[turbofan] Infer some receiver maps when lowering loads/stores.

[turbofan] Infer some receiver maps when lowering loads/stores.

The type feedback for load and store ICs might be polluted because of
different/conflicting call sites, but yet we can utilize some of the
static information that is available in the graph to determine the
effective receiver map, or at least filter out impossible receiver maps
from the feedback by looking at the transition trees.

This is similar to what Crankshaft does in ComputeReceiverTypes, but
more general and less hacky.

R=jarin@chromium.org

Committed: https://crrev.com/9c5d12e5fb3d489cafba69e8962613847479c731
Cr-Commit-Position: refs/heads/master@{#36167}

[Benedikt Meurer, 2016-05-11 09:44:45]

Hey Jaro,

Here's an initial shot to improve RayTrace and some other benchmarks.
Please take a look.

Thanks,
Benedikt

[Jarin, 2016-05-11 09:58:51]

lgtm

[Benedikt Meurer, 2016-05-11 09:59:02]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2016-05-11 09:59:10]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1972563002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1972563002/1

[commit-bot: I haz the power, 2016-05-11 10:11:20]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-05-11 10:11:21]

Try jobs failed on following builders:
  v8_linux64_avx2_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_avx2_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng_trig...)

[Benedikt Meurer, 2016-05-11 10:45:35]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2016-05-11 10:45:35]

The patchset sent to the CQ was uploaded after l-g-t-m from jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/1972563002/#ps20001
(title: "Make tests happy.")

[commit-bot: I haz the power, 2016-05-11 10:45:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1972563002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1972563002/20001

[commit-bot: I haz the power, 2016-05-11 11:14:35]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-05-11 11:15:14]

Description was changed from

==========
[turbofan] Infer some receiver maps when lowering loads/stores.

The type feedback for load and store ICs might be polluted because of
different/conflicting call sites, but yet we can utilize some of the
static information that is available in the graph to determine the
effective receiver map, or at least filter out impossible receiver maps
from the feedback by looking at the transition trees.

This is similar to what Crankshaft does in ComputeReceiverTypes, but
more general and less hacky.

R=jarin@chromium.org
==========

to

==========
[turbofan] Infer some receiver maps when lowering loads/stores.

The type feedback for load and store ICs might be polluted because of
different/conflicting call sites, but yet we can utilize some of the
static information that is available in the graph to determine the
effective receiver map, or at least filter out impossible receiver maps
from the feedback by looking at the transition trees.

This is similar to what Crankshaft does in ComputeReceiverTypes, but
more general and less hacky.

R=jarin@chromium.org

Committed: https://crrev.com/9c5d12e5fb3d489cafba69e8962613847479c731
Cr-Commit-Position: refs/heads/master@{#36167}
==========

[commit-bot: I haz the power, 2016-05-11 11:15:15]

Patchset 2 (id:??) landed as
https://crrev.com/9c5d12e5fb3d489cafba69e8962613847479c731
Cr-Commit-Position: refs/heads/master@{#36167}