This CL adds special case for FutureOr<T> type (which is a union
of Future<T> and T) into type flow analysis.
As type flow analysis doesn't retain information about generic type
arguments yet, FutureOr<T> is treated as AnyType for now.
Issue: https://github.com/dart-lang/sdk/issues/30480
Change-Id: Ieb4d862b5a993fe46c9662d9bcb0e1f06058a166
Reviewed-on: https://dart-review.googlesource.com/40463
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Implemented more efficient way to iterate ClassSet in class_hierarchy.dart,
ClassSet becomes Iterable<Class>. This considerably improves speed of
ClosedWorldClassHierarchy.getSingleTargetForInterfaceInvocation() and
CHA-based devirtualization.
Time of CHA-based devirtualization on Flutter Gallery reduced
from ~13.4s to ~1.2s.
Also, ClassHierarchy used in devirtualization is configured to ignore
ambiguous supertypes to avoid crashes on Flutter Gallery.
Change-Id: Iaca251ca268aad3d8bb652247650096f19c25839
Reviewed-on: https://dart-review.googlesource.com/38960
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This relands 75a9579ea0.
The approach works as follows:
Step 1: Kernel transform. Under the closed world assumption compute the
set of selectors dispatched dynamically, then mark all procedures that don't
match any of those selectors as 'not-dispatched-dynamically'.
Step 2: VM backend. When building IR for a function if this function was
marked as not-dispatched-dynamically then omit type checks for any parameter
that is not marked as generic-covariant-impl, as such arguments are guaranteed
to be checked on the caller side (by front-end).
+------------------------+------------+----------+-----------+--------------+
| benchmark | baseline | current | with opt | improved by |
+------------------------+------------+----------+-----------+--------------+
| stock_layout_iteration | 2366.3786 | 2724.3 | 2562.75 | -5.93% |
| stock_build_iteration | 3824.3 | 4914.8 | 4681 | -4.76% |
+------------------------+------------+----------+-----------+--------------+
* Flutter gallery Instructions size is reduced by 11% (8748720 bytes to 7846368 bytes).
Baseline is at 6196496 bytes.
Alternatively to annotating individual procedures, I considered annotating Program node
with a set of dynamically dispatched selectors. Decoding and passing this information
around proved to be quite cumbersome in the "streaming" world, so I opted for a simpler
approach where all individual procedures are annotated.
Bug: https://github.com/dart-lang/sdk/issues/3179
Change-Id: I2f32a609e3872c74d5ae7bbd97555453aaedf15f
Reviewed-on: https://dart-review.googlesource.com/38125
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
This reverts commit 75a9579ea0.
Reason for revert: function_subtype_bound_closure7_test fails in darkp
Original change's description:
> [vm/kernel/aot] Skip unnecessary type checks on parameters of instance methods.
>
> The approach works as follows:
>
> Step 1: Kernel transform. Under the closed world assumption compute the
> set of selectors dispatched dynamically, then mark all procedures that don't
> match any of those selectors as 'not-dispatched-dynamically'.
>
> Step 2: VM backend. When building IR for a function if this function was
> marked as not-dispatched-dynamically then omit type checks for any parameter
> that is not marked as generic-covariant-impl, as such arguments are guaranteed
> to be checked on the caller side (by front-end).
>
> # Performance Impact
>
> +------------------------+------------+----------+-----------+--------------+
> | benchmark | baseline | current | with opt | improved by |
> +------------------------+------------+----------+-----------+--------------+
> | stock_layout_iteration | 2366.3786 | 2724.3 | 2562.75 | -5.93% |
> | stock_build_iteration | 3824.3 | 4914.8 | 4681 | -4.76% |
> +------------------------+------------+----------+-----------+--------------+
>
> * Flutter gallery Instructions size is reduced by 11% (8748720 bytes to 7846368 bytes).
> Baseline is at 6196496 bytes.
>
> # Alternative Implementations
>
> Alternatively to annotating individual procedures, I considered annotating Program node
> with a set of dynamically dispatched selectors. Decoding and passing this information
> around proved to be quite cumbersome in the "streaming" world, so I opted for a simpler
> approach where all individual procedures are annotated.
