basic block.  This is amazingly common in code generated by the C/C++ front-ends.
This change makes it not have to insert ANY phi nodes, whereas before it would insert
a ton of dead ones which DCE would have to clean up.

Thus, this fix improves compile-time performance of these trivial allocas in two ways:
  1. It doesn't have to do the walking and book-keeping for renaming
  2. It does not insert dead phi nodes for them which would have to
     subsequently be cleaned up.

On my favorite testcase from 252.eon, this special case handles 305 out of
356 promoted allocas in the function.  It speeds up the mem2reg pass from 7.5256s
to 1.2505s.  It inserts 677 fewer dead PHI nodes, which speeds up a subsequent
-dce pass from 18.7524s to 2.4806s.

There are still 120 trivially dead PHI nodes being inserted for variables used
in multiple basic blocks, but they are not handled by this patch.

llvm-svn: 8881

llvm-svn: 8865

The VersionNumbers vector is only used during PHI placement.  Turn it into an argument, allowing us to get rid of the vector.

llvm-svn: 8864

*** Revamp the code which handled unreachable code in the function.  Now the
    code is much more efficient for high-degree basic blocks, such as those
    that occur in the 252.eon SPEC benchmark.

For the interested, the time to promote a SINGLE alloca in _ZN7mrScene4ReadERSi
function used to be > 3.5s.  Now it is < .075s.  The function has a LOT of
allocas in it, so it appeared to be infinite looping, this should make it much
nicer.  :)

llvm-svn: 8863

llvm-svn: 8862

work-list of value definitions.  This allows elimination of the explicit
'iterative' step of the algorithm, and also reuses temporary memory better.

llvm-svn: 8861

need to be an instance variable!

llvm-svn: 8860

* Fuse two parallel loops
* Use a more specific type for AllocaLookup

llvm-svn: 8859

 * Do not insert a new entry into NewPhiNodes during the rename pass if there are no PHIs in a block.
 * Do not compute WriteSets in parallel

llvm-svn: 8858

* Eliminate the KillList instance variable, instead, just delete loads and
  stores as they are "renamed", and delete allocas when they are done
* Make the 'visited' set an instance variable to avoid passing it on the stack.

llvm-svn: 8857

llvm-svn: 8667

llvm-svn: 8666

... by making sure to update PHI nodes to take into consideration the
extra edges we get if we inline a call instruction through an invoke.

llvm-svn: 8664

Rename Function::getEntryNode -> getEntryBlock

llvm-svn: 8623

llvm-svn: 8514

llvm-svn: 8456

llvm-svn: 8411

llvm-svn: 8409

Fix bug where we considered function types equivalent even if they had differing numbers of arguments

llvm-svn: 8178

llvm-svn: 8144

llvm-svn: 8130

llvm-svn: 8129

llvm-svn: 8126

in terms of it and llvm.exc.setcurrent.
Rework the intrinsics.

llvm-svn: 8111

llvm-svn: 8106

llvm-svn: 8099

llvm-svn: 8093

llvm-svn: 8084

llvm-svn: 8083

llvm-svn: 8080

llvm-svn: 8059

The process of linking types can cause their addresses to become invalid.  For this reason, we must use TypeHandles!

llvm-svn: 8057

Hopefully this will fix the 176.gcc spec test as well.

llvm-svn: 8051

llvm-svn: 7944

Implements SimplifyCFG/2003-08-17-FoldSwitch.ll

llvm-svn: 7923

llvm-svn: 7921

This fixes testcase: SimplifyCFG/2003-08-17-BranchFold.ll

llvm-svn: 7919

Fix bug: SimplifyCFG/2003-08-05-MishandleInvoke.ll

llvm-svn: 7599

llvm-svn: 6476

internal phi nodes, and returns a new vector of basic blocks.

llvm-svn: 6431