Ahmed Bougacha cd5bbd8bad [X86] Also create+widen FMIN/FMAX nodes for v2f32. ... This happens in the HINT benchmark, where the SLP-vectorizer created v2f32 fcmp/select code. The "correct" solution would have been to teach the vectorizer cost model that v2f32 isn't legal (because really, it isn't), but if we can vectorize we might as well do so. We legalize these v2f32 FMIN/FMAX nodes by widening to v4f32 later on. v3f32 were already widened to v4f32 by the generic unroll-and-build-vector legalization. rdar://15763436 Differential Revision: http://reviews.llvm.org/D6557 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225691 91177308-0d34-0410-b5e6-96231b3b80d8		2015-01-12 20:31:30 +00:00
..
Analysis	Fix PR22179.	2015-01-10 23:41:24 +00:00
AsmParser	IR: Split GenericMDNode into MDTuple and UniquableMDNode	2015-01-12 20:09:34 +00:00
Bitcode	IR: Split GenericMDNode into MDTuple and UniquableMDNode	2015-01-12 20:09:34 +00:00
CodeGen	RegisterCoalescer: Turn some impossible conditions into asserts	2015-01-12 19:10:17 +00:00
DebugInfo	Implement a very basic colored syntax highlighting for llvm-dwarfdump.	2015-01-06 16:50:25 +00:00
ExecutionEngine	Fix undefined behavior (shift of negative value) in RuntimeDyldMachOAArch64::encodeAddend.	2015-01-10 00:46:38 +00:00
IR	IR: Move creation logic to MDNodeFwdDecl, NFC	2015-01-12 20:21:37 +00:00
IRReader
LineEditor
Linker	Use the DiagnosticHandler to print diagnostics when reading bitcode.	2015-01-10 00:07:30 +00:00
LTO	Use the DiagnosticHandler to print diagnostics when reading bitcode.	2015-01-10 00:07:30 +00:00
MC	Add r224985 back with two fixes.	2015-01-12 18:13:07 +00:00
Object	Don't loop endlessly for MachO files with 0 ncmds	2015-01-06 17:08:26 +00:00
Option
ProfileData
Support	Triple: Add amdgcn triple	2015-01-06 18:00:00 +00:00
TableGen
Target	[X86] Also create+widen FMIN/FMAX nodes for v2f32.	2015-01-12 20:31:30 +00:00
Transforms	IR: Split GenericMDNode into MDTuple and UniquableMDNode	2015-01-12 20:09:34 +00:00
CMakeLists.txt
LLVMBuild.txt
Makefile