>
> Change-Id: I363db6d5dd1138fe25917646313c16d0b213c3b4
> Reviewed-on: https://dart-review.googlesource.com/37822
> Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Reviewed-by: Samir Jindel <sjindel@google.com>
TBR=vegorov@google.com,alexmarkov@google.com,sjindel@google.com
Change-Id: I0fd6bf3e6edfc3e605da2b996f9a0da6c409d01c
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/37802
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
The approach works as follows:
Step 1: Kernel transform. Under the closed world assumption compute the
set of selectors dispatched dynamically, then mark all procedures that don't
match any of those selectors as 'not-dispatched-dynamically'.
Step 2: VM backend. When building IR for a function if this function was
marked as not-dispatched-dynamically then omit type checks for any parameter
that is not marked as generic-covariant-impl, as such arguments are guaranteed
to be checked on the caller side (by front-end).
# Performance Impact
+------------------------+------------+----------+-----------+--------------+
| benchmark | baseline | current | with opt | improved by |
+------------------------+------------+----------+-----------+--------------+
| stock_layout_iteration | 2366.3786 | 2724.3 | 2562.75 | -5.93% |
| stock_build_iteration | 3824.3 | 4914.8 | 4681 | -4.76% |
+------------------------+------------+----------+-----------+--------------+
* Flutter gallery Instructions size is reduced by 11% (8748720 bytes to 7846368 bytes).
Baseline is at 6196496 bytes.
# Alternative Implementations
Alternatively to annotating individual procedures, I considered annotating Program node
with a set of dynamically dispatched selectors. Decoding and passing this information
around proved to be quite cumbersome in the "streaming" world, so I opted for a simpler
approach where all individual procedures are annotated.
Change-Id: I363db6d5dd1138fe25917646313c16d0b213c3b4
Reviewed-on: https://dart-review.googlesource.com/37822
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
* _ReceiverTypeBuilder class is introduced to create receiver type more
efficiently (without using union operations and creating intermediate
types).
* Certain tracing prints are hidden behind kPrintTrace flag to avoid
expensive string interpolations.
* Also, added debugging flag for dumping class hierarchy.
This CL reduces analysis time of Flutter gallery by ~2x.
Change-Id: Ia918b88a4d902895e77aac306dc6a5cf9cecad14
Reviewed-on: https://dart-review.googlesource.com/37644
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
This CL fixes the case when an invocation is invalidated in the
middle of processing. In such case, its result should not be retained
and it should be re-processed again.
Also:
* _WorkList.pending Set is replaced with LinkedList (reduces analysis time).
* ClassHierarchy is configured to ignore ambiguous supertypes.
* _AssertionError._throwNew entry point is added (as it is implicitly
inserted by flow graph builder).
Change-Id: Ie2d5d29352a5463fbe602910f9ae7f6f77a917d7
Reviewed-on: https://dart-review.googlesource.com/37126
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
* Invocation handlers and InvocationAnalyzer are revised to reduce
complexity of the code, improve precision and efficiency of analysis.
* Invocation now has 2 subclasses: DirectInvocation and
DispatchableInvocation; processing of invocations is moved to these
subclasses and simplified.
DirectInvocation is responsible for processing particular
methods and fields, while DispatchableInvocation processes polymorphic
calls by collecting targets and accurate receiver types for each target,
and delegating to DirectInvocations for processing of each target.
* Dynamic tear-offs and dynamic calls via getter/field are
supported.
* Invalidation of field initializers is implemented.
* Implicit 'return null' at the end of a function is correctly
detected.
Issue: https://github.com/dart-lang/sdk/issues/30480
Change-Id: I3d2ea69fbc95e3eb9aee244ee909016dfcf79bee
Reviewed-on: https://dart-review.googlesource.com/34507
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This is a pure refactoring which extracts the base class for
Devirtualization transformation. It is needed to add a new implementation
of devirtualization later, in addition to the current CHA-based
implementation.
Change-Id: Ifbc6160150a842f5322cf5aea13a7c2180507c96
Reviewed-on: https://dart-review.googlesource.com/30620
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
* The package:kernel/transformations/precompiler.dart is moved to
package:vm and split into 2 libraries: metadata/direct_call and
transformations/cha_devrtualization.
* Fasta 'target-options' command line argument and
package:kernel/target/implementation_option.dart are cleaned up
as they are no longer used.
Issue: https://github.com/dart-lang/sdk/issues/30480
Change-Id: I21a7b6bc62a036a9694e62a5dae890f6120d79ab
Reviewed-on: https://dart-review.googlesource.com/26360
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Vyacheslav Egorov