Select sbfx for ((x << k) >> k) in ARM64 instruction selector, and similarly
for ubfx. This is a more generic version of the previous sxtb/h selector.
BUG=
Review URL: https://codereview.chromium.org/1135543002
Cr-Commit-Position: refs/heads/master@{#28318}
This revives the Terminate operator and removes the weird Always
operator. As a first step we let the ControlReducer connect non
terminating loops via Terminate. The next step will be to change the
graph builder to insert Terminate nodes into every loop.
Review URL: https://codereview.chromium.org/1123213002
Cr-Commit-Position: refs/heads/master@{#28259}
An AdvancedReducer is basically a regular Reducer with an editor
that can perform graph editing operations beyond changing or
replacing the node that is currently being reduced. The GraphReducer
is the default implementation of the AdvancedReducer::Editor interface.
The ControlReducerImpl is now just an AdvancedReducer, which
temporarily requires a Finish method in the reducer to implement
the dead node trimming until we move that to the GraphReducer
(which in turn requires that all loops are connected to End).
Review URL: https://codereview.chromium.org/1122423003
Cr-Commit-Position: refs/heads/master@{#28251}
Tail calls are matched on the graph, with a dedicated tail call
optimization that is actually testable. The instruction selection can
still fall back to a regular if the platform constraints don't allow to
emit a tail call (i.e. the return locations of caller and callee differ
or the callee takes non-register parameters, which is a restriction that
will be removed in the future).
Also explicitly limit tail call optimization to stubs for now and drop
the global flag.
BUG=v8:4076
LOG=n
Review URL: https://codereview.chromium.org/1114163005
Cr-Commit-Position: refs/heads/master@{#28219}
This introduces a simplified allocation operator which can be used to
model inline allocations in TurboFan. It is currently used for context
allocations, but still disabled because change lowering introduces
floating allocations outside the effect chain that interfere.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/1109773002
Cr-Commit-Position: refs/heads/master@{#28195}
This CL contains the first steps towards tail call optimization:
* Structurally detect tail calls during instruction selection,
looking for special return/call combinations.
* Added new architecture-specific instructions for tail calls which
jump instead of call and take care of frame adjustment.
* Moved some code around.
Currently we restrict tail calls to callees which only use registers
for arguments/return value and to call sites which are explicitly
marked as being OK for tail calls. This excludes, among other things,
call sites in sloppy JS functions and our IC machinery (both need in
general to be able to access the caller's frame).
All this is behind a flag --turbo-tail-calls, which is currently off
by default, so it can easily be toggled.
Review URL: https://codereview.chromium.org/1108563002
Cr-Commit-Position: refs/heads/master@{#28150}
- allows the optimization of emitted gap move code since the representation of the value in the register is known
- necessary preparation for vector register allocation
- prepare for slot sharing for any value of the same byte width
TBR=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1111323003
Cr-Commit-Position: refs/heads/master@{#28140}
- allows the optimization of emitted gap move code since the representation of the value in the register is known
- necessary preparation for vector register allocation
- prepare for slot sharing for any value of the same byte width
BUG=
Review URL: https://codereview.chromium.org/1087793002
Cr-Commit-Position: refs/heads/master@{#28137}
These two test cases "InlineCreateArrayLiteral" and "InlineCreateObjectLiteral" are
added in d1597b7d22. They invokes Linkage::GetStubCallDescriptor
which is unimplemented for turbofan unsupported platform.
BUG=
Review URL: https://codereview.chromium.org/1095793007
Cr-Commit-Position: refs/heads/master@{#28090}
This uses explicit operators instead of intrinsic runtime calls to
create literals froms boilerplates. It allows for easier access of
static parameters and syncs it with other allocating operators.
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/1104453006
Cr-Commit-Position: refs/heads/master@{#28062}
This introduces a JSCreateClosure operator which can be lowered by the
typed pipeline to the aforementioned stub. It also allows for further
optimizations of closure creation.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/1105513002
Cr-Commit-Position: refs/heads/master@{#28058}
Implements the strong mode proposal's restrictions on
implicit conversions for binary arithmetic operations, not
including the + special case. Adds some infrastructure
for future implementation of the restrictions for other
operators.
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/1092353002
Cr-Commit-Position: refs/heads/master@{#28045}
Now all nodes that care about deoptimization always take frame state
inputs no matter whether deoptimization is enabled for a particular
function. In case that deoptimization is off, the AstGraphBuilder just
inserts the empty frame state. This greatly simplifies the logic in
various places and makes testing easier as well, and is probably the
first step towards enabling --turbo-deoptimization by default.
