Commit Graph

15 Commits

Author SHA1 Message Date
bmeurer
91a2ea815b [turbofan] We never use Int64Constant for branch/select conditions.
Remove dead code to optimize Int64Constants as branch/select conditions,
because we either have tagged booleans or bits represented as word32.

R=jarin@chromium.org

Review-Url: https://codereview.chromium.org/1994533002
Cr-Commit-Position: refs/heads/master@{#36308}
2016-05-18 07:20:19 +00:00
jarin
bb2a830deb [turbofan] Make MachineType a pair of enums.
MachineType is now a class with two enum fields:
- MachineRepresentation
- MachineSemantic

Both enums are usable on their own, and this change switches some places from using MachineType to use just MachineRepresentation. Most notably:
- register allocator now uses just the representation.
- Phi and Select nodes only refer to representations.

Review URL: https://codereview.chromium.org/1513543003

Cr-Commit-Position: refs/heads/master@{#32738}
2015-12-10 09:03:53 +00:00
titzer
27433918f5 Move machine-type.h from src/compiler to src/.
R=bmeurer@chromium.org,jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1498833002

Cr-Commit-Position: refs/heads/master@{#32564}
2015-12-03 13:34:02 +00:00
bmeurer
ce0431d665 [turbofan] Also update the BranchHint when merging a BooleanNot.
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1218443002

Cr-Commit-Position: refs/heads/master@{#29324}
2015-06-26 12:08:26 +00:00
bmeurer
7337021e54 [turbofan] Add support for pushing returns into merges.
This will enable tail call optimization even across inlining. Plus it
might enable some other interesting optimizations as well. In order to
avoid blowing up the generated code, we can still canonicalize the
epilogue in the CodeGenerator, similar to what fullcodegen does.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1215623002

Cr-Commit-Position: refs/heads/master@{#29311}
2015-06-26 08:21:01 +00:00
bmeurer
35eb3a0260 [turbofan] Optimize BooleanNot conditions to Branch nodes.
Also remove the weird work-around for this missing optimization in
CHECK_DATE in macros.py.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1205353002

Cr-Commit-Position: refs/heads/master@{#29288}
2015-06-25 11:07:09 +00:00
bmeurer
733a246386 [turbofan] Proper dead code elimination as regular reducer.
The three different concerns that the ControlReducer used to deal with
are now properly separated into

  a.) DeadCodeElimination, which is a regular AdvancedReducer, that
      propagates Dead via control edges,
  b.) CommonOperatorReducer, which does strength reduction on common
      operators (i.e. Branch, Phi, and friends), and
  c.) GraphTrimming, which removes dead->live edges from the graph.

This will make it possible to run the DeadCodeElimination together with
other passes that actually introduce Dead nodes, i.e. typed lowering;
and it opens the door for general inlining without two stage fix point
iteration.

To make the DeadCodeElimination easier and more uniform, we basically
reverted the introduction of DeadValue and DeadEffect, and changed the
Dead operator to produce control, value and effect. Note however that
this is not a requirement, but merely a way to make dead propagation
easier and more uniform. We could always go back and decide to have
different Dead operators if some other change requires that.

Note that there are several additional opportunities for cleanup now,
i.e. OSR deconstruction could be a regular reducer now, and we don't
need to use TheHole as dead value marker in the GraphReducer. And we can
actually run the dead code elimination together with the other passes
instead of using separate passes over the graph.  We will do this in
follow up CLs.

R=jarin@chromium.org, mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/1193833002

Cr-Commit-Position: refs/heads/master@{#29146}
2015-06-19 12:07:26 +00:00
bmeurer
92e6bcf13c [turbofan] Improve interplay of ControlReducer and CommonOperatorReducer.
This turns the CommonOperatorReducer into an AdvancedReducer and makes
it independent of JSGraph (which was used only because it was convienent),
and let's the CommonOperatorReducer run together with the ControlReducer.

The ControlReducer is still not able to run together with other reducers,
but we're getting closer. The plan is to split the ControlReducer into
two parts: The dead code elimination part and the common operator
reduction part. This separation will help to avoid tricky bugs in the
future and should make testing a *lot* easier.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1192063002

Cr-Commit-Position: refs/heads/master@{#29105}
2015-06-18 09:15:42 +00:00
bmeurer
189609e197 [turbofan] Float32Abs and Float64Abs are supported by all backends.
The Float32Abs and Float64Abs operators are supported by all TurboFan
backends, so we no longer need the flags for them.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1132033002

Cr-Commit-Position: refs/heads/master@{#28310}
2015-05-08 08:25:05 +00:00
Ross McIlroy
063fc25122 Replace OVERRIDE->override and FINAL->final since we now require C++11.
R=jochen@chromium.org

Review URL: https://codereview.chromium.org/1088993003

Cr-Commit-Position: refs/heads/master@{#27937}
2015-04-20 13:08:14 +00:00
bmeurer
35f6c0fdbf [turbofan] Optimize silent hole checks on legacy const context slots.
Currently we always generate a diamond in the graph builder for every
legacy const context slot, which we cannot get rid of until late control
reduction, even if we know after context specialization that the slot is
already initialized.

Now we generate a select instead, which the CommonOperatorReducer
happily removes during typed lowering. This greatly speeds up asm.js
code generated by Emscripten with the new POINTER_MASKING mode.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/1072353002

Cr-Commit-Position: refs/heads/master@{#27739}
2015-04-10 10:28:12 +00:00
Benedikt Meurer
9af9f1d026 [turbofan] Add new Float32Abs and Float64Abs operators.
These operators compute the absolute floating point value of some
arbitrary input, and are implemented without any branches (i.e. using
vabs on arm, and andps/andpd on x86).

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1066393002

Cr-Commit-Position: refs/heads/master@{#27662}
2015-04-08 11:55:04 +00:00
bmeurer
99f8d57f3c [turbofan] Introduce optional Float64Min and Float64Max machine operators.
Basically recognize certain x < y ? x : y constructs and turn that into
Float64Min/Float64Max operations, if the target machine supports that.
On x86 we lower to (v)minsd/(v)maxsd.

R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/998283002

Cr-Commit-Position: refs/heads/master@{#27160}
2015-03-12 14:07:39 +00:00
bmeurer
e53845d41c [turbofan] Cleanup Graph and related classes.
- Move NodeMarker to its own file, and introduce a non
  templatized base class.
- Cleanup the include hell.
- Sanitize the Node construction methods now that we
  got rid of that GenericNode/GenericGraph stuff.
- Protect against NodeId overflow in Graph.
- Various minor cleanups.

TEST=cctest,mjsunit,unittests

Review URL: https://codereview.chromium.org/838783002

Cr-Commit-Position: refs/heads/master@{#25977}
2015-01-07 14:42:49 +00:00
Benedikt Meurer
ee98a1d760 [turbofan] Introduce CommonOperatorReducer.
The CommonOperatorReducer currently takes care of redundant Phis,
EffectPhis and Selects. This functionality overlaps with ControlReducer,
but is required to make certain optimizations effective, since the
ControlReducer only runs really early and really late in the pipeline
and therefore other reducers aren't reapplied properly after redundant
phi/select elimination.

TEST=unittests
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/817243003

Cr-Commit-Position: refs/heads/master@{#25922}
2014-12-22 13:06:43 +00:00