Optimize identical-comparisons based on propagated type information.

Optimize identical-comparisons based on propagated type information.

If the propagated type allows, eliminate the check for numbers at
identical operations in optimized code.

The previous optimizaton was done before full type propagation was run
and was therefore not very effective.

* Made flow-graph printing having no side-effects: calling Type() lazily
initializes the type_ and reaching_type_ fields of Value and Definition.
Access fields directly when printing instead.

Also, fixed two random spelling mistakes.

R=srdjan@google.com

Committed: https://code.google.com/p/dart/source/detail?r=28851

[srdjan, 2013-10-17 17:06:45]

lgtm

https://codereview.chromium.org/27727002/diff/6001/runtime/vm/intermediate_la...
File runtime/vm/intermediate_language.cc (right):

https://codereview.chromium.org/27727002/diff/6001/runtime/vm/intermediate_la...
runtime/vm/intermediate_language.cc:1656: } else if
(!MaybeNumber(left()->Type()) || !MaybeNumber(right()->Type())) {
Wouldn't it be enough to just check for MayBeNumber, i.e., Type() should have
enough information, we do not need to check the exact cid but the subtype
relation only?

[Florian Schneider, 2013-10-18 10:55:27]

https://codereview.chromium.org/27727002/diff/6001/runtime/vm/intermediate_la...
File runtime/vm/intermediate_language.cc (right):

https://codereview.chromium.org/27727002/diff/6001/runtime/vm/intermediate_la...
runtime/vm/intermediate_language.cc:1656: } else if
(!MaybeNumber(left()->Type()) || !MaybeNumber(right()->Type())) {
On 2013/10/17 17:06:45, srdjan wrote:
> Wouldn't it be enough to just check for MayBeNumber, i.e., Type() should have
> enough information, we do not need to check the exact cid but the subtype
> relation only?

It would work, but the check on the AbstractType does not handles smis
separately. I guess I could do that with AbstractType, too, but the check for
cid is more efficient.

[Florian Schneider, 2013-10-18 10:59:35]

@Kevin: Could you have a quick look at il_printer.cc? I added the change to make
flow-graph printing have no side-effects on the type information.

[Kevin Millikin (Google), 2013-10-18 11:06:59]

Excellent.

[Florian Schneider, 2013-10-18 11:14:07]

Committed patchset #3 manually as r28851 (presubmit successful).