[turbofan] Propagate minus-zero truncation in representation inference.

[turbofan] Propagate minus-zero truncation in representation inference.

This introduces a new truncation bit for truncation of minus-zero to zero.

At the moment it is only used to handle the limit cases of deopt, such as the
one in the Google maps workload (see simplified version below), where the -q
(which is desugared to q * -1.0) currently deoptimizes because the result would
produce minus zero. To handle this situation, we exploit the knowledge that
righthand side of + cannot be -0, so even if lefthand side was -0, the result
would still be 0 (so the + operation cannot distinguish between left hand side
0 and -0).


function f(q) {
  q -= 4;
  return (-q) + q;
}

f(10);
f(10);
%OptimizeFunctionOnNextCall(f);
f(4);

Review-Url: https://codereview.chromium.org/2734253002
Cr-Commit-Position: refs/heads/master@{#43661}
Committed: https://chromium.googlesource.com/v8/v8/+/18f169d46cb4602fe3ee92900501aaf469c76000

[Jarin, 2017-03-07 22:50:10]

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

[commit-bot: I haz the power, 2017-03-07 22:50:14]

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

[Jarin, 2017-03-07 22:54:32]

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

[Jarin, 2017-03-07 22:54:32]

Could you take a look, please?

[commit-bot: I haz the power, 2017-03-07 23:15:50]

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

[commit-bot: I haz the power, 2017-03-07 23:15:51]

Dry run: This issue passed the CQ dry run.

[Benedikt Meurer, 2017-03-08 03:56:41]

Nice. LGTM with nits.

[Benedikt Meurer, 2017-03-08 03:56:53]

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
File src/compiler/simplified-lowering.cc (right):

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:1229: IdentifyZeros lhs_identify_zeros =
truncation.identify_zeros();
Nit: left instead of lbs for consistency.

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:1230: if (node->opcode() ==
IrOpcode::kSpeculativeNumberAdd &&
This deserves a comment, maybe just repeating the example from the CL
description.

[Jarin, 2017-03-08 05:46:51]

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
File src/compiler/simplified-lowering.cc (right):

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:1229: IdentifyZeros lhs_identify_zeros =
truncation.identify_zeros();
On 2017/03/08 03:56:53, Benedikt Meurer wrote:
> Nit: left instead of lbs for consistency.

Done.

https://codereview.chromium.org/2734253002/diff/1/src/compiler/simplified-low...
src/compiler/simplified-lowering.cc:1230: if (node->opcode() ==
IrOpcode::kSpeculativeNumberAdd &&
On 2017/03/08 03:56:53, Benedikt Meurer wrote:
> This deserves a comment, maybe just repeating the example from the CL
> description.

Done.

[Jarin, 2017-03-08 05:48:38]

The CQ bit was checked by jarin@chromium.org

[Jarin, 2017-03-08 05:48:39]

The patchset sent to the CQ was uploaded after l-g-t-m from bmeurer@chromium.org
Link to the patchset: https://codereview.chromium.org/2734253002/#ps20001
(title: "Address reviewer comments")

[commit-bot: I haz the power, 2017-03-08 05:48:45]

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, 2017-03-08 06:11:38]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1488952118706900,
"parent_rev": "99aaa69b29c64bb05b6cf318b01bfac6214de340", "commit_rev":
"18f169d46cb4602fe3ee92900501aaf469c76000"}

[commit-bot: I haz the power, 2017-03-08 06:11:43]

Description was changed from

==========
[turbofan] Propagate minus-zero truncation in representation inference.

This introduces a new truncation bit for truncation of minus-zero to zero.

At the moment it is only used to handle the limit cases of deopt, such as the
one in the Google maps workload (see simplified version below), where the -q
(which is desugared to q * -1.0) currently deoptimizes because the result would
produce minus zero. To handle this situation, we exploit the knowledge that
righthand side of + cannot be -0, so even if lefthand side was -0, the result
would still be 0 (so the + operation cannot distinguish between left hand side
0 and -0).


function f(q) {
  q -= 4;
  return (-q) + q;
}

f(10);
f(10);
%OptimizeFunctionOnNextCall(f);
f(4);
==========

to

==========
[turbofan] Propagate minus-zero truncation in representation inference.

This introduces a new truncation bit for truncation of minus-zero to zero.

At the moment it is only used to handle the limit cases of deopt, such as the
one in the Google maps workload (see simplified version below), where the -q
(which is desugared to q * -1.0) currently deoptimizes because the result would
produce minus zero. To handle this situation, we exploit the knowledge that
righthand side of + cannot be -0, so even if lefthand side was -0, the result
would still be 0 (so the + operation cannot distinguish between left hand side
0 and -0).


function f(q) {
  q -= 4;
  return (-q) + q;
}

f(10);
f(10);
%OptimizeFunctionOnNextCall(f);
f(4);

Review-Url: https://codereview.chromium.org/2734253002
Cr-Commit-Position: refs/heads/master@{#43661}
Committed:
https://chromium.googlesource.com/v8/v8/+/18f169d46cb4602fe3ee92900501aaf469c...
==========

[commit-bot: I haz the power, 2017-03-08 06:11:44]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/v8/v8/+/18f169d46cb4602fe3ee92900501aaf469c...