[WASM SIMD] Remove opcodes that are slow on some platforms.

[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine. Clients are better off
  synthesizing these from splats, multiplies and adds.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum. Clients are
  better off synthesizing these or using the reciprocal approximations,
  possibly with a refinement step.

LOG=N
BUG=v8:6020
Review-Url: https://codereview.chromium.org/2827143002
Cr-Commit-Position: refs/heads/master@{#44784}
Committed: https://chromium.googlesource.com/v8/v8/+/dddfcfd0a965199bdc41954d88f6e57e6cae17e1

[bbudge, 2017-04-19 20:45:26]

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

[commit-bot: I haz the power, 2017-04-19 20:45:34]

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

[bbudge, 2017-04-19 20:50:24]

Description was changed from

==========
[WASM SIMD] Remove opcodes that have performance cliffs on some platforms.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine.

LOG=N
BUG=v8:6020
==========

to

==========
[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine.

LOG=N
BUG=v8:6020
==========

[commit-bot: I haz the power, 2017-04-19 20:53:04]

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

[commit-bot: I haz the power, 2017-04-19 20:53:04]

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...)

[bbudge, 2017-04-19 21:00:57]

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

[commit-bot: I haz the power, 2017-04-19 21:01:05]

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

[bbudge, 2017-04-19 21:10:51]

bbudge@chromium.org changed reviewers:
+ aseemgarg@chromium.org, bmeurer@chromium.org, bradnelson@chromium.org,
dusan.simicic@imgtec.com

[bbudge, 2017-04-19 21:10:52]

Benedikt for compiler
Dusan for MIPS
Aseem for simd-scalar-lowering

[bbudge, 2017-04-19 21:15:46]

Description was changed from

==========
[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine.

LOG=N
BUG=v8:6020
==========

to

==========
[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine. Clients are better off
  synthesizing these from splats, multiplies and adds.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum. Clients are
  better off synthesizing these or using the reciprocal approximations,
  possibly with a refinement step.

LOG=N
BUG=v8:6020
==========

[commit-bot: I haz the power, 2017-04-19 21:24:09]

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

[commit-bot: I haz the power, 2017-04-19 21:24:10]

Dry run: This issue passed the CQ dry run.

[dusan.simicic, 2017-04-20 10:31:32]

LGTM

[aseemgarg, 2017-04-20 22:37:13]

lgtm

[bbudge, 2017-04-21 20:36:46]

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

[bbudge, 2017-04-21 20:36:47]

+Mircea for compiler

[Mircea Trofin, 2017-04-21 21:02:40]

On 2017/04/21 20:36:47, bbudge wrote:
> +Mircea for compiler

lgtm

[bbudge, 2017-04-21 21:03:51]

The CQ bit was checked by bbudge@chromium.org

[commit-bot: I haz the power, 2017-04-21 21:03:54]

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, 2017-04-21 21:34:45]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1492808631988820,
"parent_rev": "75ce09b53392eee23f3c0c6727463f9d9ef9c73f", "commit_rev":
"dddfcfd0a965199bdc41954d88f6e57e6cae17e1"}

[commit-bot: I haz the power, 2017-04-21 21:34:53]

Description was changed from

==========
[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine. Clients are better off
  synthesizing these from splats, multiplies and adds.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum. Clients are
  better off synthesizing these or using the reciprocal approximations,
  possibly with a refinement step.

LOG=N
BUG=v8:6020
==========

to

==========
[WASM SIMD] Remove opcodes that are slow on some platforms.
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine. Clients are better off
  synthesizing these from splats, multiplies and adds.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum. Clients are
  better off synthesizing these or using the reciprocal approximations,
  possibly with a refinement step.

LOG=N
BUG=v8:6020

Review-Url: https://codereview.chromium.org/2827143002
Cr-Commit-Position: refs/heads/master@{#44784}
Committed:
https://chromium.googlesource.com/v8/v8/+/dddfcfd0a965199bdc41954d88f6e57e6ca...
==========

[commit-bot: I haz the power, 2017-04-21 21:34:54]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/v8/v8/+/dddfcfd0a965199bdc41954d88f6e57e6ca...