There seems to be no noticable performance impact on asm.js programs.
Also fix the graph replay in order to regenerate the scheduler unittests.
Review URL: https://codereview.chromium.org/1106613003
Cr-Commit-Position: refs/heads/master@{#28026}
This allows the JSIntrinsicLowering to optimize the cloning of literal
boilerplate objects using either the FastCloneShallowArrayStub or the
FastCloneShallowObjectStub when applicable.
R=mvstanton@chromium.org
Review URL: https://codereview.chromium.org/1097963002
Cr-Commit-Position: refs/heads/master@{#27974}
- make ParallelMove into a ZoneVector, removing an annoying level of indirection
- make MoveOperands hold InstructionOperands instead of pointers, so there's no more operand aliasing for moves
- opens up possibility of storing MachineType in allocated operands
R=bmeurer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1081373002
Cr-Commit-Position: refs/heads/master@{#27842}
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}
- ConstantOperand was using a too-small field too store its virtual register
- drop ConvertTo, replace it with simple copy
- split AllocatedOperand off from Immediate and Constant to make assignment clearer, also paving the way for small Immediates
- put zone first in *Operand::New
- driveby: drop delayed ssa deconstruction experiment
R=titzer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1050803002
Cr-Commit-Position: refs/heads/master@{#27692}
Port 9af9f1d026
Original commit message:
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).
BUG=
Review URL: https://codereview.chromium.org/1073463003
Cr-Commit-Position: refs/heads/master@{#27679}
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}
This enables eager optimization of top-level code with TurboFan and
extends test coverage by triggering it with the --always-opt flag.
Script contexts are now also properly allocated in TurboFan.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/1053063003
Cr-Commit-Position: refs/heads/master@{#27633}
Support sxtb and sxth extend operators on add and subtract, as we've
done for ubtx/h. This is similar to ARM support for sxtab/h.
BUG=
Review URL: https://codereview.chromium.org/1064813003
Cr-Commit-Position: refs/heads/master@{#27624}
This CL primarily makes the loop peeling algorithm more robust; it no longer damages the graph if the loops are improperly closed.
R=bmeurer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1052753004
Cr-Commit-Position: refs/heads/master@{#27620}
We can use xorps/xorpd on Intel CPUs to flip the sign bit. Ideally we'd
use an absolute 128-bit constant in the code object, as OCaml/GCC
does, however that requires 128-bit alignment for code objects,
which is not yet implemented. So for now we materialize the mask
inline.
As drive-by-fix, don't hardcode xmm0 as scratch double register.
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/1064833002
Cr-Commit-Position: refs/heads/master@{#27618}
We can use xorps/xorpd on Intel CPUs to flip the sign bit. Ideally we'd
use a RIP-relative 128-bit constant in the code object, as OCaml/GCC
does, however that requires 128-bit alignment for code objects, which is
not yet implemented. So for now we materialize the mask inline.
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/1046893002
Cr-Commit-Position: refs/heads/master@{#27611}
Currently, this doesn't really help to generate better code,
nevertheless this is the right thing to do. When our type system(s)
are fixed, this should avoid falling back to floating point operations
in various cases.
Review URL: https://codereview.chromium.org/1057813002
Cr-Commit-Position: refs/heads/master@{#27578}
This changes the IrOpcode::kThrow operator to have kNoThrow property,
which sounds unintuitive, but holds for our graphs. The operators is
used to indicate exceptional control flow out of the function, but in
itself does not throw, the throwing is done by a runtime call.
R=titzer@chromium.org
TEST=unittests/CommonOperatorTest/CommonSharedOperatorTest.Properties
Review URL: https://codereview.chromium.org/1046173002
Cr-Commit-Position: refs/heads/master@{#27541}
Currently this only sets branch hints, so we get unlikely code "out of
the way", but in the long run the register allocator needs some love
to treat the unlikely code as, well, unlikely. :-)
Review URL: https://codereview.chromium.org/1048063002
Cr-Commit-Position: refs/heads/master@{#27540}
This adds the basics necessary to support float32 operations in TurboFan.
The actual functionality required to detect safe float32 operations will
be added based on this later. Therefore this does not affect production
code except for some cleanup/refactoring.
