forgetest1
/
antlr
forked from jasder/antlr
1
0
Fork 0
Code Issues Pull Requests 1 Releases Wiki Activity

Merge pull request #1791 from parrt/use-jol-for-object-size 

use JOL to compute buffer sizes. Dumps footprint now too.
This commit is contained in:
Terence Parr 2017-03-27 11:35:13 -07:00 committed by GitHub
parent 00d5781bee eb49080720
commit de4d129921
3 changed files with 188 additions and 49 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
runtime-testsuite/pom.xml
View File

@ -77,6 +77,11 @@
<version>1.0.4</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.openjdk.jol</groupId>
<artifactId>jol-core</artifactId>
<version>0.8</version>
</dependency>
</dependencies>
<build>

24
runtime-testsuite/test/org/antlr/v4/test/runtime/java/api/perf/Instrumentor.java Normal file
View File

@ -0,0 +1,24 @@
/*
* Copyright (c) 2012-2016 The ANTLR Project. All rights reserved.
* Use of this file is governed by the BSD 3-clause license that
* can be found in the LICENSE.txt file in the project root.
*/
package org.antlr.v4.test.runtime.java.api.perf;
import java.lang.instrument.Instrumentation;
/** Just a hook so we can call {@link java.lang.instrument.Instrumentation}
* methods like sizeof(). Start the Java VM with -javaagent instrumentor.jar
* if instrumentor.jar is where you put the .class file for this code.
* MANIFEST.MF for that jar must have "Premain-Class:Instrumentor".
*
* I'm not using at moment but I'm adding in case.
*/
public class Instrumentor {
public static Instrumentation instrumentation;
public static void premain(String args, Instrumentation I) {
instrumentation = I;
}
}

208
runtime-testsuite/test/org/antlr/v4/test/runtime/java/api/perf/TimeLexerSpeed.java
View File

