSwitch Added test for commit r212802 that was missing 2014-07-11 10:36:00 +00:00
Mem2Reg
MemCpyOpt Don't eliminate memcpy's when the address of the pointer may itself be relevant. Fixes PR18304. Patch by David Wiberg! 2014-07-14 18:52:02 +00:00
MergeFunc MergeFunc patch from Björn Steinbrink. 2014-07-15 10:46:51 +00:00
MetaRenamer Use "weak alias" instead of "alias weak" 2014-07-30 22:51:54 +00:00
ObjCARC
PartiallyInlineLibCalls PartiallyInlineLibCalls: Check sqrt result type before transforming it. 2014-08-01 23:21:21 +00:00
PhaseOrdering
PruneEH
Reassociate [Reassociation] Add support for reassociation with unsafe algebra. 2014-08-14 15:23:01 +00:00
Reg2Mem
SampleProfile
Scalarizer
ScalarRepl
SCCP SCCP: update for cmpxchg returning { iN, i1 } now. 2014-06-13 14:54:09 +00:00
SeparateConstOffsetFromGEP/NVPTX Partially revert r210444 due to performance regression 2014-07-16 23:25:00 +00:00
SimplifyCFG [SimplifyCFG] fix accessing deleted PHINodes in switch-to-table conversion. 2014-08-02 23:41:54 +00:00
Sink
SLPVectorizer SLPVectorizer: Use the type of the value loaded/stored to get the ABI alignment 2014-08-07 22:47:27 +00:00
SROA Fix a case in SROA where lifetime intrinsics could inhibit alloca promotion. In 2014-08-07 21:07:35 +00:00
StripSymbols
StructurizeCFG
TailCallElim We may visit a call that uses an alloca multiple times in callUsesLocalStack, sometimes with IsNocapture true and sometimes with IsNocapture false. We accidentally skipped work we needed to do in the IsNocapture=false case if we were called with IsNocapture=true the first time. Fixes PR20405! 2014-07-23 06:24:49 +00:00
TailDup Reduce verbiage of lit.local.cfg files 2014-06-09 22:42:55 +00:00
Util utils: Fix segfault in flattencfg 2014-08-13 20:31:53 +00:00