Transform LdaNull/LdaUndefined followed by StrictEquality to TestNull/TestUndefined.
This would avoid a call to the compare IC. In the bytecode-graph builder these are
mapped to StrictEqual javascript operator. When reducing this operator, we already
optimize the cases for null/undefined.
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2554723004
Cr-Commit-Position: refs/heads/master@{#41768}
This introduces an explicit struct for the communication channel between
the {ArrayLiteral} AST node and the corresponding runtime methods. Those
methods take a pair of {ElementsKind} as well as an array (can either be
a FixedArray or a FixedDoubleArray) of constant values.
For bonus points it also reduces the size of the involved heap object by
one word (i.e. length field of FixedArray not needed anymore).
R=mvstanton@chromium.org
Review-Url: https://codereview.chromium.org/2581683003
Cr-Commit-Position: refs/heads/master@{#41752}
Allocate the registers used as arguments to a call on-demand after visiting the
argument (or reciever). This means that the visited expression can use registers
that would otherwise have been allocated for arguments which haven't been
visited yet.
The reason for doing this is to avoid keeping things live in registers
unecessarily for chained function calls, which avoids a memory leak for
functions which chain a large number of calls with large temporary arguments /
recievers.
BUG=chromium:672027
Review-Url: https://codereview.chromium.org/2557173004
Cr-Commit-Position: refs/heads/master@{#41714}
The evaluation order of this argument was accidentally changed when the
special-case was added for super calls with a final spread argument.
Review-Url: https://codereview.chromium.org/2563423002
Cr-Commit-Position: refs/heads/master@{#41633}
The patch was reverted due to a bug - we failed to evict OSR-optimized
code in the case where the SharedFunctionInfo OptimizedCodeMap was
empty/cleared.
Since we OSR code rarely, it makes sense to store it and look for it on the native context rather than the SharedFunctionInfo. This makes the OptimizedCodeMap data structure more space efficient, as it doesn't have to store an ast ID for the OSR entry point.
Review-Url: https://codereview.chromium.org/2561083002
Cr-Commit-Position: refs/heads/master@{#41584}
-- Moves promiseHasHandlerSymbol to inobject property
-- Ports PromiseResolveClosure to TF
-- Fix a non spec async-await test which fails now because we do a map
check for native promise check (instead of IsPromise). Changing the
constructor (in the test) invalidates the map check.
This patch results in a 7.1% performance improvement in the bluebird
benchmark (over 5 runs).
BUG=v8:5343
Review-Url: https://codereview.chromium.org/2541283002
Cr-Commit-Position: refs/heads/master@{#41569}
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}
Since we OSR code rarely, it makes sense to store it and look for it on the native context rather than the SharedFunctionInfo. This makes the OptimizedCodeMap data structure more space efficient, as it doesn't have to store an ast ID for the OSR entry point.
BUG=
Review-Url: https://codereview.chromium.org/2549753002
Cr-Commit-Position: refs/heads/master@{#41554}
This allows to detect a static property also named 'name', and also makes sure 'name' is added last, to be standards-compliant.
BUG=v8:4199
Review-Url: https://codereview.chromium.org/2423053002
Cr-Commit-Position: refs/heads/master@{#41546}
This just calls into a runtime function for implementation currently.
Intermediate step in speeding up constructor calls containing a spread.
The NewWithSpread bytecode will probably end up having different arguments with future CLs - the constructor and the new.target should have their own regs. For now we are calling into the runtime function, so we need the regs together.
BUG=v8:5659
Review-Url: https://codereview.chromium.org/2541113004
Cr-Commit-Position: refs/heads/master@{#41542}
Equality with null/undefined is equivalent to a check on the undetectable bit
on the map of the object. This would be more efficient than performing the entire
comparison operation.
This cl introduces:
1. A new bytecode called TestUndetectable that checks if the object is null/undefined.
2. Updates peeophole optimizer to emit TestUndetectable when a LdaNull/Undefined
precedes equality check.
4. TestUndetectable is transformed to ObjectIsUndetectable operator when building
turbofan graph.