In detail, this patchset contains the following features:
- Add support for float32 operations to arm, arm64, ia32 and x64
backends.
- Add float32 machine operators.
- Add support for float32 constants to simplified lowering.
- Handle float32 representation for phis in simplified lowering.
In addition, contains the following (related) cleanups:
- Fix/unify naming of backend instructions.
- Use AVX comparisons when available.
- Extend ArchOpcodeField to 9 bits (required for arm64).
- Refactor some code duplication in instruction selectors.
BUG=v8:3589
LOG=n
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/1044793002
Cr-Commit-Position: refs/heads/master@{#27509}
With this change, we remember the types of frame state inputs (in a new
operator, called TypedStateValues). Instead of inferring the value types
when building translations, we used the recorded types.
The original approach was not reliable because the passes after
simplified lowering can change node types, and this in turn confuses
the translation builder.
BUG=chromium:468727
LOG=n
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/1015423002
Cr-Commit-Position: refs/heads/master@{#27310}
Add support for appending extend modes uxtb or uxth to add and subtract
instructions, and using them in the instruction selector.
BUG=
Review URL: https://codereview.chromium.org/1021533002
Cr-Commit-Position: refs/heads/master@{#27303}
This reduces the overhead of typed lowering, because we lower
JSToBoolean/JSUnaryNot directly if possible, instead of first lowering
to AnyToBoolean, and then letting the SimplifiedOperatorReducer do the
further lowering.
Also remove some obsolete tests from the cctest suite that have since
been removed by proper unittests. And improve unitttest coverage for the
typed lowering cases.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/999173003
Cr-Commit-Position: refs/heads/master@{#27295}
The control input is only relevant for operations that may "write" (to
prevent hoisting) or "throw" (because they are part of the control chain).
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1015353004
Cr-Commit-Position: refs/heads/master@{#27288}
The change introduces a second frame state (for the state before
the operation) for the StoreProperty nodes. If the store writes
into a typed array, the frame state is used for lazy deopt from
the to-number conversion that is performed by the store.
BUG=v8:3963
LOG=n
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/997983004
Cr-Commit-Position: refs/heads/master@{#27285}
- Use representation information provided by the type system to skip SMI
checks.
- Fix combining of ChangeTaggedToFloat64 with JSToNumber now that JS
operators can produce control.
- Remove the unnecessary abstraction of smi/field offsets.
- Improve unit test coverage.
- Various cosmetic fixes.
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/1018873002
Cr-Commit-Position: refs/heads/master@{#27250}
This change introduces a liveness analyzer for local variables in frame states.
The main idea is to use the AstGraphBuilder::Environment class to build the control flow graph, and record local variable loads, stores and checkpoints in the CFG basic blocks (LivenessAnalyzerBlock class).
After the graph building finishes, we run a simple data flow analysis over the CFG to figure out liveness of each local variable at each checkpoint. Finally, we run a pass over all the checkpoints and replace dead local variables in the frame states with the 'undefined' value.
Performance numbers for Embenchen are below.
----------- box2d.js
Current --turbo-deoptimization: EmbenchenBox2d(RunTime): 11265 ms.
d8-master --turbo-deoptimization: EmbenchenBox2d(RunTime): 11768 ms.
d8-master: EmbenchenBox2d(RunTime): 10996 ms.
----------- bullet.js
Current --turbo-deoptimization: EmbenchenBullet(RunTime): 17049 ms.
d8-master --turbo-deoptimization: EmbenchenBullet(RunTime): 17384 ms.
d8-master: EmbenchenBullet(RunTime): 16153 ms.
----------- copy.js
Current --turbo-deoptimization: EmbenchenCopy(RunTime): 4877 ms.
d8-master --turbo-deoptimization: EmbenchenCopy(RunTime): 4938 ms.
d8-master: EmbenchenCopy(RunTime): 4940 ms.
----------- corrections.js
Current --turbo-deoptimization: EmbenchenCorrections(RunTime): 7068 ms.
d8-master --turbo-deoptimization: EmbenchenCorrections(RunTime): 6718 ms.
d8-master: EmbenchenCorrections(RunTime): 6858 ms.
----------- fannkuch.js
Current --turbo-deoptimization: EmbenchenFannkuch(RunTime): 4167 ms.
d8-master --turbo-deoptimization: EmbenchenFannkuch(RunTime): 4608 ms.
d8-master: EmbenchenFannkuch(RunTime): 4149 ms.
