[turbofan] Consume feedback types for NumberMax and NumberMin.

[turbofan] Consume feedback types for NumberMax and NumberMin.

For nodes

  NumberMin(lhs, rhs)
  NumberMax(lhs, rhs)

we might have feedback types for lhs and rhs that would allow us to
generate unsigned32 or signed32 versions of this operator, which is way
more efficient that going to the full Float64Min/Float64Max operator.
However we cannot promise word32 truncations in this case, since we
based this decision on the feedback types.

This allows us to generate better code for Math.min and Math.max when
one of the inputs is a speculative number operator that provides better
typing during representation selection. We've seen such code in the
hottest function on Google Maps for example.

BUG=v8:5267
R=jarin@chromium.org,mvstanton@chromium.org

Review-Url: https://codereview.chromium.org/2734193003
Cr-Commit-Position: refs/heads/master@{#43660}
Committed: https://chromium.googlesource.com/v8/v8/+/99aaa69b29c64bb05b6cf318b01bfac6214de340

[Benedikt Meurer, 2017-03-07 19:19:35]

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

[commit-bot: I haz the power, 2017-03-07 19:19:37]

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, 2017-03-07 19:46:33]

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

[commit-bot: I haz the power, 2017-03-07 19:46:34]

Dry run: This issue passed the CQ dry run.

[Jarin, 2017-03-07 23:00:51]

lgtm.

https://codereview.chromium.org/2734193003/diff/1/src/compiler/simplified-low...
File src/compiler/simplified-lowering.cc (right):

https://codereview.chromium.org/2734193003/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:2083: VisitBinop(node,
UseInfo::TruncatingFloat64AsWord32(),
Hmm, now that I have been staring at this for longer, I believe we could
actually already look at the feedback type in the cases above (lines 2064, 2070)
because the feedback type can only get smaller, so this can only change the
truncation to a more specific one. Obviously, this reasoning would have to be
well described in the comments. What do you think?

[Benedikt Meurer, 2017-03-08 04:44:07]

https://codereview.chromium.org/2734193003/diff/1/src/compiler/simplified-low...
File src/compiler/simplified-lowering.cc (right):

https://codereview.chromium.org/2734193003/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:2083: VisitBinop(node,
UseInfo::TruncatingFloat64AsWord32(),
Yes, I think it's sufficient to look at the feedback type. But we can just
always pass Float64 truncation, since the Word32 truncation doesn't change
anything if the upper bound is already Signed32 or Unsigned32.

I've changed it to always pass Float64 truncation now (we should really rename
this BTW).

[Benedikt Meurer, 2017-03-08 04:53:52]

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

[commit-bot: I haz the power, 2017-03-08 04:54:02]

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, 2017-03-08 05:16:45]

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

[commit-bot: I haz the power, 2017-03-08 05:16:46]

Dry run: This issue passed the CQ dry run.

[Benedikt Meurer, 2017-03-08 05:39:54]

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

[commit-bot: I haz the power, 2017-03-08 05:40:06]

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

[Benedikt Meurer, 2017-03-08 05:43:17]

The CQ bit was unchecked by bmeurer@chromium.org

[Benedikt Meurer, 2017-03-08 05:43:22]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2017-03-08 05:43:23]

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/2734193003/#ps40001
(title: "Use jarin@s suggestion.")

[commit-bot: I haz the power, 2017-03-08 05:43:25]

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, 2017-03-08 06:05:42]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1488951802933930,
"parent_rev": "76224f7e4994af40e896069a1372373e25699a7c", "commit_rev":
"99aaa69b29c64bb05b6cf318b01bfac6214de340"}

[commit-bot: I haz the power, 2017-03-08 06:05:48]

Description was changed from

==========
[turbofan] Consume feedback types for NumberMax and NumberMin.

For nodes

  NumberMin(lhs, rhs)
  NumberMax(lhs, rhs)

we might have feedback types for lhs and rhs that would allow us to
generate unsigned32 or signed32 versions of this operator, which is way
more efficient that going to the full Float64Min/Float64Max operator.
However we cannot promise word32 truncations in this case, since we
based this decision on the feedback types.

This allows us to generate better code for Math.min and Math.max when
one of the inputs is a speculative number operator that provides better
typing during representation selection. We've seen such code in the
hottest function on Google Maps for example.

BUG=v8:5267
R=jarin@chromium.org,mvstanton@chromium.org
==========

to

==========
[turbofan] Consume feedback types for NumberMax and NumberMin.

For nodes

  NumberMin(lhs, rhs)
  NumberMax(lhs, rhs)

we might have feedback types for lhs and rhs that would allow us to
generate unsigned32 or signed32 versions of this operator, which is way
more efficient that going to the full Float64Min/Float64Max operator.
However we cannot promise word32 truncations in this case, since we
based this decision on the feedback types.

This allows us to generate better code for Math.min and Math.max when
one of the inputs is a speculative number operator that provides better
typing during representation selection. We've seen such code in the
hottest function on Google Maps for example.

BUG=v8:5267
R=jarin@chromium.org,mvstanton@chromium.org

Review-Url: https://codereview.chromium.org/2734193003
Cr-Commit-Position: refs/heads/master@{#43660}
Committed:
https://chromium.googlesource.com/v8/v8/+/99aaa69b29c64bb05b6cf318b01bfac6214...
==========

[commit-bot: I haz the power, 2017-03-08 06:05:49]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/v8/v8/+/99aaa69b29c64bb05b6cf318b01bfac6214...