The results are too noisy ATM. This switches off validate-asm for
default comparisons. We can add back dedicated jobs later that switch
it on.
BUG=chromium:663714
NOTRY=true
TBR=bradnelson@chromium.org,titzer@chromium.org
Review-Url: https://codereview.chromium.org/2640743004
Cr-Commit-Position: refs/heads/master@{#42467}
The new ignition config will be used as baseline comparison in new
jobs, e.g. against ignition_turbo. We'll keep --validate-asm off
in ignition_turbo for now as it is very chatty.
BUG=chromium:673246
NOTRY=true
Review-Url: https://codereview.chromium.org/2640043002
Cr-Commit-Position: refs/heads/master@{#42464}
- Refactors many FP, integer, and signed integer instructions where
possible.
LOG=N
BUG=v8:4124
Review-Url: https://codereview.chromium.org/2639443002
Cr-Commit-Position: refs/heads/master@{#42463}
CodeAssemblerLabel's destructor was not calling RawMachineLabel's destructor, because label_ is zone-allocated. RawMachineLabel's destructor contains a DCHECK that would make debugging easier. If the DCHECK is not triggered, things will go awry in the register allocation phase, making it harder to debug the issue.
BUG=
Review-Url: https://codereview.chromium.org/2641863002
Cr-Commit-Position: refs/heads/master@{#42461}
A future linear version of the hash table will only need the element count and
deleted element count. Hence moving them to the beginning of the underlying
fixed array makes the transition easier.
BUG=v8:5717
Review-Url: https://codereview.chromium.org/2630373002
Cr-Commit-Position: refs/heads/master@{#42459}
Due to outstanding performance issues, it's safer to revert these
changes.
Revert "PPC/s390: [TypeFeedbackVector] DeclareGlobals needs a literals array"
Revert "[TypeFeedbackVector] DeclareGlobals needs a literals array"
Revert "PPC/s390: [builtins] CompileLazy bailed out to the runtime entirely too often."
Revert "[builtins] CompileLazy bailed out to the runtime entirely too often."
Another CL is to follow that removes the main part of this work
(after that revert, performance regressions should disappear.
TBR=mstarzinger@chromium.org, leszeks@chromium.org, bmeurer@chromium.org
BUG=680637
Review-Url: https://codereview.chromium.org/2645543002
Cr-Commit-Position: refs/heads/master@{#42456}
Also, emit a NewWithSpread bytecode for CallNew AST nodes where possible, rather than desugaring in the parser.
BUG=v8:5511
Review-Url: https://codereview.chromium.org/2629363002
Cr-Commit-Position: refs/heads/master@{#42455}
This CL fixes the check for empty elements in keys.cc. Previously we would
accidentally bail out of the fast path because the check would always fail.
As a consequence for-in loops that would initialize the enum-cache of an object
with own-only fast properties would never be optimized properly.
Review-Url: https://codereview.chromium.org/2638323002
Cr-Commit-Position: refs/heads/master@{#42454}
Lower JSLoadMessage and JSStoreMessage in JSTypedLowering to allow for
optimizations across try-catch, try-finally, for-of, destructuring,
etc.
BUG=v8:5448
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2641843002
Cr-Commit-Position: refs/heads/master@{#42453}
Breakpoints are always re-set by the embedder after compilation, so we
don't want to store the corresponding breakpoint objects.
Also don't serialize WASM_INTERPRETER_ENTRY stubs as they are replaced
by ordinary WASM_FUNCTION code at instantiation anyway, and skip
WASM_TO_JS wrappers which are recompiled on each instantiation.
Instead, we serialize the Illegal builtin, and also use that one
instead of the placeholder when compiling the wasm code initially.
R=titzer@chromium.org, yangguo@chromium.org
BUG=v8:5822
Review-Url: https://codereview.chromium.org/2629853004
Cr-Commit-Position: refs/heads/master@{#42451}
As no one will ever try to allocate a Thread directly, we can just make
Thread a proxy of ThreadImpl by reinterpret_casting between both types.
This allows to not mention ThreadImpl in the header at all, and to
define it in an anonymous namespace in the implementation, allowing for
more optimizations.
It also saves runtime, as no memory load is needed to forward from
Thread to ThreadImpl, and we do not need to allocate ThreadImpl objects
on the heap.