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2547043002
Cr-Commit-Position: refs/heads/master@{#41514}
Unfortunately we have to split this up into two cases: those with exactly one spread argument as the final argument, and all others, due to any side-effects of evaluation being visible.
This is in preparation for a new bytecode which handles super calls.
BUG=v8:5659
Review-Url: https://codereview.chromium.org/2540593003
Cr-Commit-Position: refs/heads/master@{#41415}
Removed a redundant check: If completion is not normal, then #iterator cannot be undefined.
Review-Url: https://codereview.chromium.org/2533803002
Cr-Commit-Position: refs/heads/master@{#41321}
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}
Collect string feedback for compare operations. Without this,
functions which have a lot of string compare operations end up with
a high generic type percentage, and don't get optimized until very
late.
Currently TurboFan doesn't use this String feedback for compare
operations, but this could be done in future work if it is useful.
BUG=chromium:660947
Review-Url: https://codereview.chromium.org/2506013005
Cr-Commit-Position: refs/heads/master@{#41078}
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}
Adds a bytecode to set and retrieve the pending message. This avoids a
runtime call in finally blocks, and also ensures that TurboFan builds a
graph using the SetMessage / LoadMessage nodes instead of inserting a
runtime call.
BUG=chromium:662334
Review-Url: https://codereview.chromium.org/2501503005
Cr-Commit-Position: refs/heads/master@{#41023}
Avoid using the iterator for arrays with fast elements where the iterator has
not been modified.
Only deals with the case where there is a single spread argument.
Improves the six-speed "spread" benchmark to 1.5x slower than baseline es5 implementation, compared to 19x slower previously.
BUG=v8:5511
Review-Url: https://codereview.chromium.org/2465253011
Cr-Commit-Position: refs/heads/master@{#40998}
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}
This removes the POSSIBLY_EVAL_CALL call type, and instead uses OTHER_CALL
or WITH_CALL to decide whether to do the special LOOKUP_SLOT_CALL runtime
call to find the callee and possibly update the receiver with the with-object.
This means that eval calls out of 'with' blocks can now just do a normal
LdaLookupGlobalSlot operation, which can check the context chain for eval
extentions and fast-path the lookup if none exist.
BUG=661556
Review-Url: https://codereview.chromium.org/2487483004
Cr-Commit-Position: refs/heads/master@{#40965}
This replaces LOOKUP_SLOT_CALL with WITH_CALL, and relies on regular lookup-slot handling in variable load to support other lookup slots (variables resolved in the context of sloppy eval). This allows optimizations for such variable loads to kick in for calls as well. We only need special handling for function calls in the context of with, since it changes the receiver of the call from undefined/global to the with-object.
This currently doesn't yet make it work for the direct eval call itself, since the POSSIBLY_EVAL_CALL flag is also used to deal with direct eval later.
BUG=
Review-Url: https://codereview.chromium.org/2480253006
Cr-Commit-Position: refs/heads/master@{#40962}
- A new runtime function (%create_resolving_functions) is installed to
call the CreateResolvingFunctions builtin from JS.
- Three new builtins are created - resolve and reject functions and a
third function that creates a new JSFunctions from these
resolve/reject builtins.
- The promise reject function is installed on the context temporarily
as internal_promise_reject. This should go away once we remove
PromiseSet.
BUG=v8:5343
Review-Url: https://codereview.chromium.org/2459283004
Cr-Commit-Position: refs/heads/master@{#40903}
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 introduces two new bytecodes LdaModuleVariable and StaModuleVariable,
replacing the corresponding runtime calls.
Support in the bytecode graph builder exists only in the form of runtime calls.
BUG=v8:1569
Review-Url: https://codereview.chromium.org/2471033004
Cr-Commit-Position: refs/heads/master@{#40825}
This makes use of the newly introduced cell indices to speed up variable
accesses. Imports and local exports are now directly stored in (separate)
arrays. In the future, we may merge the two arrays into a single one, or
even into the module context.
This CL also replaces the LoadImport and LoadExport runtime functions with
a single LoadVariable taking a variable index as argument (rather than a
name).
