Jakob Stoklund Olesen 22a1df6bf2 Keep track of which stage produced a live range, and bypass earlier stages when revisiting. ... This effectively disables the 'turbo' functionality of the greedy register allocator where all new live ranges created by splitting would be reconsidered as if they were originals. There are two reasons for doing this, 1. It guarantees that the algorithm terminates. Early versions were prone to infinite looping in certain corner cases. 2. It is a 2x speedup. We can skip a lot of unnecessary interference checks that won't lead to good splitting anyway. The problem is that region splitting only gets one shot, so it should probably be changed to target multiple physical registers at once. Local live range splitting is still 'turbo' enabled. It only accounts for a small fraction of compile time, so it is probably not necessary to do anything about that. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@126781 91177308-0d34-0410-b5e6-96231b3b80d8		2011-03-01 21:10:07 +00:00
..
Analysis	Optimize "icmp pred (urem X, Y), Y" --> true/false depending on pred. There's	2011-03-01 08:15:50 +00:00
Archive	Revert PathV2 changes, as sys::fs::unique_file is not finished yet.	2011-03-01 19:50:55 +00:00
AsmParser
Bitcode
CodeGen	Keep track of which stage produced a live range, and bypass earlier stages when revisiting.	2011-03-01 21:10:07 +00:00
CompilerDriver
ExecutionEngine
Linker
MC	Make all static functions become static class methods. Move shared (duplicated) functions to new MCELF class.	2011-02-28 21:45:04 +00:00
Object
Support
Target	Add datalayout information for the IEEE quad precision fp128 type.	2011-03-01 20:56:50 +00:00
Transforms	Make InstCombiner::FoldAndOfICmps create a ConstantRange that's the	2011-03-01 15:05:01 +00:00
VMCore	Added missing va_end().	2011-03-01 18:00:49 +00:00
CMakeLists.txt
Makefile