If we inline something inside a loop which might throw, the throwing blocks
are currently spliced in the middle of the loop. This CL moves those blocks
to the very end.
This improves a number of typed data benchmarks in dart-aot mode by 5-10%.
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: I5dc86291240d8dac61798ff873ffa7205edc0007
Reviewed-on: https://dart-review.googlesource.com/c/85263
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Rationale:
Rather than relying on a separate loop detector, rely
on the new loop framework, which avoids code duplication
and ensures any improvement in loop detection/handling
will benefit this phase too. Note, most of the time, the
same loops are discovered with a few exceptions (which
is okay, since this is "just" heuristic usage). This CL
also simplifies loop detection a bit.
https://github.com/dart-lang/sdk/issues/34473
Change-Id: I1a1b19b99a698c74822473d2a1fe370287c1ade4
Reviewed-on: https://dart-review.googlesource.com/c/80523
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
The extra call specialization pass after constant propagation is able to
replace more calls in AOT, as constant propagation removes unreachable
code and subsequent type propagation infers more accurate types.
Performance improvement (armv8):
DartMicroBench.IntPower3 196.4%
DartMicroBench.IntPower2 273.1%
Change-Id: I7076608f828484c4045e7a7b605ed8ff7981a868
Reviewed-on: https://dart-review.googlesource.com/75564
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Some optimizations were checking for a number of successors of graph
entry - but that count includes unchecked entry and indirect entries.
Change-Id: Ib163a7710aaba3e68023a9cd04ca194c03262c33
Reviewed-on: https://dart-review.googlesource.com/71302
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Test Plan:
Behavioral correctness should be ensured by existing tests. Tests in vm/dart/entrypoints
ensure that the unchecked entrypoint is used in cases where the optimization should trigger.
Bug: https://github.com/dart-lang/sdk/issues/31798
Change-Id: I5b880b2dfa6343b4bb0a96ad23562facff73e41f
Cq-Include-Trybots: luci.dart.try:vm-kernel-win-release-x64-try,vm-kernel-optcounter-threshold-linux-release-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-linux-release-simarm64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/69741
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Rationale:
Running an extra call specialization pass a bit later
in the compiler passes stream (after types are propagated)
recognizes more operators, which execute more efficiently.
In addition, unboxing phis seems useful on all archs, not
just 32-bit. These minor tweeks combined improve the prime
number benchmark 3 fold.
https://github.com/flutter/flutter/issues/19677
Change-Id: Ib2102ce807c2f0a9f801542e0c4bc6a7673f552b
Reviewed-on: https://dart-review.googlesource.com/69240
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
If the left-hand side of the store is an allocation and no GC-triggering
instructions have been executed since, it is safe to elide the store barrier.
This gives a 1.4% code size reduction on Flutter Gallery ARM32.
Change-Id: Ib7227d3ef9d798d5e30f238f3f789f9a2e637d6a
Cq-Include-Trybots: luci.dart.try: vm-kernel-optcounter-threshold-linux-release-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-linux-release-simarm64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/64687
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
If either of print-flow-graph/print-flow-graph-optimized is passed then
print the flow-graph after the register allocation.
Change-Id: If1ad9117ee2c1d5bf7d3608110f9f72eee7393e2
Reviewed-on: https://dart-review.googlesource.com/50726
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Passes bodies are moved into compiler_pass{.cc,.h}.
Invoking a pass is just INVOKE_PASS(Name) now, instead of putting a bunch of
if-s and calls in compiler.cc or precompiler.cc.
We also consolidate ability to print IL and enable-disable passes under a
single flag --compiler-passes, e.g.
--compiler-passes=-Inlining,-CSE disable inlining and CSE passes
--compiler-passes=Inlining+ print IL after Inlining pass and all
subsequent passes
Change-Id: I90ff54b04a54f20099f5bf38dd45b16b8e3c4781
Reviewed-on: https://dart-review.googlesource.com/43968
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Aart Bik