/* * Copyright (C) 2015 Google 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 benchmarks.regression; import com.google.caliper.BeforeExperiment; import junit.framework.Assert; /** * Benchmarks to measure the performance of String.equals for Strings of varying lengths. * Each benchmarks makes 5 measurements, aiming at covering cases like strings of equal length * that are not equal, identical strings with different references, strings with different endings, * interned strings, and strings of different lengths. */ public class StringEqualsBenchmark { private final String long1 = "Ahead-of-time compilation is possible as the compiler may just" + "convert an instruction thus: dex code: add-int v1000, v2000, v3000 C code: setIntRegter" + "(1000, call_dex_add_int(getIntRegister(2000), getIntRegister(3000)) This means even lid" + "instructions may have code generated, however, it is not expected that code generate in" + "this way will perform well. The job of AOT verification is to tell the compiler that" + "instructions are sound and provide tests to detect unsound sequences so slow path code" + "may be generated. Other than for totally invalid code, the verification may fail at AOr" + "run-time. At AOT time it can be because of incomplete information, at run-time it can e" + "that code in a different apk that the application depends upon has changed. The Dalvik" + "verifier would return a bool to state whether a Class were good or bad. In ART the fail" + "case becomes either a soft or hard failure. Classes have new states to represent that a" + "soft failure occurred at compile time and should be re-verified at run-time."; private final String veryLong = "Garbage collection has two phases. The first distinguishes" + "live objects from garbage objects. The second is reclaiming the rage of garbage object" + "In the mark-sweep algorithm used by Dalvik, the first phase is achievd by computing the" + "closure of all reachable objects in a process known as tracing from theoots. After the" + "trace has completed, garbage objects are reclaimed. Each of these operations can be" + "parallelized and can be interleaved with the operation of the applicationTraditionally," + "the tracing phase dominates the time spent in garbage collection. The greatreduction i" + "pause time can be achieved by interleaving as much of this phase as possible with the" + "application. If we simply ran the GC in a separate thread with no other changes, normal" + "operation of an application would confound the trace. Abstractly, the GC walks the h o" + "all reachable objects. When the application is paused, the object graph cannot change." + "The GC can therefore walk this structure and assume that all reachable objects live." + "When the application is running, this graph may be altered. New nodes may be addnd edge" + "may be changed. These changes may cause live objects to be hidden and falsely recla by" + "the GC. To avoid this problem a write barrier is used to intercept and record modifion" + "to objects in a separate structure. After performing its walk, the GC will revisit the" + "updated objects and re-validate its assumptions. Without a card table, the garbage" + "collector would have to visit all objects reached during the trace looking for dirtied" + "objects. The cost of this operation would be proportional to the amount of live data." + "With a card table, the cost of this operation is proportional to the amount of updateat" + "The write barrier in Dalvik is a card marking write barrier. Card marking is the proce" + "of noting the location of object connectivity changes on a sub-page granularity. A car" + "is merely a colorful term for a contiguous extent of memory smaller than a page, common" + "somewhere between 128- and 512-bytes. Card marking is implemented by instrumenting all" + "locations in the virtual machine which can assign a pointer to an object. After themal" + "pointer assignment has occurred, a byte is written to a byte-map spanning the heap whic" + "corresponds to the location of the updated object. This byte map is known as a card ta" + "The garbage collector visits this card table and looks for written bytes to reckon the" + "location of updated objects. It then rescans all objects located on the dirty card," + "correcting liveness assumptions that were invalidated by the application. While card" + "marking imposes a small burden on the application outside of a garbage collection, the" + "overhead of maintaining the card table is paid for by the reduced time spent inside" + "garbage collection. With the concurrent garbage collection thread and a write barrier" + "supported by the interpreter, JIT, and Runtime we modify garbage collection"; private final String[][] shortStrings = new String[][] { // Equal, constant comparison { "a", "a" }, // Different constants, first character different { ":", " :"}, // Different constants, last character different, same length { "ja M", "ja N"}, // Different constants, different lengths {"$$$", "$$"}, // Force execution of code beyond reference equality check {"hi", new String("hi")} }; private final String[][] mediumStrings = new String[][] { // Equal, constant comparison { "Hello my name is ", "Hello my name is " }, // Different constants, different lengths { "What's your name?", "Whats your name?" }, // Force execution of code beyond reference equality check { "Android Runtime", new String("Android Runtime") }, // Different constants, last character different, same length { "v3ry Cre@tiVe?