Follow up of https://dart-review.googlesource.com/c/sdk/+/117547
This gets rid of unnecessary allocations in hot loops with indexed loads and stores.
Issue: https://github.com/dart-lang/sdk/issues/38172
Change-Id: I37a4b1aba00084e465d47cce79bb9963e1afc104
Cq-Include-Trybots: luci.dart.try:vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,app-kernel-linux-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-dartkb-linux-debug-simarm64-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-dartkb-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-abi-try,vm-kernel-precomp-android-release-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-simarm64-try,vm-kernel-mac-debug-simdbc64-try,vm-kernel-precomp-android-release-arm_x64-try,vm-kernel-reload-mac-release-simdbc64-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-precomp-mac-release-simarm_x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119645
Reviewed-by: Martin Kustermann <kustermann@google.com>
Trying it out in flutter (for flutter test runs):
Running "flutter test" in flutter/packages/flutter:
BEFORE ENABLING
===============
$ flutter test --local-engine=host_debug
[...]
03:38 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug
[...]
03:11 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug
[...]
03:10 +5149 ~27: All tests passed!
With coverage:
$ flutter test --local-engine=host_debug --coverage
[...]
08:37 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug --coverage
[...]
08:21 +5149 ~27: All tests passed!
WITH THIS ENABLED
=================
$ flutter test --local-engine=host_debug
[...]
01:39 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug
[...]
01:32 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug
[...]
01:31 +5149 ~27: All tests passed!
With coverage:
$ flutter test --local-engine=host_debug --coverage
[...]
08:33 +5149 ~27: All tests passed!
$ flutter test --local-engine=host_debug --coverage
[...]
08:34 +5149 ~27: All tests passed!
So, without coverage more than twice as fast!
With coverage ~no change (which makes sense considering it's the VM thats spending extra time collecting coverage, so the compilation is not the bottleneck).
Running "flutter test" in a new project created with "flutter create" (i.e. with a single test):
BEFORE ENABLING
===============
$ time flutter test --local-engine=host_debug
00:05 +1: All tests passed!
real 0m6.360s
user 0m8.783s
sys 0m0.987s
$ time flutter test --local-engine=host_debug
00:02 +1: All tests passed!
real 0m2.790s
user 0m3.089s
sys 0m0.962s
$ time flutter test --local-engine=host_debug
00:02 +1: All tests passed!
real 0m2.693s
user 0m2.872s
sys 0m0.933s
WITH THIS ENABLED
=================
$ time flutter test --local-engine=host_debug
00:05 +1: All tests passed!
real 0m6.356s
user 0m8.253s
sys 0m1.161s
$ time flutter test --local-engine=host_debug
00:02 +1: All tests passed!
real 0m3.035s
user 0m3.449s
sys 0m0.880s
$ time flutter test --local-engine=host_debug
00:02 +1: All tests passed!
real 0m3.055s
user 0m3.534s
sys 0m0.885s
So maybe ~10% improvement on non-first runs (where it reuses the serialization from the dill it initialized from).
~no change on a clean run.
Change-Id: I2817fba9ca60190966194aa0b1ae8e16701137a0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/121121
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
In order to support causal-async-stacks, the following synthetic calls
are inserted: _asyncStackTraceHelper, _setAsyncThreadStackTrace and
_clearAsyncThreadStackTrace.
_asyncStackTraceHelper is always generated in the async transformation.
_setAsyncThreadStackTrace and _clearAsyncThreadStackTrace are inserted
during bytecode generation (or in the flow graph builder in case of AST
pipeline). If bytecode generation options are set inconsistently with
--causal-async-stacks VM option, then it was possible that
_asyncStackTraceHelper was generated but other calls were not generated,
causing incorrect (full, non-truncated) async stack traces. This also
causes performance regression as collecting full stack traces takes
much more time.
This change makes generation of all these calls more consistent: if
causal async stacks are disabled during bytecode generation, then all
three calls are omitted. Also, in case these calls were generated and
present in bytecode, but VM option --causal-async-stacks is disabled,
then bytecode flow graph builder turns all these three calls into no-ops.
Change-Id: I93eb1d83c675ee093799bb8e37ca3d60a3c5c19d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120927
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Previously, calls through field/getter were generated as dynamic calls
because InterfaceCall with field/getter target doesn't correctly
represent what such call should do. This results in a quite inefficient
code in AOT mode.
