Introduces:
- a new AST node representing the GetIterator() algorithm in the specification, to be used by ForOfStatement, YieldExpression (in the case of delegating yield*), and the future `for-await-of` loop proposed in http://tc39.github.io/proposal-async-iteration/#sec-async-iterator-value-unwrap-functions.
- a new opcode (JumpIfJSReceiver), which is useful for `if Type(object) is not Object` checks which are common throughout the specification. This node is easily eliminated by TurboFan.
The AST node is desugared specially in bytecode, rather than manually when building the AST. The benefit of this is that desugaring in the BytecodeGenerator is much simpler and easier to understand than desugaring the AST.
This also reduces parse time very slightly, and allows us to use LoadIC rather than KeyedLoadIC, which seems to have better baseline performance. This results in a ~20% improvement in test/js-perf-test/Iterators micro-benchmarks, which I believe owes to the use of the slightly faster LoadIC as opposed to the KeyedLoadIC in the baseline case. Both produce identical optimized code via TurboFan when the type check can be eliminated, and the load can be replaced with a constant value.
BUG=v8:4280
R=bmeurer@chromium.org, rmcilroy@chromium.org, adamk@chromium.org, neis@chromium.org, jarin@chromium.orgTBR=rossberg@chromium.org
Review-Url: https://codereview.chromium.org/2557593004
Cr-Commit-Position: refs/heads/master@{#41555}
This moves the location of the bytecode-offset translation that turns
offsets of back jumps into offsets of loop headers. This translation is
now done by the {BytecodeGraphBuilder} after loop analysis has been
performed. It safes one redudant iteration over the bytecode array. Note
that this changes the semantics of the BailoutId used as an {osr_ast_id}
throughout the compiler pipeline for OSR from Ignition.
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2465913002
Cr-Commit-Position: refs/heads/master@{#41431}
Replaces the graph-based liveness analyzer in the bytecode graph builder
with an initial bytecode-based liveness analysis pass, which is added to
the existing loop extent analysis.
Now the StateValues in the graph have their inputs initialised to
optimized_out, rather than being modified after the graph is built.
Review-Url: https://codereview.chromium.org/2523893003
Cr-Commit-Position: refs/heads/master@{#41355}
Reason for revert:
Breaks the build:
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20shared/builds/14886
Original issue's description:
> [ignition/turbo] Perform liveness analysis on the bytecodes
>
> Replaces the graph-based liveness analyzer in the bytecode graph builder
> with an initial bytecode-based liveness analysis pass, which is added to
> the existing loop extent analysis.
>
> Now the StateValues in the graph have their inputs initialised to
> optimized_out, rather than being modified after the graph is built.
>
> Committed: https://crrev.com/1852300954c216c29cf93444430681d213e87925
> Cr-Commit-Position: refs/heads/master@{#41344}
TBR=jarin@chromium.org,rmcilroy@chromium.org,yangguo@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review-Url: https://codereview.chromium.org/2541443002
Cr-Commit-Position: refs/heads/master@{#41346}
Replaces the graph-based liveness analyzer in the bytecode graph builder
with an initial bytecode-based liveness analysis pass, which is added to
the existing loop extent analysis.
Now the StateValues in the graph have their inputs initialised to
optimized_out, rather than being modified after the graph is built.
Review-Url: https://codereview.chromium.org/2523893003
Cr-Commit-Position: refs/heads/master@{#41344}
The AstGraphBuilder pipeline is only used for asm.js now, so the whole
type feedback mechanism is essentially dead code currently, thus we
better nuke it.
BUG=v8:5267,v8:5657
Review-Url: https://codereview.chromium.org/2523953002
Cr-Commit-Position: refs/heads/master@{#41201}
Now that we have a JumpLoop bytecode, we can heavily simplify the
branch/loop analysis by assuming that only JumpLoop bytecodes are
backwards edges, and performing the loop analysis as a single
(backwards) pass.
This allows us to get rid of the branch analysis entirely, and builds a
framework to do liveness analysis in the same pass.
Review-Url: https://codereview.chromium.org/2519983002
Cr-Commit-Position: refs/heads/master@{#41194}
Add bytecode for defining data properties, which initially just calls the runtime function.
BUG=v8:5624
Review-Url: https://codereview.chromium.org/2510743002
Cr-Commit-Position: refs/heads/master@{#41101}
The reasons are:
1) The names dictionaries in the feedback metadata seems to consume a lot of memory
and the idea didn't payoff.
2) The absence of a name parameter blocks data handlers support in LoadGlobalIC.
This CL reverts a part of r37278 (https://codereview.chromium.org/2096653003/).