----------- fasta.js
Current --turbo-deoptimization: EmbenchenFasta(RunTime): 9981 ms.
d8-master --turbo-deoptimization: EmbenchenFasta(RunTime): 9848 ms.
d8-master: EmbenchenFasta(RunTime): 9640 ms.
----------- lua_binarytrees.js
Current --turbo-deoptimization: EmbenchenLuaBinaryTrees(RunTime): 11571 ms.
d8-master --turbo-deoptimization: EmbenchenLuaBinaryTrees(RunTime): 13089 ms.
d8-master: EmbenchenLuaBinaryTrees(RunTime): 10957 ms.
----------- memops.js
Current --turbo-deoptimization: EmbenchenMemOps(RunTime): 7766 ms.
d8-master --turbo-deoptimization: EmbenchenMemOps(RunTime): 7346 ms.
d8-master: EmbenchenMemOps(RunTime): 7738 ms.
----------- primes.js
Current --turbo-deoptimization: EmbenchenPrimes(RunTime): 7459 ms.
d8-master --turbo-deoptimization: EmbenchenPrimes(RunTime): 7453 ms.
d8-master: EmbenchenPrimes(RunTime): 7451 ms.
----------- skinning.js
Current --turbo-deoptimization: EmbenchenSkinning(RunTime): 15564 ms.
d8-master --turbo-deoptimization: EmbenchenSkinning(RunTime): 15611 ms.
d8-master: EmbenchenSkinning(RunTime): 15583 ms.
----------- zlib.js
Current --turbo-deoptimization: EmbenchenZLib(RunTime): 10825 ms.
d8-master --turbo-deoptimization: EmbenchenZLib(RunTime): 11180 ms.
d8-master: EmbenchenZLib(RunTime): 10823 ms.
BUG=
Review URL: https://codereview.chromium.org/949743002
Cr-Commit-Position: refs/heads/master@{#27232}
Instead of the current approach of storing flat vectors in frame states (and possibly reusing the last vector in AST graph builder), this change list builds a tree for the values and tries to reuse the nodes for different frame states. At the moment, we only use this for the local variable part of frame state, but nothing prevents us from using this for all parts.
This change provides two new classes: one for creating the tree (StateValuesCache) and one for iterating the trees (StateValuesAccess).
BUG=
Review URL: https://codereview.chromium.org/1008213002
Cr-Commit-Position: refs/heads/master@{#27222}
The CL enables the same instructions are selected for Word32 and Word64 compare
operations which is possible due to a fact 32-bit inputs and produced values
are always sign-extended.
TEST=
BUG=
Review URL: https://codereview.chromium.org/1005123002
Cr-Commit-Position: refs/heads/master@{#27212}
Reduces the amount of custom support code for Math functions in TurboFan
and allows for more general inlining (i.e. independent of parameter
types).
BUG=v8:3952
LOG=n
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1004083002
Cr-Commit-Position: refs/heads/master@{#27172}
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}
Provide an intrinsic %MathFloor / %_MathFloor that is used to optimize
both Math.ceil and Math.floor, and use the JS inlining mechanism to
inline Math.ceil into TurboFan code. Although we need to touch code
outside of TurboFan to make this work, this does not affect the way we
handle Math.ceil and/or Math.floor in CrankShaft, because for CrankShaft
the old-style builtin function id based inlining still kicks in first.
Once this solution is stabilized, we can use it for Math.floor as well.
And once that is settled, we can establish it as the unified way to
inline builtins, and get rid of the specialized builtin function id
based inlining at some point.
Note that "builtin" applies to basically every piece of internal
JavaScript/intrinsics based code, so this also applies to the yet to be
defined JavaScript based code stubs and handlers.
BUG=v8:3953
LOG=n
R=yangguo@chromium.org,svenpanne@chromium.org
Review URL: https://codereview.chromium.org/990963003
Cr-Commit-Position: refs/heads/master@{#27086}
This adds support for the double bits intrinsics to TurboFan, and is
a first step towards fast Math functions inlined into TurboFan code
or even compiled by themselves with TurboFan.
Review URL: https://codereview.chromium.org/974313002
Cr-Commit-Position: refs/heads/master@{#27006}
Without this shortcut we generate one BooleanNot per JSUnaryNot with
number input, which is quite common in asm.js. These dead nodes then
survive until the late control reducer runs, and may prevent
optimizations in the mean time.