****", "v3ry Cre@tiVe?***." }, // Different constants, first character different, same length { "!@#$%^&*()_++*^$#@", "0@#$%^&*()_++*^$#@" } }; private final String[][] longStrings = new String[][] { // Force execution of code beyond reference equality check { long1, new String(long1) }, // Different constants, last character different, same length { long1 + "fun!", long1 + "----" }, // Equal, constant comparison { long1 + long1, long1 + long1 }, // Different constants, different lengths { long1 + "123456789", long1 + "12345678" }, // Different constants, first character different, same length { "Android Runtime" + long1, "android Runtime" + long1 } }; private final String[][] veryLongStrings = new String[][] { // Force execution of code beyond reference equality check { veryLong, new String(veryLong) }, // Different constants, different lengths { veryLong + veryLong, veryLong + " " + veryLong }, // Equal, constant comparison { veryLong + veryLong + veryLong, veryLong + veryLong + veryLong }, // Different constants, last character different, same length { veryLong + "77777", veryLong + "99999" }, // Different constants, first character different { "Android Runtime" + veryLong, "android Runtime" + veryLong } }; private final String[][] endStrings = new String[][] { // Different constants, medium but different lengths { "Hello", "Hello " }, // Different constants, long but different lengths { long1, long1 + "x"}, // Different constants, very long but different lengths { veryLong, veryLong + "?"}, // Different constants, same medium lengths { "How are you doing today?", "How are you doing today " }, // Different constants, short but different lengths { "1", "1." } }; private final String tmpStr1 = "012345678901234567890" + "0123456789012345678901234567890123456789" + "0123456789012345678901234567890123456789" + "0123456789012345678901234567890123456789" + "0123456789012345678901234567890123456789"; private final String tmpStr2 = "z012345678901234567890" + "0123456789012345678901234567890123456789" + "0123456789012345678901234567890123456789" + "0123456789012345678901234567890123456789" + "012345678901234567890123456789012345678x"; private final String[][] nonalignedStrings = new String[][] { // Different non-word aligned medium length strings { tmpStr1, tmpStr1.substring(1) }, // Different differently non-word aligned medium length strings { tmpStr2, tmpStr2.substring(2) }, // Different non-word aligned long length strings { long1, long1.substring(3) }, // Different non-word aligned very long length strings { veryLong, veryLong.substring(1) }, // Equal non-word aligned constant strings { "hello", "hello".substring(1) } }; private final Object[] objects = new Object[] { // Compare to Double object new Double(1.5), // Compare to Integer object new Integer(9999999), // Compare to String array new String[] {"h", "i"}, // Compare to int array new int[] {1, 2, 3}, // Compare to Character object new Character('a') }; // Check assumptions about how the compiler, new String(String), and String.intern() work. // Any failures here would invalidate these benchmarks. @BeforeExperiment protected void setUp() throws Exception { // String constants are the same object Assert.assertSame("abc", "abc"); // new String(String) makes a copy Assert.assertNotSame("abc" , new String("abc")); // Interned strings are treated like constants, so it is not necessary to // separately benchmark interned strings. Assert.assertSame("abc", "abc".intern()); Assert.assertSame("abc", new String("abc").intern()); // Compiler folds constant strings into new constants Assert.assertSame(long1 + long1, long1 + long1); } // Benchmark cases of String.equals(null) public void timeEqualsNull(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < mediumStrings.length; i++) { mediumStrings[i][0].equals(null); } } } // Benchmark cases with very short (<5 character) Strings public void timeEqualsShort(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < shortStrings.length; i++) { shortStrings[i][0].equals(shortStrings[i][1]); } } } // Benchmark cases with medium length (10-15 character) Strings public void timeEqualsMedium(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < mediumStrings.length; i++) { mediumStrings[i][0].equals(mediumStrings[i][1]); } } } // Benchmark cases with long (>100 character) Strings public void timeEqualsLong(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < longStrings.length; i++) { longStrings[i][0].equals(longStrings[i][1]); } } } // Benchmark cases with very long (>1000 character) Strings public void timeEqualsVeryLong(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < veryLongStrings.length; i++) { veryLongStrings[i][0].equals(veryLongStrings[i][1]); } } } // Benchmark cases with non-word aligned Strings public void timeEqualsNonWordAligned(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < nonalignedStrings.length; i++) { nonalignedStrings[i][0].equals(nonalignedStrings[i][1]); } } } // Benchmark cases with slight differences in the endings public void timeEqualsEnd(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < endStrings.length; i++) { endStrings[i][0].equals(endStrings[i][1]); } } } // Benchmark cases of comparing a string to a non-string object public void timeEqualsNonString(int reps) { for (int rep = 0; rep < reps; ++rep) { for (int i = 0; i < mediumStrings.length; i++) { mediumStrings[i][0].equals(objects[i]); } } } }