BUG=chromium:576312, v8:5561
Review-Url: https://codereview.chromium.org/2510653002
Cr-Commit-Position: refs/heads/master@{#41046}
SourcePosition::InliningId() refers to a the new table DeoptimizationInputData::InliningPositions(), which provides the following data for every inlining id:
- The inlined SharedFunctionInfo as an offset into DeoptimizationInfo::LiteralArray
- The SourcePosition of the inlining. Recursively, this yields the full inlining stack.
Before the Code object is created, the same information can be found in CompilationInfo::inlined_functions().
If SourcePosition::InliningId() is SourcePosition::kNotInlined, it refers to the outer (non-inlined) function.
So every SourcePosition has full information about its inlining stack, as long as the corresponding Code object is known. The internal represenation of a source position is a positive 64bit integer.
All compilers create now appropriate source positions for inlined functions. In the case of Turbofan, this required using AstGraphBuilderWithPositions for inlined functions too. So this class is now moved to a header file.
At the moment, the additional information in source positions is only used in --trace-deopt and --code-comments. The profiler needs to be updated, at the moment it gets the correct script offsets from the deopt info, but the wrong script id from the reconstructed deopt stack, which can lead to wrong outputs. This should be resolved by making the profiler use the new inlining information for deopts.
I activated the inlined deoptimization tests in test-cpu-profiler.cc for Turbofan, changing them to a case where the deopt stack and the inlining position agree. It is currently still broken for other cases.
The following additional changes were necessary:
- The source position table (internal::SourcePositionTableBuilder etc.) supports now 64bit source positions. Encoding source positions in a single 64bit int together with the difference encoding in the source position table results in very little overhead for the inlining id, since only 12% of the source positions in Octane have a changed inlining id.
- The class HPositionInfo was effectively dead code and is now removed.
- SourcePosition has new printing and information facilities, including computing a full inlining stack.
- I had to rename compiler/source-position.{h,cc} to compiler/compiler-source-position-table.{h,cc} to avoid clashes with the new src/source-position.cc file.
- I wrote the new wrapper PodArray for ByteArray. It is a template working with any POD-type. This is used in DeoptimizationInputData::InliningPositions().
- I removed HInlinedFunctionInfo and HGraph::inlined_function_infos, because they were only used for the now obsolete Crankshaft inlining ids.
- Crankshaft managed a list of inlined functions in Lithium: LChunk::inlined_functions. This is an analog structure to CompilationInfo::inlined_functions. So I removed LChunk::inlined_functions and made Crankshaft use CompilationInfo::inlined_functions instead, because this was necessary to register the offsets into the literal array in a uniform way. This is a safe change because LChunk::inlined_functions has no other uses and the functions in CompilationInfo::inlined_functions have a strictly longer lifespan, being created earlier (in Hydrogen already).
BUG=v8:5432
Review-Url: https://codereview.chromium.org/2451853002
Cr-Commit-Position: refs/heads/master@{#40975}
We seem to get some small wins from avoiding the Ldr bytecodes, probably due
to reduced icache pressure since there are less bytecode handlers. Replace
the Ldr bytecodes with Star lookahead inlined into the Lda versions.
Also fixes IsAccumulatorLoadWithoutEffects to include LdaContextSlot and
LdaCurrentContextSlot
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2489513005
Cr-Commit-Position: refs/heads/master@{#40883}
The Ldr[Named/Keyed]Property bytecodes are problematic for the deoptimizer when
inlining accessors in TurboFan. Remove them and replace with a Star lookahead
in the bytecode handlers for Lda[Named/Keyed]Property.
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2485383002
Cr-Commit-Position: refs/heads/master@{#40860}
This removes the first stack check in inlined functions in the bytecode
graph builder, to match the behaviour of the AST graph builder.
I measure a ~1% statistically significant (p < 0.01) improvement on
Mandreel with --ignition-staging --turbo (on my x64 machine, YMMV).
Review-Url: https://codereview.chromium.org/2392333002
Cr-Commit-Position: refs/heads/master@{#40715}
The majority of context slot accesses are to the local context (current context
register and depth 0), so this adds bytecodes to optimise for that case.
This cuts down bytecode size by roughly 1% (measured on Octane and Top25).
Review-Url: https://codereview.chromium.org/2459513002
Cr-Commit-Position: refs/heads/master@{#40641}
This is a new bytecode which behaves (for now) exactly like Call,
except that in turbofan graph building we can set the
ConvertReceiverMode to NotNullOrUndefined.
I observe a 1% improvement on Box2D, I'd expect a similar improvement on
other OOP heavy code.
