Evan Cheng d51425a82d Use movaps / movapd (instead of movss / movsd) to do FR32 / FR64 reg to reg ... transfer. According to the Intel P4 Optimization Manual: Moves that write a portion of a register can introduce unwanted dependences. The movsd reg, reg instruction writes only the bottom 64 bits of a register, not to all 128 bits. This introduces a dependence on the preceding instruction that produces the upper 64 bits (even if those bits are not longer wanted). The dependence inhibits register renaming, and thereby reduces parallelism. Not to mention movaps is shorter than movss. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26226 91177308-0d34-0410-b5e6-96231b3b80d8		2006-02-16 01:50:02 +00:00
..
Analysis	add an assert	2006-02-14 20:14:17 +00:00
Archive	more C++ daintiness	2005-12-26 14:31:26 +00:00
AsmParser	Convert this over to work with the new makefiles	2006-02-15 07:22:58 +00:00
Bytecode	Fix auto-upgrade of intrinsics to work properly with both assembly and	2006-01-27 11:49:27 +00:00
CodeGen	Remove an unused function parameter.	2006-02-15 22:12:35 +00:00
Debugger	dum de dum	2005-12-26 10:24:15 +00:00
ExecutionEngine	simplify this code now that each constant pool entry is not separately allocated	2006-02-09 04:49:59 +00:00
Linker	Rename method	2006-01-24 04:14:29 +00:00
Support	Let bugpoint work on sparc with v9 instructions enabled.	2006-02-04 05:02:27 +00:00
System	Add AddSymbol() method to DynamicLibrary to work around Windows limitation	2006-01-30 04:33:51 +00:00
Target	Use movaps / movapd (instead of movss / movsd) to do FR32 / FR64 reg to reg	2006-02-16 01:50:02 +00:00
Transforms	fix a bug where we unswitched the wrong way	2006-02-16 01:24:41 +00:00
VMCore	Another work around for the 'symbols with different types can have the same	2006-02-13 21:43:26 +00:00
Makefile	Add the Linker library	2004-11-14 21:54:41 +00:00