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/963713002
Cr-Commit-Position: refs/heads/master@{#26911}
This implements a special case of block cloning to recognize constructs like
if (a ? b : c) { ... }
that happen to be generated by Emscripten quite often.
Review URL: https://codereview.chromium.org/947963002
Cr-Commit-Position: refs/heads/master@{#26808}
This adds a new ControlFlowOptimizer that - for now - recognizes chains
of Branches generated by the SwitchBuilder for a subset of javascript
switches into Switch nodes. Those Switch nodes are then lowered to
either table or lookup switches.
Also rename Case to IfValue (and introduce IfDefault) for consistency.
BUG=v8:3872
LOG=n
Review URL: https://codereview.chromium.org/931623002
Cr-Commit-Position: refs/heads/master@{#26691}
A CompilationInfo constructed from just an Isolate* and a Zone* is in
weird an inconsistent state (calling e.g. flags() on it will crash),
so we need to avoid them. This CL removes almost all of them, the
remaining 2 call sites in (for testing only) will be handled in a
separate CL. Things which have been changed:
* Linkage is basically a decorator for CallDescriptor now.
* ChangeLowering doesn't need Linkage at all.
* JSGenericLowering doesn't need a full CompilationInfo*, just a
single flag.
* JSContextSpecializer doesn't need the full CompilationInfo, just a
Context.
* Removed unused CompilationInfo from SimplifiedLoweringTester.
This nicely decouples things already a bit more, but there's still
work to do...
Review URL: https://codereview.chromium.org/899803003
Cr-Commit-Position: refs/heads/master@{#26580}
Adds Switch and Case operators to TurboFan and handles them
appropriately in instruction selection and code generation.
BUG=v8:3872
LOG=n
Review URL: https://codereview.chromium.org/892513003
Cr-Commit-Position: refs/heads/master@{#26515}
Contribution of PowerPC port (continuation of 422063005, 817143002 and 866843003)
This patch brings the ppc directories up to date with our repo. We have
removed 5 individual optimizations which require changes in both the ppc and common
directories so they can be more easily reviewed on their own in subsequent patches.
Subsequent patches will cover:
- individual optimizations for PPC (5)
- remaining AIX changes not resolved by 4.8 compiler (4.8 is only recently available
for AIX)
- incremental updates required to ppc directories due to platform specific changes
made in google repos while we complete the above steps.
With the update there are still some timeouts seen when run in simulated mode which
may be a result of the missing optimizations. Once we have the optimizations in
we will review the simulation results and address/exclude tests as necessary so that
the simulated runs are clean.
new file: src/compiler/ppc/code-generator-ppc.cc
new file: src/compiler/ppc/instruction-codes-ppc.h
new file: src/compiler/ppc/instruction-selector-ppc.cc
new file: src/compiler/ppc/linkage-ppc.cc
modified: src/ic/ppc/handler-compiler-ppc.cc
modified: src/ic/ppc/ic-compiler-ppc.cc
modified: src/ic/ppc/ic-ppc.cc
modified: src/ic/ppc/stub-cache-ppc.cc
modified: src/ppc/assembler-ppc.cc
modified: src/ppc/assembler-ppc.h
modified: src/ppc/builtins-ppc.cc
modified: src/ppc/code-stubs-ppc.cc
modified: src/ppc/code-stubs-ppc.h
modified: src/ppc/codegen-ppc.cc
modified: src/ppc/constants-ppc.h
modified: src/ppc/deoptimizer-ppc.cc
modified: src/ppc/disasm-ppc.cc
modified: src/ppc/full-codegen-ppc.cc
modified: src/ppc/interface-descriptors-ppc.cc
modified: src/ppc/lithium-codegen-ppc.cc
modified: src/ppc/lithium-codegen-ppc.h
modified: src/ppc/lithium-ppc.cc
modified: src/ppc/lithium-ppc.h
modified: src/ppc/macro-assembler-ppc.cc
modified: src/ppc/macro-assembler-ppc.h
modified: src/ppc/regexp-macro-assembler-ppc.cc
modified: src/ppc/regexp-macro-assembler-ppc.h
modified: src/ppc/simulator-ppc.cc
modified: src/ppc/simulator-ppc.h
new file: test/unittests/compiler/ppc/instruction-selector-ppc-unittest.cc
R=danno@chromium.org, svenpanne@chromium.org
BUG=
Review URL: https://codereview.chromium.org/901083004
Cr-Commit-Position: refs/heads/master@{#26471}
It doesn't do anything for now, but it implies strict mode. Added tests to
test-parsing.cc to test that.
