Improvements in the Inliner.
- Change from a depth limit to a total number of HInstructions
inlined limit. Remove the dex2oat depth limit argument.
- Add more stats to diagnose reasons for not inlining.
- Clean up logging to easily parse output.
Individual Ritz benchmarks improve from 3 to 10%.
No change in other heuristics. There was already an instruction budget.
Note that the instruction budget is rarely hit in the "apps" I've tried
with.
Compile-times improve from 5 to 15%.
Code size go from 4% increase (Gms) to 1% decrease (Docs).
bug:35724239
test: test-art-host test-art-target
Change-Id: I5a35c4bd826cf21fead77859709553c5b57608d6
diff --git a/compiler/optimizing/optimizing_compiler.cc b/compiler/optimizing/optimizing_compiler.cc
index 23ccd9e..3c6d2d6 100644
--- a/compiler/optimizing/optimizing_compiler.cc
+++ b/compiler/optimizing/optimizing_compiler.cc
@@ -499,7 +499,8 @@
handles,
stats,
number_of_dex_registers,
- /* depth */ 0);
+ /* total_number_of_instructions */ 0,
+ /* parent */ nullptr);
} else if (opt_name == HSharpening::kSharpeningPassName) {
return new (arena) HSharpening(graph, codegen, dex_compilation_unit, driver, handles);
} else if (opt_name == HSelectGenerator::kSelectGeneratorPassName) {
@@ -607,8 +608,7 @@
VariableSizedHandleScope* handles) const {
OptimizingCompilerStats* stats = compilation_stats_.get();
const CompilerOptions& compiler_options = driver->GetCompilerOptions();
- bool should_inline = (compiler_options.GetInlineDepthLimit() > 0)
- && (compiler_options.GetInlineMaxCodeUnits() > 0);
+ bool should_inline = (compiler_options.GetInlineMaxCodeUnits() > 0);
if (!should_inline) {
return;
}
@@ -623,7 +623,8 @@
handles,
stats,
number_of_dex_registers,
- /* depth */ 0);
+ /* total_number_of_instructions */ 0,
+ /* parent */ nullptr);
HOptimization* optimizations[] = { inliner };
RunOptimizations(optimizations, arraysize(optimizations), pass_observer);