Rationale:
With limited integers, signed/unsigned 64-bit int typed
data is just a bit pattern, no need to sign/zero extend
these into bigger ints on load. This enables more
intrinsification of indexed stores/loads.
Note:
Still TBD, inline these indexed operations too.
Performance:
About 10x improvement on micro benchmarks.
https://github.com/dart-lang/sdk/issues/33205
Change-Id: I640c324a7d91e57fb4edc025e0dd456ad34fe906
Reviewed-on: https://dart-review.googlesource.com/60403
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
The stub code for calling into the runtime is now responsible for saving any
live registers. Since it doesn't know which registers are live (it's shared
across many safepoints), it saves all the registers. The stackmaps and
environments for safepoints which call this stub are extended to cover the stack
slots in the stub's frame which correspond to registers which are actually live
at the safepoint.
Currently this is only supported in precompiled X64 for the `CheckNull`
slow-path. By creating a separate stub for the CheckNull slow path we are able
to eliminate the size overhead of preparing arguments for the runtime call stub,
and reduce the slow-path size on X64 to 1 instruction.
# Test Plan
The stackmap for a safepoint in the `CheckNull` slowpath is only used when
handling the exception causes a GC, which is extremely unlikely. To induce this,
we add a flag to force this behavior in debug-mode and a test which exercises
this. The rest of the changes inherit existing coverage.
Change-Id: I3d8f2a7c3560c432599344d7420d42ddbee7a89d
Reviewed-on: https://dart-review.googlesource.com/59660
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Samir Jindel <sjindel@google.com>
Rationale:
Rather than forced inlining of clamped convertors
(the "saturated" method _toClampedUint8() from
dart:typed_data), exposing the fact that it always
returns smi values (since they check fail on null inputs)
yields much better Uint8ClampedListView performance
(it avoids re-compilation due to a speculative CheckSmi).
Note:
In the long run, we may still want them inlined
and improve range analysis to deal with clamping.
Performance:
About 2.8x faster than previous optimized version,
about 4.5x faster than original.
https://github.com/dart-lang/sdk/issues/33205
Change-Id: I86a06525d2f2ea0476effd3c3d856ff8d9ab1d87
Reviewed-on: https://dart-review.googlesource.com/60201
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Rationale:
Handles remaining polymorphic reason for typed_data
setters and getters (internal vs. external) during inlining.
Also introduces high level flow graph utilities that
can be reused throughout the compiler to reduce
future code duplication. Disables type speculation
for 64-bit AOT Dart2 to make all work.
Performance:
About 4x speedup on micro benchmarks (AOT64).
https://github.com/dart-lang/sdk/issues/33205
Change-Id: I678426719e49cd8aa1e5051523da12178120b3ba
Reviewed-on: https://dart-review.googlesource.com/59000
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
To reduce size, PC descriptors corresponding to deoptimization points
were omitted if FLAG_optimization_counter_threshold < 0.
However, even if generation of optimized code is disabled with
'--optimization-counter-threshold=-1' option, we could still load optimized
code from snapshot. The loaded code can trigger deoptimization and
PC descriptors corresponding to deoptimization points could be needed.
This change removes this incorrect optimization of PC descriptors size when
optimizing compiler is disabled.
Fixes https://github.com/dart-lang/sdk/issues/33389
Change-Id: I74b0fab34ca1a46672e5a8d5c45ea9b506eab9b1
Reviewed-on: https://dart-review.googlesource.com/59481
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
The `@pragma` annotation on `vmservice_io.main` is modified to only mark it as a
root on non-product builds. Since the annotation is evaluated in the constant
transformation, we need to thread a flag through `gen_kernel.dart` indicating
whether the Kernel produced should target a product precompiler.
# Test Plan
Tested locally on Flutter. Unfortunately it's not possible to test the scenario where
`vmservice_io` is not included, since we don't have any "product" bots. Soon the
flag will be introduced into Flutter, and we will have some coverage from there.
Change-Id: I39f12343038fa221aafe5a55241bff4b8d14ec19
Reviewed-on: https://dart-review.googlesource.com/57501
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
So far the frontend (parser, flow graph builder, ssa construction) were
aware of the actual frame layout.
This CL makes the indices we assign to [LocalVariable]s logical
indices, assigning:
* M parameters the indices 1 ... M
* N local variables the indices 0 -1 ... -(N-1)
The scope building, flow graph builder and ssa construction operate on
those logical indices.