BUG=
Review URL: https://codereview.chromium.org/898983002
Cr-Commit-Position: refs/heads/master@{#26460}
This enables adding more language modes in the future.
For maximum flexibility, LanguageMode is a bitmask, so we're not restricted to
use a sequence of language modes which are progressively stricter, but we can
express the language mode as combination of features.
For now, LanguageMode can only be "sloppy" or "strict", and there are
STATIC_ASSERTS in places which need to change when more modes are added.
LanguageMode is a bit like the old LanguageMode when "extended" mode was still
around (see https://codereview.chromium.org/8417035 and
https://codereview.chromium.org/181543002 ) except that it's transmitted through
all the layers (there's no StrictModeFlag).
BUG=
Review URL: https://codereview.chromium.org/894683003
Cr-Commit-Position: refs/heads/master@{#26419}
If a (pure) node has two or more uses, but there exists a path from the
common dominator of these uses to end, which does not contain a use,
then we split the node such that no unnecessary computation takes place.
Note however, that this only applies if the node cannot be hoisted out
of a loop.
BUG=v8:3864
LOG=n
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/899433005
Cr-Commit-Position: refs/heads/master@{#26404}
Up until now we used a special Terminate node to artifically connect non
terminating loops to the End node, but this was kind of adhoc and didn't
work for the CFG. So without all kinds of weird hacks, the end block in
the CFG will not be connected to NTLs, which makes it impossible to
compute post dominance / control dependence in the current setting.
So instead of Terminate, we add a special Branch to NTLs, whose
condition is the special Always node, which corresponds to True, except
that it cannot be folded away. This way we don't need any special
machinery in the scheduler, since it's just a regular Branch.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/875263004
Cr-Commit-Position: refs/heads/master@{#26294}
Along the way:
- Thread isolate parameter explicitly through code that used to
rely on getting it from the zone.
- Canonicalize the parameter position of isolate and zone for
affected code
- Change Hydrogen New<> instruction templates to automatically
pass isolate
R=mstarzinger@chromium.org
LOG=N
Review URL: https://codereview.chromium.org/868883002
Cr-Commit-Position: refs/heads/master@{#26252}
Reason for revert:
Breaks test-types/Maybe, i.e.
out/Release/cctest --random-seed=-707413401 test-types/Maybe
started failing afterwards
Original issue's description:
> Steps towards unification of number bitset and range types.
>
> - New invariant on union types: if the union has a range then the number
> bits in the bitset must be cleared.
>
> - Various tweaks in intersection and union to satisfy the invariant.
>
> - Exposed and used representation bits in range types (and the Limits
> helper class).
>
> - Implemented Glb for ranges so that the Is predicate handles
> ranges correctly.
>
> - Change typer weakening so that it does not rely on GetRange.
> However, the code still seems to be a bit fragile.
>
> - Removed the Smi types from the type system core, instead introduced
> Signed31, Unsigned30 and created constructors for Small(Un)Signed
> that point to the right type for the architecture.
>
> - Punched a hole in the config to be able to get to the isolate so
> that it is possible to allocate heap numbers for newly created
> ranges.
>
> Patch by jarin@chromium.prg, original review here:
> https://codereview.chromium.org/795713003/
>
> TBR=jarin@chromium.org
> BUG=
>
> Committed: https://crrev.com/2764fd8d1a266a9136c987c2483492113b0c8d80
> Cr-Commit-Position: refs/heads/master@{#26197}
TBR=jkummerow@chromium.org,rossberg@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/868583002
Cr-Commit-Position: refs/heads/master@{#26207}
- New invariant on union types: if the union has a range then the number
bits in the bitset must be cleared.
- Various tweaks in intersection and union to satisfy the invariant.
- Exposed and used representation bits in range types (and the Limits
helper class).
- Implemented Glb for ranges so that the Is predicate handles
ranges correctly.
- Change typer weakening so that it does not rely on GetRange.
However, the code still seems to be a bit fragile.
- Removed the Smi types from the type system core, instead introduced
Signed31, Unsigned30 and created constructors for Small(Un)Signed
that point to the right type for the architecture.
- Punched a hole in the config to be able to get to the isolate so
that it is possible to allocate heap numbers for newly created
ranges.
