Andrew Trick 31b941a60d Enable SCEV-based unrolling by default. ... This changes loop unrolling to use the same mechanism for trip count computation as indvars. This is a stronger check that tends to unroll more loops. A very common side-effect is that many single iteration loops will be removed sooner. The real goal was simply to remove dependence on canonical IVs. x86 is break even. ARM performance changes to expect (+ is good): External/SPEC/CFP2000/183.equake/183.equake +13% SingleSource/Benchmarks/Dhrystone/fldry +21% MultiSource/Applications/spiff/spiff +3% SingleSource/Benchmarks/Stanford/Puzzle -14% The Puzzle regression is actually an improvement in loop optimization that defeats GVN: rdar://problem/10065079. llvm-svn: 139009		2011-09-02 17:26:28 +00:00
clang	[analyzer] Remove TransferFuncs.h, then deal with the fallout.	2011-09-02 08:02:59 +00:00
compiler-rt	Change ARM vfp assembly functions to use unified syntax.	2011-08-23 16:40:18 +00:00
debuginfo-tests	Testcase for r133065	2011-06-15 17:57:23 +00:00
libcxx	Up version number by 1. This is a 'minor version' update.	2011-09-01 23:13:01 +00:00
libcxxabi	reformatted to match Clang style; thanks to John McCall for the nudge	2011-08-15 18:06:47 +00:00
lldb	Added alpha .debug_names and .debug_types support in the DWARF parser.	2011-09-02 04:03:59 +00:00
llvm	Enable SCEV-based unrolling by default.	2011-09-02 17:26:28 +00:00
polly	Fix crashes due to unaligned parameters	2011-08-25 08:40:59 +00:00