With this change, calls through fields/getters are decomposed into two
calls, while keeping the same order of evaluation.
Change-Id: Ieb33927c1b39d1b4d8eafb7bd542675820474068
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120628
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Original CL in patchset 1.
Fix for simdbc in patchset 4.
Fix for arm32 precompiled in: https://dart-review.googlesource.com/c/sdk/+/120660/
This CL optimizes Pointer operations in hot loops for Pointer<NativeInteger/NativeDouble/Pointer> (not for structs).
Design: go/dart-ffi-pointers-il
It provides roughly a 100x speedup for the FfiMemory benchmark. The next 5x speedup is to get rid of allocations due to `load` and `store` not being inlined.
FFI API is changed to enable optimizations:
* Disable dynamic invocations of Pointer.load / Pointer.store.
* Disallow implicit downcast of argument passed to Pointer.store.
* Stop zeroing out Pointer.address on Pointer.free().
Issue: https://github.com/dart-lang/sdk/issues/38172
Related issues:
Closes: https://github.com/dart-lang/sdk/issues/35902 (Disallowing dynamic invocations of Pointer ops.)
Closes: https://github.com/dart-lang/sdk/issues/37385 (Function variance checking)
Change-Id: I3921a595fd05026d6ca565ace496771d7c1d877b
Cq-Include-Trybots: luci.dart.try:vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,app-kernel-linux-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-dartkb-linux-debug-simarm64-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-dartkb-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-abi-try,vm-kernel-precomp-android-release-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-simarm64-try,vm-kernel-mac-debug-simdbc64-try,vm-kernel-precomp-android-release-arm_x64-try,vm-kernel-reload-mac-release-simdbc64-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-precomp-mac-release-simarm_x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120661
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
To support JavaScript compilation, the frontend_server will require a dependency on the dev_compiler. To avoid conflating this with the vm specific functionality, the frontend server will be split from its current location.
This change will require a small corresponding update in flutter/engine, documented in the patches directory
Change-Id: I47923765546f7f6fa43e36ef38f8f466d3a7b2fa
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120321
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
* AOT call specializer handles devirtualized static calls to integer
and double operations more efficiently, so direct call metadata
should be preferred over specialized bytecode instructions.
* Avoid clearing 'inferredTypeMetadata', which contains inferred-type
metadata for all members. Previously it was cleared after generating
the first member by mistake.
Change-Id: Ieaa6ad74cc2168c30b0c9865d17b195c49f85744
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120627
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Consume the extension flag setting for functions and the extension flag and is late
flags for fields when reading kernel file and set bit in Function/Field objects.
Change-Id: I661d30d53e817d968151d2b95474a6daf918de13
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/118000
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
After d65a60db00, we don't need AST with the bytecode to initialize a kernel compiler since the kernel compiler will initialize from a separate file.
Remove duplicated BinaryCacheMetadataRepository.
Fix generating bytecode in non-incremental mode.
Change-Id: Ica7bfb5a2d86dcc224849a9934a7a4676b01f075
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119702
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This CL optimizes Pointer operations in hot loops for Pointer<NativeInteger/NativeDouble/Pointer> (not for structs).
Design: go/dart-ffi-pointers-il
It provides roughly a 100x speedup for the FfiMemory benchmark. The next 5x speedup is to get rid of allocations due to `load` and `store` not being inlined.
FFI API is changed to enable optimizations:
* Disable dynamic invocations of Pointer.load / Pointer.store.
* Disallow implicit downcast of argument passed to Pointer.store.
* Stop zeroing out Pointer.address on Pointer.free().
Issue: https://github.com/dart-lang/sdk/issues/38172
Related issues:
Closes: https://github.com/dart-lang/sdk/issues/35902 (Disallowing dynamic invocations of Pointer ops.)
Closes: https://github.com/dart-lang/sdk/issues/37385 (Function variance checking.)
Change-Id: I96058d8b5b49052eb6999f084372e6f08b4f6f17
Cq-Include-Trybots: luci.dart.try:vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,app-kernel-linux-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-dartkb-linux-debug-simarm64-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-dartkb-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-abi-try,vm-kernel-precomp-android-release-arm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117547
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Previously, it was possible that type arguments of recursive types are
shared between unrelated recursive types if they have exactly the same shape
(e.g. Foo extends Base<Foo> and Bar extends Base<Bar> both have
type arguments <recursive-ref#0>).