Patch by jarin@chromium.prg, original review here:
https://codereview.chromium.org/795713003/TBR=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/837723006
Cr-Commit-Position: refs/heads/master@{#26197}
Use std::numeric_limits<double>::quiet_NaN() and
std::numeric_limits<float>::quiet_NaN() instead.
Review URL: https://codereview.chromium.org/864803002
Cr-Commit-Position: refs/heads/master@{#26195}
- Make Node::Inputs and Node::Uses mostly STL compliant.
- Get rid of some pre-C++11 crappiness.
- Start moving unit tests from cctest to unittests.
- TrimInputCount() now tries to reserve inputs slots for
later appending.
- Fix numerous style guide violations.
TEST=cctest,unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/851263002
Cr-Commit-Position: refs/heads/master@{#26098}
- Use C++11 range based for loops.
- Remove duplicated virtual register set in unittests.
- Don't expose implementation details of InstructionSelector.
TEST=unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/837423002
Cr-Commit-Position: refs/heads/master@{#25997}
- 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}
Also support additional number types (singleton ranges, NaN and minus
zero) for constant propagation in typed lowering.
TEST=unittests
Review URL: https://codereview.chromium.org/829303002
Cr-Commit-Position: refs/heads/master@{#25963}
Introduce a new AnyToBoolean simplified operator to handle the later
lowering of boolean conversions. Previously we tried to hack that with
the generic JSToBoolean, having its context set to zero, but that lead
to various problems/bugs and did not handle all cases.
TEST=cctest,unittests
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/800833003
Cr-Commit-Position: refs/heads/master@{#25958}
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}
This greatly reduces the number of nodes in the graph (by more than 20x in
some extreme cases) for the Emscripten python interpreter main function.
BUG=v8:3763
LOG=y
TEST=cctest,mjsunit,unittests
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/802353003
Cr-Commit-Position: refs/heads/master@{#25840}
Relanded with fix for always returning Change for PlainPrimitive even
if there was no change. The performance regression on primes.js and
corrections.js is due to unlucky loop header alignment; will be addressed
separately.
TEST=unittests
R=svenpanne@chromium.org
Committed: 75484e8d16
Review URL: https://codereview.chromium.org/799413002
Cr-Commit-Position: refs/heads/master@{#25832}
The instruction selector now selects pseudo instructions: CompareAndBranch or
TestAndBranch which are associated with their continuations so that generic
code in the code generator will treat them as branch instruction and will be
able to apply optimization like avoiding branches when the code can falltrhough.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/798553002
Cr-Commit-Position: refs/heads/master@{#25773}
Reason for revert:
For breaking the waterfall (run-json-stringify test).
Original issue's description:
> Avoid number range holes in bitset types.
>
> BUG=
TBR=rossberg@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/794663002
Cr-Commit-Position: refs/heads/master@{#25756}
Also fix the pushing of JSToBoolean into Phis and generalize it to
also include pushing into Selects.
TEST=cctest,unittests
Review URL: https://codereview.chromium.org/792463003
Cr-Commit-Position: refs/heads/master@{#25718}
Achieve more than parity with modes currently handled on ia32 in preparation for
porting the entire mechanism to ia32, including supporting mul of constants 3,
5 and 9 with "leal" instructions.
TEST=unittests
Review URL: https://codereview.chromium.org/774193003
Cr-Commit-Position: refs/heads/master@{#25677}
This is an initial version of redundant load elimination, currently
limited to LoadField operators, and implemented by walking the effect
chain.
TEST=unittests
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/782473002
Cr-Commit-Position: refs/heads/master@{#25673}
Only use "addl" and "subl" in cases that have been measured to be
faster (currently only immediate operations).
Review URL: https://codereview.chromium.org/735293004
Cr-Commit-Position: refs/heads/master@{#25580}
Resets the scaled exponent to 0 when the scaling match fails.
BUG=
Review URL: https://codereview.chromium.org/756643002
Cr-Commit-Position: refs/heads/master@{#25491}
They generally cause regressions on most modern Intel chips. Replace them with
addl/subl.
Review URL: https://codereview.chromium.org/737153003
Cr-Commit-Position: refs/heads/master@{#25466}
- Use "leal" for subtraction of integer constant when non-constant input to
subtract is used more than once.
