[builtins] Don't canonicalize results of Math double operations.

[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org

Committed: https://crrev.com/9f4f582bbfd806774f5663a87bfe7b804bf11e9b
Cr-Commit-Position: refs/heads/master@{#40573}

[Benedikt Meurer, 2016-10-26 04:16:01]

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

[commit-bot: I haz the power, 2016-10-26 04:16:05]

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

[Benedikt Meurer, 2016-10-26 04:17:09]

Hey Yang,

Here's a unification for the Math builtin canonicalization. This way the
builtins more or less agree in both baseline and optimized code wrt. the
doubleness of the result.
Please take a look.

Thanks,
Benedikt

[commit-bot: I haz the power, 2016-10-26 04:18:14]

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

[commit-bot: I haz the power, 2016-10-26 04:18:15]

Dry run: Try jobs failed on following builders:
  v8_linux_arm64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_rel_ng/build...)

[Benedikt Meurer, 2016-10-26 04:22:07]

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

[commit-bot: I haz the power, 2016-10-26 04:22:14]

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

[Benedikt Meurer, 2016-10-26 04:32:15]

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

[commit-bot: I haz the power, 2016-10-26 04:32:27]

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

[Benedikt Meurer, 2016-10-26 04:33:21]

Description was changed from

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

R=yangguo@chromium.org
==========

to

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org
==========

[Yang, 2016-10-26 04:39:53]

On 2016/10/26 04:17:09, Benedikt Meurer wrote:
> Hey Yang,
> 
> Here's a unification for the Math builtin canonicalization. This way the
> builtins more or less agree in both baseline and optimized code wrt. the
> doubleness of the result.
> Please take a look.
> 
> Thanks,
> Benedikt

lgtm.

[Benedikt Meurer, 2016-10-26 04:40:09]

The CQ bit was unchecked by bmeurer@chromium.org

[Benedikt Meurer, 2016-10-26 04:40:43]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2016-10-26 04:40:46]

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

[Benedikt Meurer, 2016-10-26 05:24:09]

The CQ bit was unchecked by bmeurer@chromium.org

[Benedikt Meurer, 2016-10-26 05:24:29]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2016-10-26 05:24:35]

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, 2016-10-26 05:56:04]

Description was changed from

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org
==========

to

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org
==========

[commit-bot: I haz the power, 2016-10-26 05:56:05]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-11-17 22:13:03]

Description was changed from

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org
==========

to

==========
[builtins] Don't canonicalize results of Math double operations.

For Math builtins that likely yield double results, i.e. Math.sin,
Math.cos and friends, don't bother trying to canonicalize the result
to Smi. The rationale behind this is that other parts of V8 use the
HeapNumber representation as a hint to assume that certain values
should be represented as double (i.e. for the array elements kind
and for double field tracking). This way the chance that we make
the ideal decision early on is better.

For Math.abs we establish the contract that if the input value is a
Smi, then we try hard to return a Smi (doesn't work for minimal Smi
value), otherwise we preserve the HeapNumberness of the input.

Same for the generic Add, Subtract, Multiply, etc. code stubs.

R=yangguo@chromium.org

Committed: https://crrev.com/9f4f582bbfd806774f5663a87bfe7b804bf11e9b
Cr-Commit-Position: refs/heads/master@{#40573}
==========

[commit-bot: I haz the power, 2016-11-17 22:13:05]

Patchset 2 (id:??) landed as
https://crrev.com/9f4f582bbfd806774f5663a87bfe7b804bf11e9b
Cr-Commit-Position: refs/heads/master@{#40573}