Review-Url: https://codereview.chromium.org/2450243002
Cr-Commit-Position: refs/heads/master@{#40610}
This makes the creation of frame states "before" and "after" node
sequences explicit in the {BytecodeGraphBuilder}. This removes some
complexity and also allows us to ellide redundant {Checkpoint} nodes
before operations that don't actually eager deoptimize.
In this change such redundant {Checkpoint} nodes have been removed for
arguments object and rest array creation bytecodes. The frame states
used in such {Checkpoint} nodes were actually bogus because they would
resume bytecode execution before the {new.target} value is assigned to
its respective variable.
R=jarin@chromium.org
Review-Url: https://chromiumcodereview.appspot.com/2437683003
Cr-Commit-Position: refs/heads/master@{#40453}
Adds a fast-path test and branch for the turbofan graph generated by
BytecodeGraphBuilder for dynamic global lookups.
BUG=v8:5263
Review-Url: https://codereview.chromium.org/2374983003
Cr-Commit-Position: refs/heads/master@{#39882}
Adds a fast path for loading DYNAMIC_GLOBAL variables, which are lookup
variables that can be globally loaded, without calling the runtime, as long as
there was no context extension by a sloppy eval along their context chain.
BUG=v8:5263
Review-Url: https://codereview.chromium.org/2347143002
Cr-Commit-Position: refs/heads/master@{#39537}
Adds a fast path for loading DYNAMIC_LOCAL variables, which are lookup
variables that can be context loaded, without calling the runtime, as
long as there was no context extension by a sloppy eval along their
context chain.
BUG=v8:5263
Review-Url: https://codereview.chromium.org/2343633002
Cr-Commit-Position: refs/heads/master@{#39473}
Add a notion of "invocation count" to the baseline compilers, which
increment a special slot in the TypeFeedbackVector for each invocation
of a given function (the optimized code doesn't currently collect this
information).
Use this invocation count to relativize the call counts on the call
sites within the function, so that the inlining heuristic has a view
of relative importance of a call site rather than some absolute numbers
with unclear meaning for the current function. Also apply the call site
frequency as a factor to all frequencies in the inlinee by passing this
to the graph builders so that the importance of a call site in an
inlinee is relative to the topmost optimized function.
Note that all functions that neither have literals nor need type
feedback slots will share a single invocation count cell in the
canonical empty type feedback vector, so their invocation count is
meaningless, but that doesn't matter since we only use the invocation
count to relativize call counts within the function, which we only have
if we have at least one type feedback vector (the CallIC slot).
See the design document for additional details on this change:
https://docs.google.com/document/d/1VoYBhpDhJC4VlqMXCKvae-8IGuheBGxy32EOgC2LnT8
BUG=v8:5267,v8:5372
R=mvstanton@chromium.org,rmcilroy@chromium.org,mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2337123003
Cr-Commit-Position: refs/heads/master@{#39410}
Extract the call counts from the type feedback vector during graph
building (either via the AstGraphBuilder or the BytecodeGraphBuilder),
and put them onto the JSCallFunction and JSCallConstruct operators,
so that they work even across inlinine through .apply and .call (which
was previously hacked by creating a temporary type feedback vector
for those).
The next logic step will be to make those call counts into real
relative call frequencies (also during graph building), so that we
can make inlining decisions that make sense for the function being
optimized (where absolute values are misleading).
R=jarin@chromium.org
BUG=v8:5267,v8:5372
Review-Url: https://codereview.chromium.org/2330883002
Cr-Commit-Position: refs/heads/master@{#39400}
This makes sure OSR entry points can be added without having special
{OsrPoll} instructions in the bytecode stream. Eventually we might end
up merging back-branches and OSR polls together. Any loop header can be
used as an OSR entry point.
R=bmeurer@chromium.org
BUG=v8:4764
Review-Url: https://codereview.chromium.org/2329933003
Cr-Commit-Position: refs/heads/master@{#39373}
This is analogous to the variable liveness analysis we do in the AstGraphBuilder, but on the bytecode registers.
BUG=
Review-Url: https://codereview.chromium.org/2307863002
Cr-Commit-Position: refs/heads/master@{#39248}
Avoid the useless strict equality comparisons with true/false being
generated for the JumpIfTrue, JumpIfFalse, JumpIfToBooleanTrue and
JumpIfToBooleanFalse bytecodes. Instead feed the accumulator (or the
outcome of ToBoolean) directly to the Branch node and do the negation
as part of the control flow.
The previous subraphs would render the loop variable analysis useless,
and would cause a lot of unnecessary bit materialization, because many
of our optimizations don't kick in.
