[turbofan] Split pure nodes in the scheduler if beneficial.

[turbofan] Split pure nodes in the scheduler if beneficial.

If a (pure) node has two or more uses, but there exists a path from the
common dominator of these uses to end, which does not contain a use,
then we split the node such that no unnecessary computation takes place.
Note however, that this only applies if the node cannot be hoisted out
of a loop.

BUG=v8:3864
LOG=n
R=jarin@chromium.org

Committed: https://chromium.googlesource.com/v8/v8/+/d8cfbc633d6f2fcc7016dfb8db8c3f7636c3151e

[Benedikt Meurer, 2015-02-03 13:40:10]

Jaro: PTAL
Michi: FYI

[Jarin, 2015-02-03 14:27:06]

LGTM with nits.

Just for the record, I am not crazy about using if (!x) ... for non-nullness
checks (because it tends to be brittle when you change x to a different type).

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.cc
File src/compiler/scheduler.cc (right):

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.c...
src/compiler/scheduler.cc:1354: BasicBlock* block = marking_queue_.front();
Please, do not shadow the 'block' parameter.

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.c...
src/compiler/scheduler.cc:1388: if (!use_node) {
Are you sure this is always initially zero?

[Benedikt Meurer, 2015-02-03 14:32:22]

Thanks, issues addressed.

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.cc
File src/compiler/scheduler.cc (right):

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.c...
src/compiler/scheduler.cc:1354: BasicBlock* block = marking_queue_.front();
On 2015/02/03 14:27:06, jarin wrote:
> Please, do not shadow the 'block' parameter.

Done.

https://codereview.chromium.org/899433005/diff/20001/src/compiler/scheduler.c...
src/compiler/scheduler.cc:1388: if (!use_node) {
Yes, that's due to the definition of operator[].

[Benedikt Meurer, 2015-02-03 14:32:32]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2015-02-03 14:33:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/899433005/40001

[Benedikt Meurer, 2015-02-03 14:50:16]

The CQ bit was unchecked by bmeurer@chromium.org

[commit-bot: I haz the power, 2015-02-03 14:51:17]

Patchset 3 (id:??) landed as
https://crrev.com/d8cfbc633d6f2fcc7016dfb8db8c3f7636c3151e
Cr-Commit-Position: refs/heads/master@{#26404}

[Benedikt Meurer, 2015-02-03 14:51:19]

Committed patchset #3 (id:40001) manually as
d8cfbc633d6f2fcc7016dfb8db8c3f7636c3151e (presubmit successful).