BUG=v8:1569
Review-Url: https://codereview.chromium.org/2465283004
Cr-Commit-Position: refs/heads/master@{#40808}
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}
Move hole check logic from full-codegen into scope analysis, and store the
"needs hole check" bit on VariableProxy. This makes it easy to re-use in
any backend: it will be trivial to extend the use of this logic in, e.g.,
full-codegen variable stores.
While changing the signatures of the variable loading/storing methods in
Ignition, I took the liberty of replacing the verb "Visit" with "Build", since these
are not part of AST visiting.
BUG=v8:5460
Review-Url: https://chromiumcodereview.appspot.com/2411873004
Cr-Commit-Position: refs/heads/master@{#40479}
Consistently collect CallIC feedback in fullcodegen and Ignition, even
for possibly direct eval calls, that were treated specially so far, for
no apparent reason. With the upcoming SharedFunctionInfo based CallIC
feedback, we might be able to even inline certain direct eval calls, if
they manage to hit the eval cache. More importantly, this patch
simplifies the collection and dealing with CallIC feedback (and as a
side effect fixes an inconsistency with feedback for super constructor
calls).
R=mvstanton@chromium.org, mythria@chromium.org
BUG=v8:2206,v8:4280,v8:5267
Review-Url: https://codereview.chromium.org/2426693002
Cr-Commit-Position: refs/heads/master@{#40397}
This adds more useful information to the v8-heap-stats tool.
BUG=v8:5489
Review-Url: https://codereview.chromium.org/2394213003
Cr-Commit-Position: refs/heads/master@{#40361}
This makes optimization of all class constructors (i.e. both base and
derived) go through TurboFan. Note some class constructors containing
Harmony language features (e.g. super constructor calls or accesses to
the new.target value) were already going through TurboFan before.
R=bmeurer@chromium.org
BUG=v8:5458
Review-Url: https://codereview.chromium.org/2397723002
Cr-Commit-Position: refs/heads/master@{#40342}
In order to optimize super constructor calls with ES6 classes, we need
some feedback for both the JSCallConstruct and the resulting JSCreate
nodes in TurboFan. Both already optimize perfectly when the see nodes
with JSFunction constants, so utilizing the existing CallIC machinery
here, enables us to optimize the super constructor calls right now w/o
a lot of effort.
Note that there are probably better ways to track this information,
for example we could do some tracking on the constructor functions;
this will however require serious changes in TurboFan and the runtime,
and would block progress on more important tasks.
R=mythria@chromium.org
BUG=v8:5517
Review-Url: https://codereview.chromium.org/2419423002
Cr-Commit-Position: refs/heads/master@{#40337}
Now that all accesses to the last match info are in C++ and TF code, we can
finally turn the last match info into a FixedArray. Similar to the ArrayList,
it uses its first field to store its length and grows dynamically in amortized
O(1) time.
Unlike previously, this means that the last match info pointer stored on the
context can actually change (in case the FixedArray needs to grow).
BUG=v8:5339
Review-Url: https://codereview.chromium.org/2415103002
Cr-Commit-Position: refs/heads/master@{#40308}
This ports RegExpInitialize, IsRegExp, InternalMatch and InternalReplace to C++
/ TurboFan. InternalMatch is in TurboFan because it calls RegExpExecStub and
needs to construct a RegExpResult (which are, respectively, a PlatformStub and
a CodeStubAssembler function).
Except for LastMatchInfo (and GetSubstitution, which could be moved to string.js
anytime), regexp.js is now completely empty.
BUG=v8:5339
Review-Url: https://codereview.chromium.org/2409513003
Cr-Commit-Position: refs/heads/master@{#40277}
This moves the implementation of @@replace from regexp.js to builtins-regexp.cc
(the TurboFan fast path) and runtime-regexp.cc (slow path). The fast path
handles all cases in which the regexp itself is an unmodified JSRegExp
instance, the given 'replace' argument is not callable and does not contain any
'$' characters (i.e. we are doing a string replacement).
BUG=v8:5339
Review-Url: https://codereview.chromium.org/2398423002
Cr-Commit-Position: refs/heads/master@{#40253}
