James Molloy 72035e9a8e Teach the SLP Vectorizer that keeping some values live over a callsite can have a cost. ... Some types, such as 128-bit vector types on AArch64, don't have any callee-saved registers. So if a value needs to stay live over a callsite, it must be spilled and refilled. This cost is now taken into account. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@214859 91177308-0d34-0410-b5e6-96231b3b80d8		2014-08-05 12:30:34 +00:00
..
ADCE
AddDiscriminators
ArgumentPromotion
AtomicExpandLoadLinked/ARM
BBVectorize
BranchFolding
CodeExtractor
CodeGenPrepare
ConstantHoisting
ConstantMerge
ConstProp
CorrelatedValuePropagation
DeadArgElim
DeadStoreElimination
DebugIR
EarlyCSE
FunctionAttrs
GCOVProfiling
GlobalDCE
GlobalOpt
GVN
IndVarSimplify
Inline
InstCombine	This patch implements transform for pattern "(A & ~B) ^ (~A) -> ~(A & B)".	2014-08-01 05:07:20 +00:00
InstMerge
InstSimplify
Internalize
IPConstantProp
JumpThreading
LCSSA
LICM
LoadCombine
LoopDeletion
LoopIdiom
LoopReroll
LoopRotate
LoopSimplify
LoopStrengthReduce
LoopUnroll
LoopUnswitch
LoopVectorize	Add diagnostics to the vectorizer cost model.	2014-08-02 00:14:03 +00:00
LowerAtomic
LowerExpectIntrinsic
LowerInvoke
LowerSwitch
Mem2Reg
MemCpyOpt
MergeFunc
MetaRenamer
ObjCARC
PartiallyInlineLibCalls	PartiallyInlineLibCalls: Check sqrt result type before transforming it.	2014-08-01 23:21:21 +00:00
PhaseOrdering
PruneEH
Reassociate
Reg2Mem
SampleProfile
Scalarizer
ScalarRepl
SCCP
SeparateConstOffsetFromGEP/NVPTX
SimplifyCFG	[SimplifyCFG] fix accessing deleted PHINodes in switch-to-table conversion.	2014-08-02 23:41:54 +00:00
Sink
SLPVectorizer	Teach the SLP Vectorizer that keeping some values live over a callsite can have a cost.	2014-08-05 12:30:34 +00:00
SROA
StripSymbols
StructurizeCFG
TailCallElim
TailDup