Note: This is only part of the problem, there are more subtle differences
in the bytecode pipeline that prevent several important optimizations to
kick in.
R=mstarzinger@chromium.org
BUG=v8:5267,v8:5348
Review-Url: https://codereview.chromium.org/2309733002
Cr-Commit-Position: refs/heads/master@{#39151}
This way, many files which only need CompilationInfo but not compiler.h
and its dependencies can include just compilation-info.h.
BUG=
Review-Url: https://codereview.chromium.org/2284313003
Cr-Commit-Position: refs/heads/master@{#39038}
Drive-by fix: the order of parameters in the BinaryOpWithFeedback TurboFan code stubs now reflects the convention of having the context at the end.
BUG=v8:5273
Review-Url: https://codereview.chromium.org/2263253002
Cr-Commit-Position: refs/heads/master@{#38832}
Collect type feedback for subtract operation in interpreter. Also use it in
bytecode-graph-bulder to set the correct Hint for subtract operation.
BUG=v8:5273
LOG=N
Review-Url: https://codereview.chromium.org/2221833002
Cr-Commit-Position: refs/heads/master@{#38467}
This completely removes translation of exception handler predictions
from the graph IR. We now rely on the runtime using deoptimization
infomation via {FrameSummary} for predictions in optimized code.
R=bmeurer@chromium.org
Review-Url: https://codereview.chromium.org/2207533002
Cr-Commit-Position: refs/heads/master@{#38250}
This gets rid of the Star bytecodes that were always dispatched to from
ToObject.
ToObject now outputs to register instead of to the accumulator and
ForInPrepare gets the receiver object from an input register.
BUG=v8:4820
LOG=n
Review-Url: https://codereview.chromium.org/2189463006
Cr-Commit-Position: refs/heads/master@{#38177}
This implements graph construction for entry via on-stack replacement
within the {BytecodeGraphBuilder}. Entry points are at loop headers
similar to previous OSR implementations. All interpreter registers are
addressable via {OsrValue} nodes in the graph. Currently we rely on
{OsrPoll} bytecodes to be placed right after loop headers (i.e. at the
targets of back edges).
R=jarin@chromium.org
BUG=v8:4764
Review-Url: https://codereview.chromium.org/2171083004
Cr-Commit-Position: refs/heads/master@{#38083}
ToName was always generated with a subsequent Star, fuse them.
Requires a few changes in the peephole optimizer as ToName cannot be
elided as easily, but must be replaced by Star.
BUG=v8:4280
LOG=n
Review-Url: https://codereview.chromium.org/2169813002
Cr-Commit-Position: refs/heads/master@{#38019}
Reason for revert:
Breaks the roll, possibly win gn:
https://codereview.chromium.org/2148863002/
Original issue's description:
> [interpreter] Reduce dependencies in bytecodes.{h,cc}
>
> This CL reduces the number of dependencies bytecodes.{h,cc} to facilitate
> generating the bytecode peephole optimizer table during build. Specifically,
> it avoids depending on v8_base.
>
> BUG=v8:4280
> LOG=N
>
> Committed: https://crrev.com/4edebb1cd870ae6c1359ad54f83e618e185883b1
> Cr-Commit-Position: refs/heads/master@{#37715}
TBR=mstarzinger@chromium.org,rmcilroy@chromium.org,oth@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2151693003
Cr-Commit-Position: refs/heads/master@{#37743}
This CL reduces the number of dependencies bytecodes.{h,cc} to facilitate
generating the bytecode peephole optimizer table during build. Specifically,
it avoids depending on v8_base.
BUG=v8:4280
LOG=N
Review-Url: https://codereview.chromium.org/2135273002
Cr-Commit-Position: refs/heads/master@{#37715}
Introduces fused bytecodes for fusing LdaSmi followed by a binary op bytecode.
The chosen bytecodes are used frequently in Octane: AddSmi, SubSmi,
BitwiseOrSmi, BitwiseAndSmi, ShiftLeftSmi, ShiftRightSmi.
There are additional code stubs for these operations that are biased towards
both the left hand and right hand operands being Smis.
BUG=v8:4280
LOG=N
Review-Url: https://codereview.chromium.org/2111923002
Cr-Commit-Position: refs/heads/master@{#37531}
This change introduces five fused bytecodes for common bytecode
sequences on popular websites. These are LdrNamedProperty,
LdrKeyedProperty, LdrGlobal, LdrContextSlot, and LdrUndefined. These
load values into a destination register operand instead of the
accumulator. They are emitted by the peephole optimizer.
