[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280
Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Committed: https://crrev.com/e5ac75c6354f6dbe0fbe217f857750ab9fb0bd45
Cr-Original-Commit-Position: refs/heads/master@{#39599}
Cr-Commit-Position: refs/heads/master@{#39637}

[rmcilroy, 2016-09-19 16:28:21]

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

[commit-bot: I haz the power, 2016-09-19 16:28:27]

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, 2016-09-19 16:30:37]

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

[commit-bot: I haz the power, 2016-09-19 16:30:38]

Dry run: Try jobs failed on following builders:
  v8_linux_mips64el_compile_rel on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_mips64el_compile_r...)
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/24372)

[rmcilroy, 2016-09-19 17:25:41]

The CQ bit was checked by rmcilroy@chromium.org

[commit-bot: I haz the power, 2016-09-19 17:25:51]

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, 2016-09-19 17:25:52]

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

[commit-bot: I haz the power, 2016-09-19 17:25:53]

No L-G-T-M from a valid reviewer yet. 
CQ run can only be started by full committers or once the patch has
received an L-G-T-M from a full committer.
Even if an L-G-T-M may have been provided, it was from a non-committer,
_not_ a full super star committer.
Committers are members of the group "project-v8-committers".
Note that this has nothing to do with OWNERS files.

[rmcilroy, 2016-09-19 17:32:33]

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

[commit-bot: I haz the power, 2016-09-19 17:32:48]

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, 2016-09-19 17:36:32]

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

[commit-bot: I haz the power, 2016-09-19 17:36:33]

Dry run: Try jobs failed on following builders:
  v8_linux_gcc_compile_rel on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)

[rmcilroy, 2016-09-19 21:41:43]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

BUG=
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

Reduces the percentage of time spent in making BytecodeArrays in 
CodeLoad from ~15% to ~11% according to perf. This increases the
CoadLoad score by around 2%.

BUG=v8:4280
==========

[rmcilroy, 2016-09-19 21:51:41]

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

[commit-bot: I haz the power, 2016-09-19 21:51:47]

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

[rmcilroy, 2016-09-19 21:51:50]

Patchset #2 (id:20001) has been deleted

[rmcilroy, 2016-09-19 21:51:55]

Patchset #1 (id:1) has been deleted

[commit-bot: I haz the power, 2016-09-19 21:56:19]

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

[commit-bot: I haz the power, 2016-09-19 21:56:20]

Dry run: Try jobs failed on following builders:
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/24412)

[rmcilroy, 2016-09-19 22:21:45]

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

[commit-bot: I haz the power, 2016-09-19 22:21:50]

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

[rmcilroy, 2016-09-19 22:21:54]

Patchset #1 (id:40001) has been deleted

[commit-bot: I haz the power, 2016-09-19 22:25:56]

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

[commit-bot: I haz the power, 2016-09-19 22:25:57]

Dry run: Try jobs failed on following builders:
  v8_win_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_rel_ng/builds/14571)

[rmcilroy, 2016-09-19 22:34:57]

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

[commit-bot: I haz the power, 2016-09-19 22:35:03]

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, 2016-09-19 22:38:33]

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

[commit-bot: I haz the power, 2016-09-19 22:38:33]

Dry run: Try jobs failed on following builders:
  v8_win_nosnap_shared_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)

[rmcilroy, 2016-09-20 08:28:54]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

Reduces the percentage of time spent in making BytecodeArrays in 
CodeLoad from ~15% to ~11% according to perf. This increases the
CoadLoad score by around 2%.

BUG=v8:4280
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and 
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the 
BytecodeNode constructor, where the decision on which operands are 
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter 
and into BytecodeRegisterOptimizer (which is the only place outside 
BytecodeGenerator which updates which registers are used). This 
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than 
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512 
bytes,
 - Make common functions in Bytecodes constexpr so that they 
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and 
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array 
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper 
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280
==========

[rmcilroy, 2016-09-20 08:30:54]

rmcilroy@chromium.org changed reviewers:
+ leszeks@chromium.org, mythria@chromium.org

[rmcilroy, 2016-09-20 08:30:55]

Mythri / Leszek, PTAL. This is a big change, but mostly fairly mechanical. I
need to look into the Windows failures (looks like the bots are on MSVC 2013
which doesn't support constexpr, but Chromium switched to MSVC 2015).

