Hal Finkel 5ecd959a9e Add a TLI hook to control when BUILD_VECTOR might be expanded using shuffles ... There are two general methods for expanding a BUILD_VECTOR node: 1. Use SCALAR_TO_VECTOR on the defined scalar values and then shuffle them together. 2. Build the vector on the stack and then load it. Currently, we use a fixed heuristic: If there are only one or two unique defined values, then we attempt an expansion in terms of SCALAR_TO_VECTOR and vector shuffles (provided that the required shuffle mask is legal). Otherwise, always expand via the stack. Even when SCALAR_TO_VECTOR is not legal, this can still be a good idea depending on what tricks the target can play when lowering the resulting shuffle. If the target can't do anything special, however, and if SCALAR_TO_VECTOR is expanded via the stack, this heuristic leads to sub-optimal code (two stack loads instead of one). Because only the target knows whether the SCALAR_TO_VECTORs and shuffles for a build vector of a particular type are likely to be optimial, this adds a new TLI function: shouldExpandBuildVectorWithShuffles which takes the vector type and the count of unique defined values. If this function returns true, then method (1) will be used, subject to the constraint that all of the necessary shuffles are legal (as determined by isShuffleMaskLegal). If this function returns false, then method (2) is always used. This commit does not enhance the current code to support expanding a build_vector with more than two unique values using shuffles, but I'll commit an implementation of the more-general case shortly. llvm-svn: 205230		2014-03-31 17:48:10 +00:00
..
Analysis	PR15967 Fix in basicaa for faulty returning no alias.	2014-03-26 21:30:19 +00:00
AsmParser	Remove the linker_private and linker_private_weak linkages.	2014-03-13 23:18:37 +00:00
Bitcode	All new elements except the last one initialized to NULL. Ideally, once parsing is complete, all elements should be non-NULL.	2014-03-27 12:08:23 +00:00
CodeGen	Add a TLI hook to control when BUILD_VECTOR might be expanded using shuffles	2014-03-31 17:48:10 +00:00
DebugInfo	DebugInfo: Support debug_loc under fission	2014-03-25 01:44:02 +00:00
ExecutionEngine	[Allocator] Lift the slab size and size threshold into template	2014-03-30 12:07:07 +00:00
IR	Intrinsics: add LLVMHalfElementsVectorType constraint	2014-03-29 07:04:54 +00:00
IRReader	Replace OwningPtr<T> with std::unique_ptr<T>.	2014-03-06 05:51:42 +00:00
LineEditor	[CMake] Use LINK_LIBS instead of target_link_libraries().	2014-02-26 06:41:29 +00:00
Linker	[Layering] Sink Linker.h into a Linker subdirectory to make it	2014-03-06 03:42:23 +00:00
LTO	This patch fixes LTO's RecordStreamer so that it records symbols in the MCExpr	2014-03-31 16:59:13 +00:00
MC	[MC] Remove an unused (and broken) variant of the setupForSymbolicDisassembly	2014-03-30 04:27:33 +00:00
Object	[yaml2obj] Add support for ELF e_flags.	2014-03-31 09:44:05 +00:00
Option	Avoid buffer copies when a Twine already is a StringRef.	2013-12-03 18:18:28 +00:00
ProfileData	InstrProf: Silence spurious warnings in GCC 4.8	2014-03-24 00:47:18 +00:00
Support	Support: generalise object type handling for Windows	2014-03-31 16:34:41 +00:00
TableGen	tblgen: Twinify PrintFatalError.	2014-03-29 17:17:15 +00:00
Target	[mips] Rewrite MipsAsmParser and MipsOperand.	2014-03-31 17:43:46 +00:00
Transforms	Add a missing break.	2014-03-30 03:26:17 +00:00
CMakeLists.txt	ProfileData: Introduce the InstrProfReader interface and a text reader	2014-03-21 17:24:48 +00:00
LLVMBuild.txt	ProfileData: Introduce the InstrProfReader interface and a text reader	2014-03-21 17:24:48 +00:00
Makefile	ProfileData: Introduce the InstrProfReader interface and a text reader	2014-03-21 17:24:48 +00:00