Andrea Di Biagio ae16ff1c42 [X86] Improved lowering of packed v8i16 vector shifts by non-constant count. ... Before this patch, the backend sub-optimally expanded the non-constant shift count of a v8i16 shift into a sequence of two 'movd' plus 'movzwl'. With this patch the backend checks if the target features sse4.1. If so, then it lets the shuffle legalizer deal with the expansion of the shift amount. Example: ;; define <8 x i16> @test(<8 x i16> %A, <8 x i16> %B) { %shamt = shufflevector <8 x i16> %B, <8 x i16> undef, <8 x i32> zeroinitializer %shl = shl <8 x i16> %A, %shamt ret <8 x i16> %shl } ;; Before (with -mattr=+avx): vmovd %xmm1, %eax movzwl %ax, %eax vmovd %eax, %xmm1 vpsllw %xmm1, %xmm0, %xmm0 retq Now: vpxor %xmm2, %xmm2, %xmm2 vpblendw $1, %xmm1, %xmm2, %xmm1 vpsllw %xmm1, %xmm0, %xmm0 retq git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@223660 91177308-0d34-0410-b5e6-96231b3b80d8		2014-12-08 14:36:51 +00:00
..
Analysis	Revert a part of r223583, for now. It seems causing different emission between stage2(gcc-clang) and stage3 clang. Investigating.	2014-12-08 02:07:22 +00:00
AsmParser	IR: Add missing tests for function-local metadata	2014-12-07 17:56:16 +00:00
Bitcode
CodeGen
DebugInfo
ExecutionEngine
IR	IR: Revert r223618 behaviour of MDNode::concatenate()	2014-12-07 20:32:11 +00:00
IRReader
LineEditor
Linker	Move the ValueMap lookup inside linkFunctionBody. NFC.	2014-12-08 14:25:26 +00:00
LTO
MC
Object
Option
ProfileData
Support
TableGen
Target	[X86] Improved lowering of packed v8i16 vector shifts by non-constant count.	2014-12-08 14:36:51 +00:00
Transforms
CMakeLists.txt
LLVMBuild.txt
Makefile