Fixes ffi/structs_test in AOT+bytecode mode.
Change-Id: I516bc34c3b5797d76e1bf087ac3c8f075f1b6424
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120380
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
The bytecode encoding allows for 8-bits of context id, but this limit is exceeded in Flutter test code. Context ids are as yet unused, but will be used to disambiguate variables at the same position in different contexts.
Change-Id: I64130c11ddf04495b2719aa6f4809fa64ff135c1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119728
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
The interpretation of the subtype check result depends on the mode. In partial
NNBD mode all type errors that wouldn't be errors in non-NNBD programs should
become warnings. In full NNBD mode all such errors are errors. This CL adds
the mode explicitly to the interface of the subtype check as a parameter.
Bug: http://dartbug.com/38673
Change-Id: I14bcdd260618530cfdd8189c01abba7469a16679
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119545
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Dmitry Stefantsov <dmitryas@google.com>
AOT transformations add vm.direct-call.metadata with devirtualization
information to AST nodes.
Bytecode generator should take this information into account when
generating bytecode in case of AOT. Bytecode format is extended with
CheckReceiverForNull and UncheckedDirectCall instructions, and
DirectCallViaDynamicForwarder constant pool entry in order to represent
devirtualized calls.
Change-Id: I697432ddd0b58d2d0413715132ba5e90eb606ec1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119201
Reviewed-by: Régis Crelier <regis@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
flutter/examples/flutter_gallery$ flutter test --local-engine host_debug -v | grep took
First run -> Second run (initialize from dill file)
AST 11539ms -> 6837ms
Bytecode before 32790ms -> 26665ms
Bytecode after 15435ms -> 11111ms
Change-Id: If1088f86a583170c9f3690778a5b8e2fd4b48ace
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/118341
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Global variable globalDebuggingNames in the front-end can accumulate a
lot of garbage (all ASTs) if toString() is called from an AST node.
Bytecode generator uses Expression.getStaticType a lot, which calls
Expression.getStaticTypeAsInstanceOf, which sometimes calls toString()
when it throws an error (these crashes are tracked in
https://github.com/dart-lang/sdk/issues/34496 and bytecode generator
has a workaround to catch them).
As a result, after bytecode generation all ASTs were hanging out of
globalDebuggingNames, which leaks a lot of memory.
Fixes https://github.com/dart-lang/sdk/issues/38553
Change-Id: Ic1516d5ec9bf60861ae678e150c66d2e9394e2cd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/118914
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
When bytecode component is created without AST, Component.uriToSource
mapping is also dropped. Front-end server relies on Component.uriToSource
in order to generated dependencies file and send list of compiled source
to flutter tools. This change fixes listing of dependencies in bytecode mode
without AST.
Change-Id: I7c2b6f48571ec3ca42de330bdc1feef3ba7425d5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117596
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Fixes the changes to assembly snapshot writer in SIMARM_X64 mode. This unfortunately can't be tested until the ELF loader is available.
Take 2 is in patchset 1.
Change-Id: Ib8b067dd4f09dcba3b142705e9cb4dfbf4e3eb53
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117726
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Sometimes we need to stop at the first CheckStack and scan call stack in
the debugger, so debugger needs to know initial context level on function
entry, even before setting up initial context.
This change moves the first scope declaration before the first CheckStack.
Parameters which are not captured are also declared in the initial scope.
After creating an initial context we may add the 2nd scope, along with
captured parameters.
Fixes flaky crashes due to context level mismatch in
service/async_generator_breakpoint_test/1 with bytecode in debug mode.
Change-Id: I11ce1407ee58954b167e51a25302272b05365386
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117296
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
When generating deopt info in SIMDBC mode (DeoptInfoBuilder::AddCopy,
DeoptInfoBuilder::ToCpuRegisterSource) kSpecialDbcRegister source
locations are not supported.
In order to avoid such deopt info, this change moves CheckStack bytecode
instruction after function type arguments handling is finished (function
type arguments handling uses arguments descriptor which resides in
kArgsDescriptorsReg special DBC register). This placement of CheckStack is
also more aligned with what AST-based flow graph builder generates.
Fixes lib_2/math/min_max_test with bytecode in dartk-reload-mac-release-simdbc64
configuration.
Change-Id: I7d9863b8290a78a6e24e5806cc5053d55c93ce5c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117221
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Daco Harkes