llvm-svn: 4281

llvm-svn: 4280

llvm-svn: 4279

llvm-svn: 4278

llvm-svn: 4277

llvm-svn: 4276

* Fix typeo in the (unused) MachineInstrInfo::isArith method

llvm-svn: 4275

llvm-svn: 4274

llvm-svn: 4273

llvm-svn: 4272

llvm-svn: 4271

information and Loop info will not have to be recomputed after this runs.

llvm-svn: 4269

Make sure to have a pass name

llvm-svn: 4268

llvm-svn: 4267

    reading bytecode files with > 255 types in them, but only when optimization
    is enabled.  This was caused by GCC shrinking an enum to a single byte
    instead of a whole word.

llvm-svn: 4266

4 bytes at a time.  Change this to read blocks of 4k

llvm-svn: 4265

llvm-svn: 4264

llvm-svn: 4263

llvm-svn: 4262

    overhead: Merge 3 parallel vectors into 1, change regsUsed hash_set to be a
    bitvector.  Sped up LLC a little less than 10% in a debug build!

llvm-svn: 4261

    future.  This prevents having spurious map entries to nodes that we really
    don't care about.  This allows us to reenable the DeadNodeElim, which was
    disabled.

llvm-svn: 4260

llvm-svn: 4259

llvm-svn: 4258

llvm-svn: 4257

20 on intel or 24 on sparc

llvm-svn: 4256

    setPreservesCFG to be less confusing.

llvm-svn: 4255

llvm-svn: 4254

llvm-svn: 4253

    function was finally able to resolve the function call.  Adding this allows
    the TD pass to actually work right!
  - Temporarily disable dead node pruning.  This will be reenabled soon.

llvm-svn: 4252

llvm-svn: 4251

something that maps through a std::map.  Since this simplified the client and
implementation code, do so now.

llvm-svn: 4250

node explosion that doesn't help anything at all.  In previous versions of
the representation this DID help, but not anymore.

llvm-svn: 4249

be printed!

llvm-svn: 4248

llvm-svn: 4247

llvm-svn: 4246

global changes, so aren't good for headers.

llvm-svn: 4245

llvm-svn: 4244

llvm-svn: 4243

llvm-svn: 4242

call instructions

llvm-svn: 4241