[Leszek Swirski, 2016-09-20 10:37:56]

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-array-builder.cc (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-array-builder.cc:381: } else {
DCHECK for strict

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-traits.h (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-traits.h:72: OperandType operand_1, OperandType
operand_2, OperandType operand_3>
Would this code benefit from template packs? e.g. 

template<int... vals>
struct SumHelper;

template <AccumulatorUse accumulator_use, OperandType... operands>
struct BytecodeTraits<accumulator_use, operands...> {

    static constexpr OperandTypeInfo kOperandTypeInfos[] = {
        OperandTraits<operands>::kOperandTypeInfo...
    };

    static constexpr int kDoubleScaleSize = SumHelper<1, OperandScaler<operands,
OperandScale::kDouble>::kSize...>::VALUE;
}

template<int val>
struct SumHelper<val> {
  enum { VALUE = val };
};

template<int val, int... vals>
struct SumHelper<val, vals...> {
  enum { VALUE = val + SumHelper<vals...>::VALUE };
};

[rmcilroy, 2016-09-20 16:57:25]

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-array-builder.cc (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-array-builder.cc:381: } else {
On 2016/09/20 10:37:56, Leszek Swirski wrote:
> DCHECK for strict

Done.

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-traits.h (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-traits.h:72: OperandType operand_1, OperandType
operand_2, OperandType operand_3>
On 2016/09/20 10:37:56, Leszek Swirski wrote:
> Would this code benefit from template packs? e.g. 
> 
> template<int... vals>
> struct SumHelper;
> 
> template <AccumulatorUse accumulator_use, OperandType... operands>
> struct BytecodeTraits<accumulator_use, operands...> {
> 
>     static constexpr OperandTypeInfo kOperandTypeInfos[] = {
>         OperandTraits<operands>::kOperandTypeInfo...
>     };
> 
>     static constexpr int kDoubleScaleSize = SumHelper<1,
OperandScaler<operands,
> OperandScale::kDouble>::kSize...>::VALUE;
> }
> 
> template<int val>
> struct SumHelper<val> {
>   enum { VALUE = val };
> };
> 
> template<int val, int... vals>
> struct SumHelper<val, vals...> {
>   enum { VALUE = val + SumHelper<vals...>::VALUE };
> };

Great suggestion! Works like a charm, thanks.

[rmcilroy, 2016-09-20 16:57:31]

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

[commit-bot: I haz the power, 2016-09-20 16:57:37]

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, 2016-09-20 16:59:26]

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

[commit-bot: I haz the power, 2016-09-20 16:59:27]

Dry run: Try jobs failed on following builders:
  v8_linux64_avx2_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_avx2_rel_ng/buil...)
  v8_linux64_gyp_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_gyp_rel_ng/build...)
  v8_linux64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng/builds/13104)
  v8_linux_dbg_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng/builds/13101)
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/24521)

[rmcilroy, 2016-09-20 20:12:16]

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

[commit-bot: I haz the power, 2016-09-20 20:12:24]

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, 2016-09-20 20:15:34]

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

[commit-bot: I haz the power, 2016-09-20 20:15:35]

Dry run: Try jobs failed on following builders:
  v8_win64_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win64_rel_ng/builds/14630)
  v8_win_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_rel_ng/builds/14665)

[rmcilroy, 2016-09-20 20:31:10]

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

[commit-bot: I haz the power, 2016-09-20 20:31:15]

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

[rmcilroy, 2016-09-20 20:34:21]

Patchset #4 (id:120001) has been deleted

[commit-bot: I haz the power, 2016-09-20 20:35:41]

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

[commit-bot: I haz the power, 2016-09-20 20:35:42]

Dry run: Try jobs failed on following builders:
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/24540)

[rmcilroy, 2016-09-20 20:47:35]

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

[commit-bot: I haz the power, 2016-09-20 20:47:41]

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, 2016-09-20 20:51:31]

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

[commit-bot: I haz the power, 2016-09-20 20:51:32]

Dry run: Try jobs failed on following builders:
  v8_win_nosnap_shared_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)

[rmcilroy, 2016-09-20 21:57:10]

Patchset #5 (id:160001) has been deleted

[rmcilroy, 2016-09-20 21:57:17]

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

[commit-bot: I haz the power, 2016-09-20 21:57:26]

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, 2016-09-20 22:02:13]

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

[commit-bot: I haz the power, 2016-09-20 22:02:14]

Dry run: Try jobs failed on following builders:
  v8_presubmit on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/24549)

[mythria, 2016-09-21 08:12:45]

Nice. lgtm.

Thanks,
Mythri.

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-pipeline.h (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-pipeline.h:277: ResetScale();
may be use UpdateScale(). I initially thought we are resetting it to default.

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-register-optimizer.cc (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-register-optimizer.cc:221:
DCHECK(!Bytecodes::IsJump(node->bytecode()));
May be a comment saying jumps are handled by WriteJump.

[Leszek Swirski, 2016-09-21 14:11:50]

lgtm from me too

[rmcilroy, 2016-09-21 14:12:34]

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-pipeline.h (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-pipeline.h:277: ResetScale();
On 2016/09/21 08:12:45, mythria wrote:
> may be use UpdateScale(). I initially thought we are resetting it to default.

