On Fri, 1 Jul 2022 18:32:18 GMT, Peter Levart <plev...@openjdk.org> wrote:
>> @plevart I've checked it with and without `@stable`, it's the same:
>>
>> with
>> Benchmark Mode Cnt Score Error Units
>> AccessParamsBenchmark.getParameter0 avgt 50 6,196 ± 0,142 ns/op
>> AccessParamsBenchmark.getParameters avgt 50 4,636 ± 0,075 ns/op
>>
>> without
>> Benchmark Mode Cnt Score Error Units
>> AccessParamsBenchmark.getParameter0 avgt 50 6,196 ± 0,142 ns/op
>> AccessParamsBenchmark.getParameters avgt 50 4,636 ± 0,075 ns/op
>>
>> This seems logical as effects of `@Stable` aren not propagated into cloned
>> array.
>> This is the benchmark I used:
>>
>> @State(Scope.Thread)
>> @BenchmarkMode(Mode.AverageTime)
>> @OutputTimeUnit(TimeUnit.NANOSECONDS)
>> @Fork(jvmArgsAppend = {"-Xms1g", "-Xmx1g"})
>> public class AccessParamsBenchmark {
>>
>> private final Method method;
>>
>> public AccessParamsBenchmark() {
>> try {
>> method = getClass().getMethod("foo", int.class, String.class);
>> } catch (NoSuchMethodException e) {
>> throw new RuntimeException(e);
>> }
>> }
>>
>> @Benchmark
>> public Object getParameters() {
>> return method.getParameters();
>> }
>>
>> @Benchmark
>> public Object getParameter0() {
>> return method.getParameters()[0];
>> }
>>
>> public void foo(int parameter1, String parameter2) {
>> }
>> }
>>
>>
>> So, should we rid the annotation from `parameters` field?
>
> @Stable is only effective if the path leading to @Stable value can be
> constant-folded by JIT. In above test, you have an instance field Method
> method. This can not be constant-folded, so neither can @stable fiels in the
> Field object, nor array elements of a @stable array. You should replace that
> with "static final Method method = ..." and re-run the test.
With `static` and without `@Stable` the benchmark yields
Benchmark Mode Cnt Score Error Units
AccessParamsBenchmark.getParameter0 avgt 10 1,212 ± 0,083 ns/op
AccessParamsBenchmark.getParameters avgt 10 2,493 ± 0,076 ns/op
and with `@Stable`
Benchmark Mode Cnt Score Error Units
AccessParamsBenchmark.getParameter0 avgt 40 0,427 ± 0,007 ns/op
AccessParamsBenchmark.getParameters avgt 40 2,123 ± 0,052 ns/op
so the annotation is useful.
-------------
PR: https://git.openjdk.org/jdk/pull/9143
