MRegisterInfo::isVirtualRegister.

llvm-svn: 11045

llvm-svn: 11044

llvm-svn: 11043

llvm-svn: 11041

llvm-svn: 11040

llvm-svn: 11038

method

llvm-svn: 11037

Fix a spello
Tighten up the assertion checking

No functionality changes.

llvm-svn: 11036

* Turn a bunch of instance variables into automatic variables

llvm-svn: 11035

Move Passes.h (which defines the interface to this file) to the top.
Move statistics to the top of the file.
Add a comment

llvm-svn: 11034

fixes the crash in 176.gcc.

llvm-svn: 11033

any test cases :-)

llvm-svn: 11032

of a linear search to find the first range for comparisons. This cuts
down the linear scan register allocator running time by a factor of 3
in 254.perlbmk and by a factor of 2.2 in 176.gcc.

llvm-svn: 11030

llvm-svn: 11029

Simplification of LiveIntervals::Interval::overlaps() and addition of
examples to overlaps() and liveAt() to make them clearer.

llvm-svn: 11028

llvm-svn: 11027

llvm-svn: 11026

llvm-svn: 11025

llvm-svn: 11024

llvm-svn: 11022

llvm-svn: 11021

FP_REG_KILL instructions at the end of blocks involved with critical edges.

Fix a bug where FP_REG_KILL instructions weren't inserted in fall through
unconditional branches.  Perhaps this will fix some linscan problems?

llvm-svn: 11019

Implement LiveVariables::getIndexMachineBasicBlock

llvm-svn: 11018

choose an ostream.

llvm-svn: 11016

llvm-svn: 11015

llvm-svn: 11006

llvm-svn: 11005

llvm-svn: 11003

llvm-svn: 11002

llvm-svn: 11001

llvm-svn: 11000

moving it to the start of removeDeadNodes.  This speeds up DSA by 2s on perlbmk
from 41s

llvm-svn: 10999

map.  This saves 5s in the TD pass, from 22->17s on perlbmk

llvm-svn: 10998

the globals directly.  This doesn't save any substantial time, however, because the
globals graph only contains globals!

llvm-svn: 10997

function to find the globals, iterate over all of the globals directly.  This
speeds the function up from 14s to 6.3s on perlbmk, reducing DSA time from
53->46s.

llvm-svn: 10996

timers by default

llvm-svn: 10993

This reduces the number of nodes allocated, then immediately merged and DNE'd
from 2193852 to 1298049.  unfortunately this only speeds DSA up by ~1.5s (of
53s), because it's spending most of its time waddling through the scalar map :(

llvm-svn: 10992

Also, use RC::merge when possible, reducing the number of nodes allocated, then immediately merged away from 2985444 to 2193852 on perlbmk.

llvm-svn: 10991

You gotta love spurious

llvm-svn: 10990

llvm-svn: 10989