@ -1,20 +1,25 @@
package org.antlr.v4.test.runtime.java.api.perf;
import org.antlr.v4.runtime.ANTLRFileStream;
import org.antlr.v4.runtime.ANTLRInputStream;
import org.antlr.v4.runtime.CharStream;
import org.antlr.v4.runtime.CharStreams;
import org.antlr.v4.runtime.CommonTokenStream;
import org.antlr.v4.runtime.Lexer;
import org.antlr.v4.test.runtime.java.api.JavaLexer;
import org.openjdk.jol.info.GraphLayout;
import java.io.BufferedReader;
import java.io.File;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.lang.management.ManagementFactory;
import java.lang.management.RuntimeMXBean;
import java.net.URL;
import java.nio.charset.StandardCharsets;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
@ -44,27 +49,98 @@ Warming up Java compiler....
*
Java VM args: -Xms2G -Xmx2G
Warming up Java compiler....
load_legacy_java_ascii average time 44us size 54301b over 3500 loads of 29038 symbols from Parser.java
load_legacy_java_utf8 average time 46us size 94675b over 3500 loads of 29038 symbols from Parser.java
load_legacy_java_utf8 average time 125us size 37490b over 3500 loads of 13379 symbols from udhr_hin.txt
load_new_utf8 average time 201us size 145893b over 3500 loads of 29038 symbols from Parser.java
load_new_utf8 average time 202us size 57482b over 3500 loads of 13379 symbols from udhr_hin.txt
load_legacy_java_ascii_file average time 53us size 58384b over 3500 loads of 29038 symbols from Parser.java
load_legacy_java_ascii_file average time 27us size 15568b over 3500 loads of 7625 symbols from RuleContext.java
load_legacy_java_ascii average time 53us size 65584b over 3500 loads of 29038 symbols from Parser.java
load_legacy_java_ascii average time 13us size 32816b over 3500 loads of 7625 symbols from RuleContext.java
load_legacy_java_utf8 average time 54us size 65584b over 3500 loads of 29038 symbols from Parser.java
load_legacy_java_utf8 average time 118us size 32816b over 3500 loads of 13379 symbols from udhr_hin.txt
load_new_utf8 average time 232us size 131232b over 3500 loads of 29038 symbols from Parser.java
load_new_utf8 average time 69us size 32928b over 3500 loads of 7625 symbols from RuleContext.java
load_new_utf8 average time 210us size 65696b over 3500 loads of 13379 symbols from udhr_hin.txt
lex_legacy_java_utf8 average time 367us over 2000 runs of 29038 symbols
lex_legacy_java_utf8 average time 907us over 2000 runs of 29038 symbols DFA cleared
lex_new_java_utf8 average time 434us over 2000 runs of 29038 symbols
lex_new_java_utf8 average time 963us over 2000 runs of 29038 symbols DFA cleared
lex_legacy_java_utf8 average time 342us over 2000 runs of 29038 symbols
lex_legacy_java_utf8 average time 890us over 2000 runs of 29038 symbols DFA cleared
lex_new_java_utf8 average time 439us over 2000 runs of 29038 symbols
lex_new_java_utf8 average time 969us over 2000 runs of 29038 symbols DFA cleared
lex_legacy_grapheme_utf8 average time 4028us over 400 runs of 6614 symbols from udhr_kor.txt
lex_legacy_grapheme_utf8 average time 4128us over 400 runs of 6614 symbols from udhr_kor.txt DFA cleared
lex_legacy_grapheme_utf8 average time 7661us over 400 runs of 13379 symbols from udhr_hin.txt
lex_legacy_grapheme_utf8 average time 7581us over 400 runs of 13379 symbols from udhr_hin.txt DFA cleared
lex_new_grapheme_utf8 average time 4026us over 400 runs of 6614 symbols from udhr_kor.txt
lex_new_grapheme_utf8 average time 4115us over 400 runs of 6614 symbols from udhr_kor.txt DFA cleared
lex_new_grapheme_utf8 average time 7547us over 400 runs of 13379 symbols from udhr_hin.txt
lex_new_grapheme_utf8 average time 7613us over 400 runs of 13379 symbols from udhr_hin.txt DFA cleared
lex_new_grapheme_utf8 average time 80us over 400 runs of 85 symbols from emoji.txt
lex_new_grapheme_utf8 average time 88us over 400 runs of 85 symbols from emoji.txt DFA cleared
lex_legacy_grapheme_utf8 average time 3971us over 400 runs of 6614 symbols from udhr_kor.txt
lex_legacy_grapheme_utf8 average time 4084us over 400 runs of 6614 symbols from udhr_kor.txt DFA cleared
lex_legacy_grapheme_utf8 average time 7542us over 400 runs of 13379 symbols from udhr_hin.txt
lex_legacy_grapheme_utf8 average time 7666us over 400 runs of 13379 symbols from udhr_hin.txt DFA cleared
lex_new_grapheme_utf8 average time 4034us over 400 runs of 6614 symbols from udhr_kor.txt
lex_new_grapheme_utf8 average time 4173us over 400 runs of 6614 symbols from udhr_kor.txt DFA cleared
lex_new_grapheme_utf8 average time 7680us over 400 runs of 13379 symbols from udhr_hin.txt
lex_new_grapheme_utf8 average time 7946us over 400 runs of 13379 symbols from udhr_hin.txt DFA cleared
lex_new_grapheme_utf8 average time 70us over 400 runs of 85 symbols from emoji.txt
lex_new_grapheme_utf8 average time 82us over 400 runs of 85 symbols from emoji.txt DFA cleared
*
* I dump footprint now too (this is 64-bit HotSpot VM):
*
Parser.java (29038 char): org.antlr.v4.runtime.ANTLRFileStream@6b8e0782d footprint:
COUNT AVG SUM DESCRIPTION
2 29164 58328 [C
1 24 24 java.lang.String
1 32 32 org.antlr.v4.runtime.ANTLRFileStream
4 58384 (total)
RuleContext.java (7625 char): org.antlr.v4.runtime.ANTLRFileStream@76fb7505d footprint:
COUNT AVG SUM DESCRIPTION
2 7756 15512 [C
1 24 24 java.lang.String
1 32 32 org.antlr.v4.runtime.ANTLRFileStream
4 15568 (total)
Parser.java (29038 char): org.antlr.v4.runtime.ANTLRInputStream@1fc1cb1d footprint:
COUNT AVG SUM DESCRIPTION
1 65552 65552 [C
1 32 32 org.antlr.v4.runtime.ANTLRInputStream
2 65584 (total)
RuleContext.java (7625 char): org.antlr.v4.runtime.ANTLRInputStream@2c6aa25dd footprint:
COUNT AVG SUM DESCRIPTION
1 32784 32784 [C
1 32 32 org.antlr.v4.runtime.ANTLRInputStream
2 32816 (total)