BUG=v8:4280
LOG=N
Review-Url: https://codereview.chromium.org/1985753002
Cr-Commit-Position: refs/heads/master@{#36507}
It's always possible to deoptimize to Ignition from TurboFan, and we
don't use Ignition for asm.js code, so there's no need to have the empty
frame state magic in the BytecodeGraphBuilder.
R=mstarzinger@chromium.org
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2008243002
Cr-Commit-Position: refs/heads/master@{#36505}
Introduces a bytecode whose handler executes the equivalent of %_IsArray and %_IsJSReceiver without a runtime call.
BUG=v8:4822
LOG=y
Review URL: https://codereview.chromium.org/1645763003
Cr-Commit-Position: refs/heads/master@{#34983}
This change introduces wide prefix bytecodes to support wide (16-bit)
and extra-wide (32-bit) operands. It retires the previous
wide-bytecodes and reduces the number of operand types.
Operands are now either scalable or fixed size. Scalable operands
increase in width when a bytecode is prefixed with wide or extra-wide.
The bytecode handler table is extended to 256*3 entries. The
first 256 entries are used for bytecodes with 8-bit operands,
the second 256 entries are used for bytecodes with operands that
scale to 16-bits, and the third group of 256 entries are used for
bytecodes with operands that scale to 32-bits.
LOG=N
BUG=v8:4747,v8:4280
Review URL: https://codereview.chromium.org/1783483002
Cr-Commit-Position: refs/heads/master@{#34955}
Reason for revert:
It is not a good idea to call CallICStub from the builtin. It might be sensitive to the frame structure. Constructing a internal frame might cause problems. It is much better to inline the code related to the type feedback vector into the builtin.
Original issue's description:
> [Interpreter] Implements calls through CallICStub in the interpreter.
>
> Calls are implemented through CallICStub to collect type feedback. Adds
> a new builtin called InterpreterPushArgsAndCallIC that pushes the
> arguments onto stack and calls CallICStub.
>
> Also adds two new bytecodes CallIC and CallICWide to indicate calls have to
> go through CallICStub.
>
> MIPS port contributed by balazs.kilvady.
>
> BUG=v8:4280, v8:4680
> LOG=N
>
> Committed: https://crrev.com/20362a2214c11a0f2ea5141b6a79e09458939cec
> Cr-Commit-Position: refs/heads/master@{#34244}
TBR=rmcilroy@chromium.org,mvstanton@chromium.org,mstarzinger@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4280, v8:4680
Review URL: https://codereview.chromium.org/1731253003
Cr-Commit-Position: refs/heads/master@{#34252}
Calls are implemented through CallICStub to collect type feedback. Adds
a new builtin called InterpreterPushArgsAndCallIC that pushes the
arguments onto stack and calls CallICStub.
Also adds two new bytecodes CallIC and CallICWide to indicate calls have to
go through CallICStub.
MIPS port contributed by balazs.kilvady.
BUG=v8:4280, v8:4680
LOG=N
Review URL: https://codereview.chromium.org/1688283003
Cr-Commit-Position: refs/heads/master@{#34244}
This CL introduces two new bytecodes TailCall and TailCallWide.
BUG=v8:4698,v8:4687
LOG=N
Review URL: https://codereview.chromium.org/1698273003
Cr-Commit-Position: refs/heads/master@{#34083}
This removes the CompilationInfo field from the BytecodeGraphBuilder
class. The intention is to reduce the risk of using uninitialized or
unavailable values from the CompilationInfo (e.g. values that are only
available after parsing).
R=rmcilroy@chromium.org
Review URL: https://codereview.chromium.org/1704913002
Cr-Commit-Position: refs/heads/master@{#34081}
Adds support for ES6 super keyword and performing loads, stores, and
calls to super class members.
Implements SetHomeObject and enables ThisFunctionVariable.
BUG=v8:4280,v8:4682
LOG=N
Review URL: https://codereview.chromium.org/1689573004
Cr-Commit-Position: refs/heads/master@{#33977}
Adds JumpIfNotHoleConstant and JumpIfNotHoleConstantWide bytecodes
and removes JumpIfHole bytecode.
In situations with large numbers of constants, the generator would
fail because an 8-bit constant could not be reserved for
JumpIfHole/JumpIfNotHole and so a 16-bit constant would be reserved.
Then when patching the bytecode the patcher would discover there was
no wide constant variant of the emitted jump.
BUG=v8:4280,v8:4680
LOG=N
Review URL: https://codereview.chromium.org/1697473002
Cr-Commit-Position: refs/heads/master@{#33952}
