Cong Hou e4f9f1583a Detecte vector reduction operations just before instruction selection. ... (This is the second attemp to commit this patch, after fixing pr26652 & pr26653). This patch detects vector reductions before instruction selection. Vector reductions are vectorized reduction operations, and for such operations we have freedom to reorganize the elements of the result as long as the reduction of them stay unchanged. This will enable some reduction pattern recognition during instruction combine such as SAD/dot-product on X86. A flag is added to SDNodeFlags to mark those vector reduction nodes to be checked during instruction combine. To detect those vector reductions, we search def-use chains starting from the given instruction, and check if all uses fall into two categories: 1. Reduction with another vector. 2. Reduction on all elements. in which 2 is detected by recognizing the pattern that the loop vectorizer generates to reduce all elements in the vector outside of the loop, which includes several ShuffleVector and one ExtractElement instructions. Differential revision: http://reviews.llvm.org/D15250 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@261804 91177308-0d34-0410-b5e6-96231b3b80d8		2016-02-24 23:40:36 +00:00
..
CMakeLists.txt	Rename TargetSelectionDAGInfo into SelectionDAGTargetInfo and move it to CodeGen/	2016-01-27 16:32:26 +00:00
DAGCombiner.cpp	[DAGCombiner] Use getBitcast helper when possible. NFCI.	2016-02-20 15:05:29 +00:00
FastISel.cpp	[NFC] Replace several manual GEP loops with gep_type_iterator.	2016-01-20 00:26:52 +00:00
FunctionLoweringInfo.cpp	Undo spurious change made in r256965	2016-01-07 04:31:35 +00:00
InstrEmitter.cpp
InstrEmitter.h
LegalizeDAG.cpp	LegalizeDAG: Fix ExpandFCOPYSIGN assuming the same type on both inputs	2016-02-19 04:44:19 +00:00
LegalizeFloatTypes.cpp
LegalizeIntegerTypes.cpp	[X86] Don't transform X << 1 to X + X during type legalization	2016-01-28 18:20:05 +00:00
LegalizeTypes.cpp
LegalizeTypes.h
LegalizeTypesGeneric.cpp
LegalizeVectorOps.cpp	[CodeGen] Document and use getConstant's splat-building feature. NFC.	2016-02-15 18:07:29 +00:00
LegalizeVectorTypes.cpp
LLVMBuild.txt
ResourcePriorityQueue.cpp
ScheduleDAGFast.cpp
ScheduleDAGRRList.cpp	rangify; NFCI	2016-02-03 22:44:14 +00:00
ScheduleDAGSDNodes.cpp	Avoid overly large SmallPtrSet/SmallSet	2016-01-30 01:24:31 +00:00
ScheduleDAGSDNodes.h
ScheduleDAGVLIW.cpp
SDNodeDbgValue.h
SelectionDAG.cpp	SelectionDAG: Use correct addrspace when lowering memcpy	2016-02-22 22:01:42 +00:00
SelectionDAGBuilder.cpp	Detecte vector reduction operations just before instruction selection.	2016-02-24 23:40:36 +00:00
SelectionDAGBuilder.h	SelectionDAG: Lower some range metadata to AssertZext	2016-02-08 16:28:19 +00:00
SelectionDAGDumper.cpp	Avoid overly large SmallPtrSet/SmallSet	2016-01-30 01:24:31 +00:00
SelectionDAGISel.cpp	Fix Clang-tidy readability-redundant-control-flow warnings; other minor fixes.	2016-02-02 18:20:45 +00:00
SelectionDAGPrinter.cpp
SelectionDAGTargetInfo.cpp	Rename TargetSelectionDAGInfo into SelectionDAGTargetInfo and move it to CodeGen/	2016-01-27 16:32:26 +00:00
StatepointLowering.cpp	[StatepointLowering] Minor non-semantic cleanups	2016-02-19 19:37:07 +00:00
StatepointLowering.h	[StatepointLowering] Change AllocatedStackSlots to use SmallBitVector	2016-02-19 17:15:26 +00:00
TargetLowering.cpp	Remove uses of builtin comma operator.	2016-02-18 22:09:30 +00:00