registers. Make sure liveinterval analysis is correctly creating live ranges
for them.

llvm-svn: 34217

comparisons.  This fixes an infinite loop on CodeGen/Generic/switch-lower.ll
and PR1197

llvm-svn: 34216

llvm-svn: 34214

llvm-svn: 34213

llvm-svn: 34212

llvm-svn: 34211

llvm-svn: 34205

llvm-svn: 34204

llvm-svn: 34202

2. Add functions to support the numberical conversion between APInt and
   double/float.

llvm-svn: 34201

module.  This speeds up the bcreader 11%.

llvm-svn: 34198

llvm-svn: 34197

as its main datastructure.  There are many improvements yet to be made, but
this speeds up opt --std-compile-opts on 447.dealII by 7.3%.

llvm-svn: 34193

llvm-svn: 34188

llvm-svn: 34187

llvm-svn: 34185

llvm-svn: 34183

llvm-svn: 34182

llvm-svn: 34180

llvm-svn: 34179

llvm-svn: 34177

llvm-svn: 34176

llvm-svn: 34175

llvm-svn: 34174

llvm-svn: 34173

llvm-svn: 34172

llvm-svn: 34171

the bc reader by 0.8%. Deemed "not worth it"

llvm-svn: 34169

by 6% on tramp3d.

llvm-svn: 34167

llvm-svn: 34166

llvm-svn: 34165

llvm-svn: 34163

llvm-svn: 34162

llvm-svn: 34161

-load-vn -gcse by 2.3%.

llvm-svn: 34160

llvm-svn: 34159

after the StructLayout object in memory.  This marginally improves locality,
speeding up -load-vn -gcse by ~0.8%.

llvm-svn: 34158

llvm-svn: 34157

llvm-svn: 34156

llvm-svn: 34154