When emitting actual code, the backend will translate those indices into
actual FP relative indices. This allows us to be more flexible in the
backend which frame layout we choose.
Issue https://github.com/dart-lang/sdk/issues/33274
Change-Id: I9a504bf97821c257aafd2b3430df9f4c9da4b442
Reviewed-on: https://dart-review.googlesource.com/57321
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Rationale:
Always inline int convertors that don't do
much more than testing and anding.
For example, force inline
v26 <- StaticCall:66( _toUint8@6027147<0> v4 T{Type: class 'int'?}) T{Type: class 'int'?}
to
v47 <- Constant(#255)
..
CheckSmi:10(v4 T{Type: class 'int'?})
v45 <- BinarySmiOp:10(&, v4 T{_Smi}, v47 T{_Smi}) T{_Smi}
https://github.com/dart-lang/sdk/issues/33205
Change-Id: I595d9a64365e16ae244480b5e27f8be23c43d164
Reviewed-on: https://dart-review.googlesource.com/58061
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Aart Bik <ajcbik@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
- Remove a level of indirection when accessing code units.
- Require a finalizer when constructing an external string.
Change-Id: I3f65246bf0ac50ffad900e2c338623a7684a9d3d
Reviewed-on: https://dart-review.googlesource.com/54300
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Relanding 4be50d6fa1 with fixes to DBC
and location summaries: AssertAssignable must save FPU registers.
For now we are limiting this to type checks against type parameter types.
In Dart 1 mode Dart2JS compiles itself in 28s when running from source
and in 23s when running from ideal app-jit snapshot (trained on the
same workload).
Before this change in Dart 2 mode numbers were 51s and 57s respectively.
After this change in Dart 2 mode numbers are 38s and 32s. Meaning
that regression is reduced by 50%.
Issue https://github.com/dart-lang/sdk/issues/31798
Issue https://github.com/dart-lang/sdk/issues/33257
Change-Id: Ifb55f86453bfdf36a2e03bcd7f3197cfde257103
Reviewed-on: https://dart-review.googlesource.com/57980
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
This reverts commit 4be50d6fa1.
Reason for revert: Failures on SIMDBC64 and Analyzer bots.
Original change's description:
> [vm] Enable type stubs based type checks in JIT mode for some types.
>
> For now we are limiting this to type checks against type parameter types.
>
> # Performance improvements
>
> In Dart 1 mode Dart2JS compiles itself in 28s when running from source
> and in 23s when running from ideal app-jit snapshot (trained on the
> same workload).
>
> Before this change in Dart 2 mode numbers were 51s and 57s respectively.
>
> After this change in Dart 2 mode numbers are 38s and 32s. Meaning
> that regression is reduced by 50%.
>
> Issue https://github.com/dart-lang/sdk/issues/31798
> Issue https://github.com/dart-lang/sdk/issues/33257
>
> Change-Id: I34bf5385a5cc3c7702dc281c6dfa89da85d3dde1
> Reviewed-on: https://dart-review.googlesource.com/57601
> Reviewed-by: Régis Crelier <regis@google.com>
> Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
TBR=vegorov@google.com,kustermann@google.com,regis@google.com
Change-Id: I85a30c962b0cd556310e19193f5993ab76ecf2e7
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/57840
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
For now we are limiting this to type checks against type parameter types.
# Performance improvements
In Dart 1 mode Dart2JS compiles itself in 28s when running from source
and in 23s when running from ideal app-jit snapshot (trained on the
same workload).
Before this change in Dart 2 mode numbers were 51s and 57s respectively.
After this change in Dart 2 mode numbers are 38s and 32s. Meaning
that regression is reduced by 50%.
Issue https://github.com/dart-lang/sdk/issues/31798
Issue https://github.com/dart-lang/sdk/issues/33257
Change-Id: I34bf5385a5cc3c7702dc281c6dfa89da85d3dde1
Reviewed-on: https://dart-review.googlesource.com/57601
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Rationale:
Some native typed setters/getters in the typed_data library
are only called from clients that perform explicit bounds checks
first. Under strongly typed Dart2, there is no need to repeat
these tests. Avoids overhead, and reduces the polymorphicness
of these calls (preparing more inlining later).
https://github.com/dart-lang/sdk/issues/33205
Tests:
./tools/test.py --mode release --arch x64 -r vm
tools/test.py -m release -r vm -c dartk --strong standalone_2/typed_data_test
Change-Id: I81771a4f4b41a21e344f2d50745bbc30480b2a6c
Reviewed-on: https://dart-review.googlesource.com/57460
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Fixes#33170: We had an assert that type variables are not bounded by "dynamic".
