[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.

[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64 on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing for platforms with simple aliasing.

LOG=N
BUG=v8:4124
Committed: https://crrev.com/4b76dc85979094ac26499f2e527e422d4f0f6e58
Cr-Commit-Position: refs/heads/master@{#37388}

[bbudge, 2016-06-27 18:17:23]

Description was changed from

==========
[Turbofan] Make RegisterAllocation great again.
- Changes InstructionOperand canonicalization back to map all FP operands
to kFloat64.
- Makes interference calculation simple on archs with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Make RegisterAllocation great again.
- Changes InstructionOperand canonicalization back to map all FP operands to
kFloat64 on Intel and other platforms with simple aliasing.
- Bypass interference calculations and fixed live range handling for platforms
with simple aliasing.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-06-28 23:24:56]

Description was changed from

==========
[Turbofan] Make RegisterAllocation great again.
- Changes InstructionOperand canonicalization back to map all FP operands to
kFloat64 on Intel and other platforms with simple aliasing.
- Bypass interference calculations and fixed live range handling for platforms
with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Operand canonicalization only for non-simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-06-29 00:06:46]

Description was changed from

==========
[Turbofan] Operand canonicalization only for non-simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] SImplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-06-29 00:12:05]

Description was changed from

==========
[Turbofan] SImplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-06-29 00:12:48]

bbudge@chromium.org changed reviewers:
+ bmeurer@chromium.org

[bbudge, 2016-06-29 00:12:49]

This should help reduce the impact of SIMD registers on non-ARM archs.

[bbudge, 2016-06-29 00:14:57]

On 2016/06/29 00:12:49, bbudge wrote:
> This should help reduce the impact of SIMD registers on non-ARM archs.

I ran the perf tests twice and these should be visible on the last patchset.

[bbudge, 2016-06-29 00:14:57]

On 2016/06/29 00:12:49, bbudge wrote:
> This should help reduce the impact of SIMD registers on non-ARM archs.

I ran the perf tests twice and these should be visible on the last patchset.

[Benedikt Meurer, 2016-06-29 03:41:53]

Nice. LGTM.

[bbudge, 2016-06-29 11:44:57]

The CQ bit was checked by bbudge@chromium.org

[commit-bot: I haz the power, 2016-06-29 11:45:09]

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, 2016-06-29 11:47:15]

Description was changed from

==========
[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

[commit-bot: I haz the power, 2016-06-29 11:47:16]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-06-29 11:49:14]

Description was changed from

==========
[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Simplify operand canonicalization on archs with simple FP aliasing.
- Changes InstructionOperand canonicalization to map all FP operands to kFloat64
on Intel and other platforms with simple aliasing.
- Bypass expensive interference calculations and fixed FP live range processing
for platforms with simple aliasing.

LOG=N
BUG=v8:4124

Committed: https://crrev.com/4b76dc85979094ac26499f2e527e422d4f0f6e58
Cr-Commit-Position: refs/heads/master@{#37388}
==========

[commit-bot: I haz the power, 2016-06-29 11:49:15]

Patchset 3 (id:??) landed as
https://crrev.com/4b76dc85979094ac26499f2e527e422d4f0f6e58
Cr-Commit-Position: refs/heads/master@{#37388}