llvm-svn: 5410

out of a large function to reduce it.

llvm-svn: 5408

llvm-svn: 5407

llvm-svn: 5406

llvm-svn: 5405

llvm-svn: 5404

llvm-svn: 5403

link with static versions of some libs to avoid having to pull in all of ipo/ipa and datastructure at all

llvm-svn: 5402

llvm-svn: 5401

llvm-svn: 5399

llvm-svn: 5397

llvm-svn: 5396

llvm-svn: 5395

llvm-svn: 5394

llvm-svn: 5386

llvm-svn: 5385

llvm-svn: 5380

llvm-svn: 5379

llvm-svn: 5378

llvm-svn: 5377

llvm-svn: 5376

* Prune down library list.  In particular, llc doesn't need all of IPO, and none of data structure and IPA completely!
* Do not keep symbols by default!

llvm-svn: 5375

llvm-svn: 5361

llvm-svn: 5360

llvm-svn: 5357

llvm-svn: 5352

output upon make clean

llvm-svn: 5347

llvm-svn: 5342

llvm-svn: 5333

llvm-svn: 5332

llvm-svn: 5330

llvm-svn: 5329

llvm-svn: 5328

llvm-svn: 5327

llvm-svn: 5326

llvm-svn: 5324

llvm-svn: 5323

llvm-svn: 5322

llvm-svn: 5321

This improves the performance of the power benchmark by a few percent.
This will be neccesary for SSE code, which requires 16 byte alignment of
the stack.

llvm-svn: 5320