This assert was previously valid because the FE replaces "extends dynamic" with
"extends Object". However, it's not necessary in any way, and it's difficult to
access the Object type from outside the FE when synthesizing type parameters for
expression compilation.
Fixes#33171: The synthetic class generated during serialization to hold the
synthetic method may need to extend core library classes if the expression's
scope was within a core library class. In this case, the class finalizer needs
to avoid applying the restrictions on class extension to the synthetic class.
To facilitate this, we now use a single class name for all synthetic classes,
rather than copying the original class name. The name of the synthetic class has
no function at all, since the synthetic method is reparented to the original
class before execution.
Fixes#33239: The arguments descriptor passed to InvokeFunction for expression
functions which capture generic type parameters was using the wrong argument
count.
Fixes#33270: Use LookupClassAllowPrivate instead of LookupClass when resolving
the expression's scope in the VM.
Test Plan:
#33170: python tools/test.py -m debug -c dartk --strong service/evaluate_function_type_parameters_test
#33171: python tools/test.py -m debug -c dartk --strong service/eval_test
#33239: Updated "evaluate_function_type_parameters_test.dart".
#33270: Updated "eval_test.dart".
Change-Id: I376926bddd2224ec2322b43685d6c13831f1a8e7
Reviewed-on: https://dart-review.googlesource.com/56060
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
When the constant transformation is enabled on annotations, we need to fix handling
of @ExternalName annotations in the kernel_loader to ensure that we are setting
is_external = false on native methods.
The original revision is in patchset 1.
# Test Plan
The only regression was on benchmarks, Golem results are pending.
Change-Id: Ib80bb9f532299056e770a3b378cc5ad9ee451f57
Reviewed-on: https://dart-review.googlesource.com/56960
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
GrowableArrayMarker was a class that implemented int and was used to
enable implementation of default List factory constructor in pure Dart:
factory List([int length = GROWABLE_ARRAY_MARKER]) {
return identical(length, GROWABLE_ARRAY_MARKER) ? new _GrowableList<E>(0)
: new _List<E>(length);
}
Its existence complicated all kinds of things in the VM and it is finally
time to remove it.
Instead we build List factory body directly in IL.
This CL also provides inlining rule for `new List(n)` case.
Change-Id: I870751658a4ac17fce649c9ac70395ff88a5436c
Reviewed-on: https://dart-review.googlesource.com/57262
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Kernel mixin transformation desugars mixin applications into normal
classes. Mixed-in type is pulled into interfaces list.
However, dart:mirrors needs to know the original mixed-in type of
a mixin application.
This change solves this problem by propagating a 'isTransformedMixinApplication'
attribute of a class through kernel AST, kernel binary and VM objects
into dart:mirrors implementation.
Fixes: https://github.com/dart-lang/sdk/issues/33240
Change-Id: I98ca69294e1ad445402a5ca91d90c30447aabcb2
Reviewed-on: https://dart-review.googlesource.com/56721
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Kevin Millikin <kmillikin@google.com>
This reverts commit 3e50ea32b5.
Reason for revert: Severe performance regressions on many aot-v2 benchmarks.
Original change's description:
> [vm] Support definition of entry-points via @pragma('vm.extern') annotations.
>
> The `@pragma` annotations are evaluated by the constants transformation and
> visible to TFA and the precompiler, which match on the "options" field of the
> annotation to determine whether to mark the class/procedure as a root.
>
> This required enabling the transformation of annotation constants by default.
>
> # Test Plan
>
> The "vmservice_io.main" entry-point is removed from `main.cc` and annotated with
> `@pragma`. All precompiler tests will crash if "vmservice_io.main" is not
> available at runtime.
>
> Debug/release precompiler bots are visible in "cl-linux" button.
