[turbofan] Use Node input iterators where possible

[turbofan] Use Node input iterators where possible

Changes some for loops to use node->inputs() instead of iterating over
InputCount and accessing InputAt(i). This saves some checks for
"has_inline_inputs" and so some branches.

Review-Url: https://codereview.chromium.org/2585713002
Cr-Commit-Position: refs/heads/master@{#42024}
Committed: https://chromium.googlesource.com/v8/v8/+/0b10c04473587aefc6ef5414473a65aa1707440a

[Leszek Swirski, 2016-12-16 09:59:28]

The CQ bit was checked by leszeks@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-12-16 09:59:34]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Leszek Swirski, 2016-12-16 10:00:54]

Hi Jaro,

I see some small but non-zero compile time improvements with this change. On
Kraken:

+------------------------------+---------+---------+--------+---------+---------+--------+-----------+
|             Test             |  mean   |         |        | median  |        
|        |  p-value  |
|                              | before  | after   | Δ      | before  | after  
| Δ      |           |
+------------------------------+---------+---------+--------+---------+---------+--------+-----------+
| untyper - time               | 0.532   | 0.481   | -9.61% | 0.525   | 0.477  
| -9.14% | 6.45e-118 |
| typer - time                 | 9.820   | 9.313   | -5.16% | 9.688   | 9.197  
| -5.06% | 1.71e-71  |
| late-optimization - time     | 17.316  | 16.489  | -4.77% | 17.312  | 16.486 
| -4.78% | 5.31e-65  |
| dead-code-elimination - time | 7.680   | 7.368   | -4.07% | 7.662   | 7.353  
| -4.03% | 1.23e-49  |
| typed-lowering - time        | 11.256  | 10.883  | -3.32% | 11.096  | 10.727 
| -3.32% | 1.6e-30   |
| generic-lowering - time      | 3.748   | 3.631   | -3.14% | 3.709   | 3.590  
| -3.21% | 8.35e-28  |
| early-optimization - time    | 19.675  | 19.109  | -2.88% | 19.640  | 19.059 
| -2.96% | 4.29e-23  |
| inlining - time              | 19.487  | 18.974  | -2.63% | 19.209  | 18.699 
| -2.66% | 1.05e-17  |
| locate-spill-slots - time    | 0.237   | 0.231   | -2.28% | 0.236   | 0.231  
| -2.12% | 4.96e-16  |
| load-elimination - time      | 19.252  | 18.843  | -2.12% | 19.020  | 18.660 
| -1.89% | 9.01e-15  |
| init-pipeline-data - time    | 0.105   | 0.108   | 3.54%  | 0.102   | 0.106  
| 3.92%  | 2.43e-14  |
| graph-creation - time        | 50.469  | 49.370  | -2.18% | 49.630  | 48.633 
| -2.01% | 8.04e-14  |
| lowering - time              | 69.835  | 68.948  | -1.27% | 69.003  | 68.150 
| -1.24% | 2.66e-06  |
| OSR-deconstruction - time    | 4.119   | 4.042   | -1.89% | 4.075   | 4.024  
| -1.26% | 4.56e-06  |
| escape-analysis - time       | 6.895   | 6.810   | -1.23% | 6.805   | 6.723  
| -1.21% | 7.92e-06  |
| totals - time                | 470.183 | 466.848 | -0.71% | 469.848 | 465.284
| -0.97% | 0.000281  |
| late-graph-trimming - time   | 5.413   | 5.459   | 0.85%  | 5.402   | 5.450  
| 0.88%  | 0.00127   |
| connect-ranges - time        | 1.876   | 1.884   | 0.46%  | 1.867   | 1.881  
| 0.72%  | 0.00128   |
+------------------------------+---------+---------+--------+---------+---------+--------+-----------+

[commit-bot: I haz the power, 2016-12-16 10:26:04]

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

[commit-bot: I haz the power, 2016-12-16 10:26:05]

Dry run: This issue passed the CQ dry run.

[Leszek Swirski, 2016-12-20 15:45:38]

leszeks@chromium.org changed reviewers:
+ bmeurer@chromium.org, jarin@chromium.org

[Leszek Swirski, 2016-12-20 15:45:38]

Hi Benedikt and Jaro,

I thought I'd sent this out for review, but looks like not. I see benefits from
using (and caching) node input iterators in graph reducer, allowing us to skip
the has_inline_inputs check while iterating.

[Benedikt Meurer, 2016-12-20 18:50:45]

LGTM from me, although I'm not sure that operator< on pointers is a good idea,
can easily result in undefined bevahior and the compiler won't warn you.

[Jarin, 2016-12-21 06:16:12]

lgtm. (Sorry, but I have to say this: oh god, this is ugly.)

[Leszek Swirski, 2016-12-21 15:34:48]

The CQ bit was checked by leszeks@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-12-21 15:34:56]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-21 15:37:46]

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

[commit-bot: I haz the power, 2016-12-21 15:37:47]

Dry run: Try jobs failed on following builders:
  v8_linux64_avx2_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng/builds/18243)
  v8_mac_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_rel_ng/builds/14660)

[Leszek Swirski, 2016-12-21 15:46:27]

The CQ bit was checked by leszeks@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-12-21 15:46:37]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-21 16:12:54]

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

[commit-bot: I haz the power, 2016-12-21 16:12:55]

Dry run: This issue passed the CQ dry run.

[Leszek Swirski, 2017-01-03 11:03:59]

The CQ bit was checked by leszeks@chromium.org

[Leszek Swirski, 2017-01-03 11:03:59]

The patchset sent to the CQ was uploaded after l-g-t-m from
bmeurer@chromium.org, jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/2585713002/#ps40001
(title: "Use == instead of DCHECK_EQ because of missing operator<< for
iterators")

[commit-bot: I haz the power, 2017-01-03 11:04:08]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-03 11:29:24]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1483441439601180,
"parent_rev": "e3ac5a31cb645974b02e1d7b238f6a9c35ab0f5b", "commit_rev":
"0b10c04473587aefc6ef5414473a65aa1707440a"}

[commit-bot: I haz the power, 2017-01-03 11:29:27]

Description was changed from

==========
[turbofan] Use Node input iterators where possible

Changes some for loops to use node->inputs() instead of iterating over
InputCount and accessing InputAt(i). This saves some checks for
"has_inline_inputs" and so some branches.
==========

to

==========
[turbofan] Use Node input iterators where possible

Changes some for loops to use node->inputs() instead of iterating over
InputCount and accessing InputAt(i). This saves some checks for
"has_inline_inputs" and so some branches.

Review-Url: https://codereview.chromium.org/2585713002
Cr-Commit-Position: refs/heads/master@{#42024}
Committed:
https://chromium.googlesource.com/v8/v8/+/0b10c04473587aefc6ef5414473a65aa170...
==========

[commit-bot: I haz the power, 2017-01-03 11:29:28]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/v8/v8/+/0b10c04473587aefc6ef5414473a65aa170...