llvm-svn: 101286

llvm-svn: 101182

llvm-svn: 101163

Re-apply 101075 and fix it properly. Just reuse the debug info of the branch instruction being optimized. There is no need to --I which can deref off start of the BB.

llvm-svn: 101162

in a nightly tester.

llvm-svn: 101158

patch by Sylvere Teissier!

llvm-svn: 101106

llvm-svn: 101085

llvm-svn: 101081

llvm-svn: 101079

llvm-svn: 101077

If we have this situation:

    jCC  L1
    jmp  L2
L1:
  ...
L2:
  ...

We can get a small performance boost by emitting this instead:

    jnCC L2
L1:
  ...
L2:
  ...

This testcase shows an example of this:

float func(float x, float y) {
    double product = (double)x * y;
    if (product == 0.0)
        return product;
    return product - 1.0;
}

llvm-svn: 101075

llvm-svn: 101023

llvm-svn: 100860

%RDI<def,dead> = MOV64rr %RAX<kill>, %EDI<imp-def>

llvm-svn: 100804

llvm-svn: 100751

ensure that the expansion is dominated by the increments of those loops.

llvm-svn: 100748

llvm-svn: 100705

explicitly split into stride-and-offset pairs. Also, add the
ability to track multiple post-increment loops on the same expression.

This refines the concept of "normalizing" SCEV expressions used for
to post-increment uses, and introduces a dedicated utility routine for
normalizing and denormalizing expressions.

This fixes the expansion of expressions which are post-increment users
of more than one loop at a time. More broadly, this takes LSR another
step closer to being able to reason about more than one loop at a time.

llvm-svn: 100699

those who don't build all targets.

llvm-svn: 100688

llvm-svn: 100637

MachineOperand::isIdenticalTo wasn't handling metadata operands.

llvm-svn: 100636

This fixes the Bullet regression on i386/nocona.

llvm-svn: 100553

llvm-svn: 100482

llvm-svn: 100455

llvm-svn: 100346

llvm-svn: 100208

in particular, they end up aligning strings at 16-byte boundaries, and
there's no way for GlobalOpt to check OptForSize.

llvm-svn: 100172

unaligned loads into aligned loads.

llvm-svn: 100166

llvm-svn: 100165

llvm-svn: 100137

- Do not try to infer GV alignment unless its type is sized. It's not possible to infer alignment if it has opaque type.

llvm-svn: 100118

llvm-svn: 100101

1. Makes it possible to lower with floating point loads and stores.
2. Avoid unaligned loads / stores unless it's fast.
3. Fix some memcpy lowering logic bug related to when to optimize a
   load from constant string into a constant.
4. Adjust x86 memcpy lowering threshold to make it more sane.
5. Fix x86 target hook so it uses vector and floating point memory
   ops more effectively.
rdar://7774704

llvm-svn: 100090

llvm-svn: 99975

llvm-svn: 99974

llvm-svn: 99959

llvm-svn: 99954

Rewrite the pmulld patterns, and make sure that they fold in loads of
arguments into the instruction.

llvm-svn: 99910

transforming it into (add (i32 GPR), 4).  This allows us to write type
generic multi patterns and have tblgen automatically drop the bitconvert
in the case when the types align.  This allows us to fold an extra load
in the changed testcase.

llvm-svn: 99756

llvm-svn: 99620