Done.

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
File src/interpreter/bytecode-register-optimizer.cc (right):

https://codereview.chromium.org/2351763002/diff/80001/src/interpreter/bytecod...
src/interpreter/bytecode-register-optimizer.cc:221:
DCHECK(!Bytecodes::IsJump(node->bytecode()));
On 2016/09/21 08:12:45, mythria wrote:
> May be a comment saying jumps are handled by WriteJump.

Done.

[rmcilroy, 2016-09-21 14:13:02]

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

[commit-bot: I haz the power, 2016-09-21 14:13:16]

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, 2016-09-21 14:39:20]

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

[commit-bot: I haz the power, 2016-09-21 14:39:21]

Dry run: This issue passed the CQ dry run.

[rmcilroy, 2016-09-21 14:53:56]

The CQ bit was checked by rmcilroy@chromium.org

[rmcilroy, 2016-09-21 14:53:57]

The patchset sent to the CQ was uploaded after l-g-t-m from
mythria@chromium.org, leszeks@chromium.org
Link to the patchset: https://codereview.chromium.org/2351763002/#ps200001
(title: "Mythri's comments")

[commit-bot: I haz the power, 2016-09-21 14:54:03]

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, 2016-09-21 15:02:39]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and 
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the 
BytecodeNode constructor, where the decision on which operands are 
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter 
and into BytecodeRegisterOptimizer (which is the only place outside 
BytecodeGenerator which updates which registers are used). This 
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than 
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512 
bytes,
 - Make common functions in Bytecodes constexpr so that they 
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and 
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array 
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper 
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and 
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the 
BytecodeNode constructor, where the decision on which operands are 
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter 
and into BytecodeRegisterOptimizer (which is the only place outside 
BytecodeGenerator which updates which registers are used). This 
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than 
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512 
bytes,
 - Make common functions in Bytecodes constexpr so that they 
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and 
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array 
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper 
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280
==========

[commit-bot: I haz the power, 2016-09-21 15:02:41]

Committed patchset #6 (id:200001)

[commit-bot: I haz the power, 2016-09-21 15:03:12]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and 
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the 
BytecodeNode constructor, where the decision on which operands are 
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter 
and into BytecodeRegisterOptimizer (which is the only place outside 
BytecodeGenerator which updates which registers are used). This 
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than 
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512 
bytes,
 - Make common functions in Bytecodes constexpr so that they 
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and 
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array 
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper 
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

[commit-bot: I haz the power, 2016-09-21 15:03:14]

Patchset 6 (id:??) landed as
https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}

[Michael Hablich, 2016-09-22 05:53:02]

A revert of this CL (patchset #6 id:200001) has been created in
https://codereview.chromium.org/2360193003/ by hablich@chromium.org.

The reason for reverting is: Prime suspect for roll blocker:
https://codereview.chromium.org/2362503002/.

[rmcilroy, 2016-09-22 10:42:44]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

[rmcilroy, 2016-09-22 11:26:52]

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

[commit-bot: I haz the power, 2016-09-22 11:26:56]

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, 2016-09-22 11:56:46]

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

[commit-bot: I haz the power, 2016-09-22 11:56:48]

Dry run: This issue passed the CQ dry run.

[rmcilroy, 2016-09-22 16:31:08]

The CQ bit was checked by rmcilroy@chromium.org

[rmcilroy, 2016-09-22 16:31:09]

The patchset sent to the CQ was uploaded after l-g-t-m from
mythria@chromium.org, leszeks@chromium.org
Link to the patchset: https://codereview.chromium.org/2351763002/#ps220001
(title: "Fix Chromium Windows bots.")

[commit-bot: I haz the power, 2016-09-22 16:31:12]

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, 2016-09-22 16:34:22]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

[commit-bot: I haz the power, 2016-09-22 16:34:23]

Committed patchset #7 (id:220001)

[commit-bot: I haz the power, 2016-09-22 16:34:40]

Description was changed from

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Cr-Commit-Position: refs/heads/master@{#39599}
==========

to

==========
[Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.

This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:

 - Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
 - Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
 - Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
 - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
 - Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
 - Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
 - Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.

I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.

This reduces the percentage of time spent in making BytecodeArrays
 in  CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.

BUG=v8:4280

Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Committed: https://crrev.com/e5ac75c6354f6dbe0fbe217f857750ab9fb0bd45
Cr-Original-Commit-Position: refs/heads/master@{#39599}
Cr-Commit-Position: refs/heads/master@{#39637}
==========

[commit-bot: I haz the power, 2016-09-22 16:34:42]

Patchset 7 (id:??) landed as
https://crrev.com/e5ac75c6354f6dbe0fbe217f857750ab9fb0bd45
Cr-Commit-Position: refs/heads/master@{#39637}