R=titzer@chromium.org
Review-Url: https://codereview.chromium.org/2643443002
Cr-Commit-Position: refs/heads/master@{#42450}
When inlining allocations for Array and String iterators, make sure the
allocations have proper types on it (rather than just Type::Any). This
allows us to escape analyze the iterators in a couple of more cases,
i.e. in this very simple case:
function f(a) {
const it = a[Symbol.iterator]();
return it.next().value;
}
R=jarin@chromium.org
BUG=v8:4586,v8:5448
Review-Url: https://codereview.chromium.org/2640743003
Cr-Commit-Position: refs/heads/master@{#42449}
This also introduces exhaustive switch-cases for instance types.
BUG=chromium:681383
Review-Url: https://codereview.chromium.org/2646433002
Cr-Commit-Position: refs/heads/master@{#42447}
This fast path check compared a tagged smi and an untagged one, meaning that the slow path was always taken.
Review-Url: https://codereview.chromium.org/2641823002
Cr-Commit-Position: refs/heads/master@{#42445}
Store breakpoint positions in the WasmSharedModuleData in order to set
them on new instantiations. Also redirect them to all live instances at
the time the breakpoint is set.
Inside the WasmDebugInfo, we store the BreakPointInfo objects to find
hit breakpoints.
R=titzer@chromium.org, yangguo@chromium.org
BUG=v8:5822
Review-Url: https://codereview.chromium.org/2626253002
Cr-Commit-Position: refs/heads/master@{#42443}
We used dynamic dispatch before, which has considerable runtime
overhead. Pimpl (a.k.a. opaque pointer) has the same memory overhead
of one additional pointer, but less runtime overhead.
R=titzer@chromium.org
Review-Url: https://codereview.chromium.org/2639013002
Cr-Commit-Position: refs/heads/master@{#42442}
Non-capturing groups so far did not have any effect, so we elided
them in the regexp parser. However, they do affect how lookarounds
are parsed: in unicode mode, lookarounds are not quantifiable. They
are however if wrapped in a group.
BUG=v8:5845
Review-Url: https://codereview.chromium.org/2636883002
Cr-Commit-Position: refs/heads/master@{#42436}
Collect Receiver feedback for abstract/strict equality in Ignition and
use it in TurboFan to optimize JSEqual and JSStrictEqual operations to
pointer equality instead of having to call Equal/StrictEqual builtins.
R=jarin@chromium.org
BUG=v8:5267,v8:5400
Review-Url: https://codereview.chromium.org/2639883002
Cr-Commit-Position: refs/heads/master@{#42435}
- Currently WebAssembly.Memory.grow() assumes that it always has an instance associated with it,
fix to grow and reflect new size when no instance is associated with memory object.
- Correctness fixes for the js api, throw range errors instead of generic errors
BUG=chromium:680938
R=bradnelson@chromium.org, titzer@chromium.org
Review-Url: https://codereview.chromium.org/2638243002
Cr-Commit-Position: refs/heads/master@{#42432}
This rewrites the rest property into a runtime call which sets up the
correct properties in the newly created object.
- Changes flag to --harmony-object-rest-spread
- Changes pattern rewriter to desugar rest property
- Adds new runtime function CopyDataPropertiesWithExcludedProperties
BUG=v8:5549
Review-Url: https://codereview.chromium.org/2620943002
Cr-Commit-Position: refs/heads/master@{#42430}
The existing implementation assumes that return nodes have exactly one
real value input. This assumption does not hold for WebAssembly. To
avoid incorrect behavior, this CL turns of the reduction of returns
with a value input count != 1.
R=titzer@chromium.org, mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2638053002
Cr-Commit-Position: refs/heads/master@{#42425}
There were multiple bugs in the apply builtins on mips and mips64 which
caused them to always take the slow path by calling into the runtime to
create the array. The other bugs went undiscovered because those paths
were never taken.
Review-Url: https://codereview.chromium.org/2634393002
Cr-Commit-Position: refs/heads/master@{#42424}
Port b83c3d29a9
Original Commit Message:
When a function is declared in global scope, the closure is created
by the DeclareGlobals runtime service. It needs a pointer to the
literals array, already allocated in the feedback vector. This
fixes a bug where it's behavior wasn't in sync with CreateClosure,
which accepts the literals from the vector.
This enables a follow-on performance improvement in the CompileLazy
builtin.
R=mvstanton@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=680637
LOG=N
Review-Url: https://codereview.chromium.org/2634423002
Cr-Commit-Position: refs/heads/master@{#42421}
Uses the structure of an IEEE float to speed up
DoubleToUint32IfEqualToSelf, similar to FastD2UI. Micro-benchmarks show
a ~1.2x-2x speed-up, depending on the input.
Review-Url: https://codereview.chromium.org/2636453003
Cr-Commit-Position: refs/heads/master@{#42420}
