PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and Math.ceil.

PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=chromium:476477, v8:2890, v8:4059
LOG=n

Committed: https://crrev.com/4e93ce4f62ab6766dad17d8a6132f5270c148f86
Cr-Commit-Position: refs/heads/master@{#35659}

[MTBrandyberry, 2016-03-30 22:53:42]

PTAL

[MTBrandyberry, 2016-03-30 22:53:49]

The CQ bit was checked by mbrandy@us.ibm.com to run a CQ dry run

[commit-bot: I haz the power, 2016-03-30 22:53:59]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1839643007/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1839643007/1

[commit-bot: I haz the power, 2016-03-30 23:16:13]

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

[commit-bot: I haz the power, 2016-03-30 23:16:13]

Dry run: This issue passed the CQ dry run.

[JoranSiu, 2016-03-31 13:18:29]

On 2016/03/30 23:16:13, commit-bot: I haz the power wrote:
> Dry run: This issue passed the CQ dry run.

lgtm

[MTBrandyberry, 2016-03-31 13:38:30]

The CQ bit was checked by mbrandy@us.ibm.com

[commit-bot: I haz the power, 2016-03-31 13:38:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1839643007/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1839643007/1

[commit-bot: I haz the power, 2016-03-31 13:43:27]

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

[commit-bot: I haz the power, 2016-03-31 13:43:28]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/13054)

[MTBrandyberry, 2016-03-31 14:06:30]

On 2016/03/31 13:43:28, commit-bot: I haz the power wrote:
> Try jobs failed on following builders:
>   v8_presubmit on tryserver.v8 (JOB_FAILED,
> http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/13054)

Need approval for src/crankshaft/hydrogen-instructions.h change.

[MTBrandyberry, 2016-03-31 17:58:43]

Description was changed from

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n
==========

to

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n
==========

[MTBrandyberry, 2016-03-31 17:58:43]

mbrandy@us.ibm.com changed reviewers:
+ danno@chromium.org

[MTBrandyberry, 2016-04-01 02:32:58]

The CQ bit was checked by mbrandy@us.ibm.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-01 02:33:05]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1839643007/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1839643007/20001

[commit-bot: I haz the power, 2016-04-01 06:08:47]

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

[commit-bot: I haz the power, 2016-04-01 06:08:48]

Dry run: This issue passed the CQ dry run.

[MTBrandyberry, 2016-04-01 15:27:49]

On 2016/04/01 06:08:48, commit-bot: I haz the power wrote:
> Dry run: This issue passed the CQ dry run.

Need approval for src/crankshaft/hydrogen-instructions.h change.

[MTBrandyberry, 2016-04-19 20:59:36]

On 2016/04/01 15:27:49, MTBrandyberry wrote:
> On 2016/04/01 06:08:48, commit-bot: I haz the power wrote:
> > Dry run: This issue passed the CQ dry run.
> 
> Need approval for src/crankshaft/hydrogen-instructions.h change.

Can I get a review for this?

[Benedikt Meurer, 2016-04-20 03:56:07]

Sorry, missed this one. LGTM.

[MTBrandyberry, 2016-04-20 13:13:53]

The CQ bit was checked by mbrandy@us.ibm.com

[MTBrandyberry, 2016-04-20 13:13:53]

The patchset sent to the CQ was uploaded after l-g-t-m from joransiu@ca.ibm.com
Link to the patchset: https://codereview.chromium.org/1839643007/#ps20001
(title: "Fix DEBUG")

[commit-bot: I haz the power, 2016-04-20 13:14:00]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1839643007/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1839643007/20001

[commit-bot: I haz the power, 2016-04-20 13:33:19]

Description was changed from

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n
==========

to

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n
==========

[commit-bot: I haz the power, 2016-04-20 13:33:20]

Committed patchset #2 (id:20001)

[nodir1, 2016-04-22 18:44:12]

Description was changed from

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n
==========

to

==========
PPC: [crankshaft] Address the deoptimization loops of Math.floor, Math.round and
Math.ceil.

Port 978ad03b9275fb338f6dd484c1ef732d1a97ccfe

Original commit message:
    Fix and re-enable the flexible representation for Math.floor (which is used
to
    implement Math.ceil) and Math.round, which allows Math.floor and Math.round
to
    return double results instead of int32, and therefore allows values outside
    the int32 range, especially -0 is now a valid result, which doesn't deopt.

    Also port this feature to x64 and ia32 when the CPU supports the SSE4.1
    extension.

    This addresses all the known deoptimization loops related to Math.round
    in the Kraken benchmark suite, and seems to also address most of the
    deoptimization loops related to Math.floor in the Oort Online benchmark.

    Drive-by-fix: Import the regression tests for the broken HMathFloorOfDiv
    optimization that caused the initial revert of the feature (for arm64 only
    back then).

R=bmeurer@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com,
michael_dawson@ca.ibm.com
BUG=chromium:476477,v8:2890,v8:4059
LOG=n

Committed: https://crrev.com/4e93ce4f62ab6766dad17d8a6132f5270c148f86
Cr-Commit-Position: refs/heads/master@{#35659}
==========

[nodir1, 2016-04-22 18:44:13]

Patchset 2 (id:??) landed as
https://crrev.com/4e93ce4f62ab6766dad17d8a6132f5270c148f86
Cr-Commit-Position: refs/heads/master@{#35659}