>
> Change-Id: I03c5d6ba7918672ed9905fcaee8dabe675a93a5d
> Reviewed-on: https://dart-review.googlesource.com/56660
> Commit-Queue: Samir Jindel <sjindel@google.com>
> Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
TBR=vegorov@google.com,alexmarkov@google.com,sjindel@google.com
Change-Id: I779c17d003659129a4b3fcf284423104948f60e2
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/56820
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Samir Jindel <sjindel@google.com>
The `@pragma` annotations are evaluated by the constants transformation and
visible to TFA and the precompiler, which match on the "options" field of the
annotation to determine whether to mark the class/procedure as a root.
This required enabling the transformation of annotation constants by default.
# Test Plan
The "vmservice_io.main" entry-point is removed from `main.cc` and annotated with
`@pragma`. All precompiler tests will crash if "vmservice_io.main" is not
available at runtime.
Debug/release precompiler bots are visible in "cl-linux" button.
Change-Id: I03c5d6ba7918672ed9905fcaee8dabe675a93a5d
Reviewed-on: https://dart-review.googlesource.com/56660
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Summary:
When evaluating constants for direct inspection (and not for injection into the
FG), we store many intermediate values in ZoneHandles. This leaks memory, since
the only need for ZoneHandles is preserving constants which are referenced by
the FG.
Test Plan:
Existing coverage is sufficient since no functionality is changed.
Change-Id: I33951eab762a4ad626f07ebf11118247c5a9ad3e
Reviewed-on: https://dart-review.googlesource.com/56345
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
# Summary
As part of commit 772c9bb5f, we changed the representation of partially
instantiated closures to facilitate partial instantiation of local functions.
The compile-time constant evaluator was not updated -- this revision implements
that update.
Fixes task 33211.
# Test Plan
Updated `partial_tearoff_instantiation_test.dart` to test the case which caused
incorrect behavior in task 33211.
Change-Id: I7bbd4fb83a5aea86ffc85c1d9952f1202f098d6b
Reviewed-on: https://dart-review.googlesource.com/56346
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Samir Jindel <sjindel@google.com>
Given IL like this:
t3 <- Constant(#null) <--\
// ... do something ... |
StoreLocal(:temp, t4) ---/
// Do something else
StoreIndexed(t1, t2, t3)
Graph builder would use Constant(#null) as an input definition for
StoreIndexed - which leads to wrong generated code in unoptimized mode
on X64 where StoreIndexed has specialization for constant values. In
optimized code SSA renaming would actually replace t3 with a correct
reaching definition.
This CL introduces a new IL instruction MakeTemp(...) which has exactly
the same effect as Constant(#null) on the expression stack, but is
not treated as a constant value by the backend.
The bug was unnoticed because of the loose assertions in the SSA
renaming which permitted replacing Constant-s with non-Constant
reaching definitions. This CL also tightens those assertions to
catch this issue.
Also cleanup the code in SSA renaming a bit to make it a bit easier
to understand.
Fixes https://github.com/dart-lang/sdk/issues/33195
Bug: 33195
Change-Id: I7c914c836d4af7a50e8a8e1a02d03e9413f87779
Reviewed-on: https://dart-review.googlesource.com/56112
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
With this change debugger requests yield(await) positions from kernel builder so it can confirm which frame handles what exceptions.
This also renames ':completer' to ':async_completer' in constructed kernel, so it is compliant with what VM expects to see(and updates test expectations accordingly).
Further it fixes async stack navigation for sync-async that became default in dart vm since first version went for the review((uses `future` getter, rather than property).
It also makes `future` getter non-debuggable to allow stepping-out of async methods.
Make Handle zone-scoped. Clean up frame counter var name.
Bug: https://github.com/dart-lang/sdk/issues/29056
Change-Id: Ia71f3c851a6b313655b57dad28c296f5dd081eda
Reviewed-on: https://dart-review.googlesource.com/54640
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Summary:
The constants table contains a concatenation of the variable-length encodings of
each constant used in the serialized component. Previously, we referenced the
constants in the code by index into the constants table, which requires reading
potentially the entire constants table to dereference. Now, we refer to the
constant by the binary offset into the constants table, which allows us to peek
into it in constant time.
Test Plan:
The constants table is used by all the precompiler modes, so the new encoding
will inherit the coverage of the existing encoding.
Results from debug and release precompiler bots are available under the
"cl-linux-try" button.
Change-Id: I4782bc0035d0b0cbd5bd6b9d086561ea601286c8
Reviewed-on: https://dart-review.googlesource.com/55470
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Alexander Markov