Skip to content

1.x: benchmark just() and its optimizations. #3495

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 5, 2015
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
197 changes: 197 additions & 0 deletions src/perf/java/rx/ScalarJustPerf.java
Original file line number Diff line number Diff line change
@@ -0,0 +1,197 @@
/**
* Copyright 2014 Netflix, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

package rx;

import java.util.concurrent.TimeUnit;

import org.openjdk.jmh.annotations.*;
import org.openjdk.jmh.infra.Blackhole;

import rx.functions.Func1;
import rx.jmh.LatchedObserver;
import rx.schedulers.Schedulers;

/**
* Benchmark the cost of just and its various optimizations.
* <p>
* gradlew benchmarks "-Pjmh=-f 1 -tu s -bm thrpt -wi 5 -i 5 -r 1 .*ScalarJustPerf.*"
* <p>
* gradlew benchmarks "-Pjmh=-f 1 -tu ns -bm avgt -wi 5 -i 5 -r 1 .*ScalarJustPerf.*"
*/
@BenchmarkMode(Mode.Throughput)
@OutputTimeUnit(TimeUnit.SECONDS)
@State(Scope.Thread)
public class ScalarJustPerf {
/** A subscriber without a CountDownLatch; use it for synchronous testing only. */
static final class PlainSubscriber extends Subscriber<Integer> {
final Blackhole bh;
public PlainSubscriber(Blackhole bh) {
this.bh = bh;
}

@Override
public void onNext(Integer t) {
bh.consume(t);
}

@Override
public void onError(Throwable e) {
bh.consume(e);
}

@Override
public void onCompleted() {
bh.consume(false);
}
}

/** This is a simple just. */
Observable<Integer> simple;
/**
* This is a simple just observed on the computation scheduler.
* The current computation scheduler supports direct scheduling and should have
* lower overhead than a regular createWorker-use-unsubscribe.
*/
Observable<Integer> observeOn;
/** This is a simple just observed on the IO thread. */
Observable<Integer> observeOnIO;

/**
* This is a simple just subscribed to on the computation scheduler.
* In theory, for non-backpressured just(), this should be the
* same as observeOn.
*/
Observable<Integer> subscribeOn;
/** This is a simple just subscribed to on the IO scheduler. */
Observable<Integer> subscribeOnIO;

/** This is a just mapped to itself which should skip the operator flatMap completely. */
Observable<Integer> justFlatMapJust;
/**
* This is a just mapped to a range of 2 elements; it tests the case where the inner
* Observable isn't a just().
*/
Observable<Integer> justFlatMapRange;

@Setup
public void setup() {
simple = Observable.just(1);

observeOn = simple.observeOn(Schedulers.computation());
observeOnIO = simple.observeOn(Schedulers.io());

subscribeOn = simple.subscribeOn(Schedulers.computation());
subscribeOnIO = simple.subscribeOn(Schedulers.io());

justFlatMapJust = simple.flatMap(new Func1<Integer, Observable<Integer>>() {
@Override
public Observable<Integer> call(Integer v) {
return Observable.just(v);
}
});

justFlatMapRange = simple.flatMap(new Func1<Integer, Observable<Integer>>() {
@Override
public Observable<Integer> call(Integer v) {
return Observable.range(v, 2);
}
});
}

/**
* Common routine to create a latched observer, subscribe it to the
* given source and spin-wait for its completion.
* <p>Don't use this with long sources. The spin-wait is there
* to avoid operating-system level scheduling-wakeup granularity problems with
* short sources.
* @param bh the black hole to sink values and prevent dead code elimination
* @param source the source observable to observe
*/
void runAsync(Blackhole bh, Observable<Integer> source) {
LatchedObserver<Integer> lo = new LatchedObserver<Integer>(bh);

source.subscribe(lo);

while (lo.latch.getCount() != 0L);
}

@Benchmark
public void simple(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
simple.subscribe(s);
}

@Benchmark
public void simpleEscape(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
bh.consume(s);
simple.subscribe(s);
}

@Benchmark
public Object simpleEscapeAll(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
bh.consume(s);
return simple.subscribe(s);
}

@Benchmark
public void observeOn(Blackhole bh) {
runAsync(bh, observeOn);
}

@Benchmark
public void observeOnIO(Blackhole bh) {
runAsync(bh, observeOnIO);
}

@Benchmark
public void subscribeOn(Blackhole bh) {
runAsync(bh, subscribeOn);
}

@Benchmark
public void subscribeOnIO(Blackhole bh) {
runAsync(bh, subscribeOnIO);
}

@Benchmark
public void justFlatMapJust(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
justFlatMapJust.subscribe(s);
}

@Benchmark
public void justFlatMapJustEscape(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
bh.consume(s);
justFlatMapJust.subscribe(s);
}

@Benchmark
public void justFlatMapRange(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
justFlatMapRange.subscribe(s);
}

@Benchmark
public void justFlatMapRangeEscape(Blackhole bh) {
PlainSubscriber s = new PlainSubscriber(bh);
bh.consume(s);
justFlatMapRange.subscribe(s);
}
}