Andrea Di Biagio 8a712ba229 [X86] Improve the lowering of BITCAST dag nodes from type f64 to type v2i32 (and vice versa). ... Before this patch, the backend always emitted a store+load sequence to bitconvert from f64 to i64 the input operand of a ISD::BITCAST dag node that performed a bitconvert from type MVT::f64 to type MVT::v2i32. The resulting i64 node was then used to build a v2i32 vector. With this patch, the backend now produces a cheaper SCALAR_TO_VECTOR from MVT::f64 to MVT::v2f64. That SCALAR_TO_VECTOR is then followed by a "free" bitcast to type MVT::v4i32. The elements of the resulting v4i32 are then extracted to build a v2i32 vector (which is illegal and therefore promoted to MVT::v2i64). This is in general cheaper than emitting a stack store+load sequence to bitconvert the operand from type f64 to type i64. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@208107 91177308-0d34-0410-b5e6-96231b3b80d8		2014-05-06 17:09:03 +00:00
..
Analysis	blockfreq: Move include to .cpp	2014-05-06 01:57:42 +00:00
AsmParser
Bitcode
CodeGen	Implememting named register intrinsics	2014-05-06 16:51:25 +00:00
DebugInfo
ExecutionEngine
IR	Special case aliases in GlobalValue::getAlignment.	2014-05-06 16:48:58 +00:00
IRReader
LineEditor
Linker	Be more strict about not calling setAlignment on global aliases.	2014-05-06 14:51:36 +00:00
LTO	Use a range loop.	2014-05-05 20:06:41 +00:00
MC	Have the SubtargetFeature help routine just not return a number and	2014-05-06 16:29:50 +00:00
Object	[ELFYAML] Group ELF header falgs to target specific blocks. Handle flags	2014-05-03 11:39:50 +00:00
Option
ProfileData
Support	Re-commit r208025, reverted in r208030, with a fix for a conformance issue	2014-05-06 01:44:26 +00:00
TableGen
Target	[X86] Improve the lowering of BITCAST dag nodes from type f64 to type v2i32 (and vice versa).	2014-05-06 17:09:03 +00:00
Transforms	Re-commit r208025, reverted in r208030, with a fix for a conformance issue	2014-05-06 01:44:26 +00:00
CMakeLists.txt
LLVMBuild.txt
Makefile