- Use "incl", "decl", and "addl" when they are smaller/cheaper than their
leal/addl/subl equivalant.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/738073002
Cr-Commit-Position: refs/heads/master@{#25439}
JavaScript shifts perform an implicit '& 0x1F' on their right operand, this
patch removes it when the underlying architecture already does it.
BUG=
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/732103002
Cr-Commit-Position: refs/heads/master@{#25438}
Don't use the generic algorithm, but instead start going into the
direction of ControlReducer, using a stack plus a revisit queue to
not miss any more possibilities for reductions anymore.
TEST=cctest,unittests
R=dcarney@chromium.org
Committed: f047507370
Committed: 6e148989a4
Review URL: https://codereview.chromium.org/726513002
Cr-Commit-Position: refs/heads/master@{#25377}
Don't use the generic algorithm, but instead start going into the
direction of ControlReducer, using a stack plus a revisit queue to
not miss any more possibilities for reductions anymore.
TEST=cctest,unittests
R=dcarney@chromium.org
Committed: f047507370
Review URL: https://codereview.chromium.org/726513002
Cr-Commit-Position: refs/heads/master@{#25345}
Don't use the generic algorithm, but instead start going into the
direction of ControlReducer, using a stack plus a revisit queue to
not miss any more possibilities for reductions anymore.
TEST=cctest,unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/726513002
Cr-Commit-Position: refs/heads/master@{#25326}
There's no need to apply the 0x1f mask to right hand sides of shifts if
the input is already in range [0,31].
TEST=cctest,unittests
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/718193003
Cr-Commit-Position: refs/heads/master@{#25313}
Lowering of NumberToUint32 and NumberToInt32 was not correctly accounting for the sign of the input and the sign of the output, emitting the wrong representation changes.
Along the way, I've found cases where MachineOperatorBuilder would break if fed a machine type for loads or stores that was not cached, requiring MachineOperatorBuilder to take zone to allocate operators for these cases.
R=bmeurer@chromium.org, jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/714613002
Cr-Commit-Position: refs/heads/master@{#25247}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25247 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This contains the following changes squashed together:
- Switch BasicBlock::loop_end to be a basic block instead of an RPO.
- Switch ScheduleLate to use dominator depth instead of RPO.
- Switch ScheduleEarly to use dominator depth instead of RPO.
- Push out absolute RPO ordering everywhere else in the scheduler.
- Keep linked list of blocks in RPO order while scheduling.
- Switch from RPO number to depth for dominator calculation.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/696363002
Cr-Commit-Position: refs/heads/master@{#25138}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25138 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Branch can now have an optional hint, when the condition is
likely true or false, and if such a hint is present the other
basic block will be marked as deferred and placed at the end
of the function.
We currently use this feature for tagging int32/uint32 in
change lowering, and for load/store bounds checks in simplified
lowering.
TEST=cctest,unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/642883003
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24802 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
- Extend verifier to check types of JS and Simplified nodes.
- Untyped nodes now contain NULL as types, enforcing hard failure.
- Typer immediately installs itself as a decorator; remove explicit decorator installation.
- Decorator eagerly types all nodes that have typed inputs
(subsumes typing of constant cache, removing its typing
side-channel and various spurious dependencies on the typer).
- Cut down typer interface to prevent inconsistently typed graphs.
- Remove verification from start, since it caused too much trouble
with semi-wellformed nodes.
- Fix a couple of bugs on the way that got uncovered.
To do: verifying machine operators. Also, various conditions in the
verifier are currently commented out, because they don't yet hold.
BUG=
R=jarin@chromium.org,titzer@chromium.org
Review URL: https://codereview.chromium.org/658543002
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24626 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This change enables non-tagged representations in FrameStates.
That allows us to run zlib with deoptimization support and have almost the same performance of the generated code (as the code with no deoptimization). Unfortunately, the frame states seem to confuse typer. As a consequence, we generate more representation changes, which in turn causes the scheduler to take a lot more time and memory (>4x). The added compiler time makes zlib with deopt be about 50% slower.
BUG=
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/614713002
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24454 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The StaticParameterTraits are broken by design, and cause way too much
trouble. The compilers usually pick the wrong specialization (i.e. the
default specialization is picked for Load and Phi even tho there is a
specialization for MachineType), which is not only the reason why GVN is
ineffective and slow, but can also lead to correctness issues in some
rare cases.
Also clean up some minor bugs/inconsistencies on the way.
TEST=cctest,unittests
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/636893002
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24437 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
