Avoid if-converting simple block that ends with unconditional branch or fallthrough unless it branches / falls to the 'false' block. Not profitable, may end up increasing code size.

llvm-svn: 37660

llvm-svn: 37659

deleteValueFromRecords and loosen the types to all it to accept
Value* instead of just Instruction*, since this is what
ScalarEvolution uses internally anyway. This allows more flexibility
for future uses.

llvm-svn: 37657

for needing the DAG node to print pre-legalize extended value types, and
to get better debug messages with target-specific nodes.

llvm-svn: 37656

Handle constants in phi nodes properly.  This fixes test/Transforms/GVNPRE/2007-06-18-ConstantInPhi.ll

llvm-svn: 37655

llvm-svn: 37651

llvm-svn: 37650

llvm-svn: 37649

Be careful to erase values from all of the appropriate sets when they're not needed anymore.  This fixes a few more memory-related issues.

llvm-svn: 37647

Thanks to Nicholas for valgrinding this.

llvm-svn: 37646

Refactor GVNPRE to use a much smart method of uniquing value sets, and centralize a lot of the value numbering information.  No functionality change.

llvm-svn: 37645

with a general target hook to identify rematerializable instructions. Some
instructions are only rematerializable with specific operands, such as loads
from constant pools, while others are always rematerializable. This hook
allows both to be identified as being rematerializable with the same
mechanism.

llvm-svn: 37644

llvm-svn: 37643

Anton for half of this patch.

llvm-svn: 37641

llvm-svn: 37640

Look for VECTOR_SHUFFLE that's identity operation on either LHS or RHS. This can happen before DAGCombiner catches it.

llvm-svn: 37636

llvm-svn: 37634

llvm-svn: 37633

llvm-svn: 37630

by making it non-recursive

llvm-svn: 37629

add operands after constant operands. The recent change to recognize
sign-extend expressions caused this to be exposed more often.

llvm-svn: 37628

llvm-svn: 37627

llvm-svn: 37626

llvm-svn: 37625

llvm-svn: 37623

llvm-svn: 37622

llvm-svn: 37621

llvm-svn: 37620

define double @test2(i64 %A) {
   %B = bitcast i64 %A to double
   ret double %B
}

$ llvm-as < t.ll | llc -march=x86-64

before:

         .align  4
         .globl  _test2
_test2:
         movd %rdi, %xmm0
         ret

after:

_test2:
         subq $8, %rsp
         movq %rdi, (%rsp)
         movsd (%rsp), %xmm0
         addq $8, %rsp
         ret

llvm-svn: 37617

1. Consider all possible ifcvt cases at once. No longer restricted to bottom
   up iterative approach.
2. Sort all possible cases based on a cost function. Perform the most profitable
   ones first invalidate others that target the same blocks.
3. Fixed a number of bugs related to block duplication.

llvm-svn: 37613

llvm-svn: 37611

Fix test/Transforms/GVNPRE/2007-06-15-InvokeInst.ll by ignoring all instructions that depend on invokes.

llvm-svn: 37610

llvm-svn: 37608

llvm-svn: 37607

llvm-svn: 37606

llvm-svn: 37602

for tied register constraints.

llvm-svn: 37601

llvm-svn: 37599

llvm-svn: 37597

integer constant SCEV.

llvm-svn: 37596