Parser.java (29038 char): org.antlr.v4.runtime.ANTLRInputStream@3d08db0bd footprint:
COUNT AVG SUM DESCRIPTION
1 65552 65552 [C
1 32 32 org.antlr.v4.runtime.ANTLRInputStream
2 65584 (total)
udhr_hin.txt (13379 char): org.antlr.v4.runtime.ANTLRInputStream@486dc6f3d footprint:
COUNT AVG SUM DESCRIPTION
1 32784 32784 [C
1 32 32 org.antlr.v4.runtime.ANTLRInputStream
2 32816 (total)
Parser.java (29038 char): org.antlr.v4.runtime.CodePointCharStream@798fe5a1d footprint:
COUNT AVG SUM DESCRIPTION
1 40 40 [C
1 131088 131088 [I
1 24 24 java.lang.String
1 48 48 java.nio.HeapIntBuffer
1 32 32 org.antlr.v4.runtime.CodePointCharStream
5 131232 (total)
RuleContext.java (7625 char): org.antlr.v4.runtime.CodePointCharStream@29cf5a20d footprint:
COUNT AVG SUM DESCRIPTION
1 40 40 [C
1 32784 32784 [I
1 24 24 java.lang.String
1 48 48 java.nio.HeapIntBuffer
1 32 32 org.antlr.v4.runtime.CodePointCharStream
5 32928 (total)
udhr_hin.txt (13379 char): org.antlr.v4.runtime.CodePointCharStream@1adb8a22d footprint:
COUNT AVG SUM DESCRIPTION
1 40 40 [C
1 65552 65552 [I
1 24 24 java.lang.String
1 48 48 java.nio.HeapIntBuffer
1 32 32 org.antlr.v4.runtime.CodePointCharStream
5 65696 (total)
*
* The "DFA cleared" indicates that the lexer was returned to initial conditions
* before the tokenizing of each file. As the ALL(*) lexer encounters new input,
@ -83,10 +159,13 @@ Warming up Java compiler....
*/
public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn test"
public static final String Parser_java_file = "Java/src/org/antlr/v4/runtime/Parser.java";
public static final String RuleContext_java_file = "Java/src/org/antlr/v4/runtime/RuleContext.java";
public static final String PerfDir = "org/antlr/v4/test/runtime/java/api/perf";
public boolean output = true;
public List<String> streamFootprints = new ArrayList<>();
public static void main(String[] args) throws Exception {
RuntimeMXBean runtimeMxBean = ManagementFactory.getRuntimeMXBean();
List<String> vmArgs = runtimeMxBean.getInputArguments();
@ -97,16 +176,21 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
}
}
System.out.println();
// System.out.println(VM.current().details());
TimeLexerSpeed tests = new TimeLexerSpeed();
tests.compilerWarmUp(100);
int n = 3500;
tests.load_legacy_java_ascii_file(Parser_java_file, n);
tests.load_legacy_java_ascii_file(RuleContext_java_file, n);
tests.load_legacy_java_ascii(Parser_java_file, n);
tests.load_legacy_java_ascii(RuleContext_java_file, n);
tests.load_legacy_java_utf8(Parser_java_file, n);
tests.load_legacy_java_utf8(PerfDir+"/udhr_hin.txt", n);
tests.load_new_utf8(Parser_java_file, n);
tests.load_new_utf8(RuleContext_java_file, n);
tests.load_new_utf8(PerfDir+"/udhr_hin.txt", n);
System.out.println();
@ -130,6 +214,10 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
tests.lex_new_grapheme_utf8("udhr_hin.txt", n, true);
tests.lex_new_grapheme_utf8("emoji.txt", n, false);
tests.lex_new_grapheme_utf8("emoji.txt", n, true);
for (String streamFootprint : tests.streamFootprints) {
System.out.print(streamFootprint);
}
}
public void compilerWarmUp(int n) throws Exception {
@ -147,6 +235,37 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
output = true;
}
public void load_legacy_java_ascii_file(String resourceName, int n) throws Exception {
URL sampleJavaFile = TimeLexerSpeed.class.getClassLoader().getResource(resourceName);
if ( sampleJavaFile==null ) {
System.err.println("Can't run load_legacy_java_ascii_file from jar (or can't find "+resourceName+")");
return; // cannot find resource
}
if ( !new File(sampleJavaFile.getFile()).exists() ) {
System.err.println("Can't run load_legacy_java_ascii_file from jar (or can't find "+resourceName+")");
return;
}
long start = System.nanoTime();
CharStream[] input = new CharStream[n]; // keep refs around so we can average memory
for (int i = 0; i<n; i++) {
input[i] = new ANTLRFileStream(sampleJavaFile.getFile());
}
long stop = System.nanoTime();
long tus = (stop-start)/1000;
int size = input[0].size();
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
GraphLayout olayout = GraphLayout.parseInstance((Object) input[0]);
long streamSize = olayout.totalSize();
streamFootprints.add(basename(resourceName)+" ("+size+" char): "+olayout.toFootprint());
if ( output ) System.out.printf("%27s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
currentMethodName,
tus/n,
streamSize,
n,
size,
basename(resourceName));
}
public void load_legacy_java_ascii(String resourceName, int n) throws Exception {
CharStream[] input = new CharStream[n]; // keep refs around so we can average memory
ClassLoader loader = TimeLexerSpeed.class.getClassLoader();
@ -154,8 +273,6 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
for (int i = 0; i<n; i++) {
streams[i] = loader.getResourceAsStream(resourceName);
}
System.gc();
long beforeFreeMem = Runtime.getRuntime().freeMemory();
long start = System.nanoTime(); // track only time to suck data out of stream
for (int i = 0; i<n; i++) {
try (InputStream is = streams[i]) {
@ -165,14 +282,13 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
long stop = System.nanoTime();
long tus = (stop-start)/1000;
int size = input[0].size();
System.gc();
long afterFreeMem = Runtime.getRuntime().freeMemory();
int avgStreamSize = (int)((beforeFreeMem-afterFreeMem) / (float)n);
long streamSize = GraphLayout.parseInstance((Object)input[0]).totalSize();
streamFootprints.add(basename(resourceName)+" ("+size+" char): "+GraphLayout.parseInstance((Object)input[0]).toFootprint());
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
if ( output ) System.out.printf("%27s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
currentMethodName,
tus/n,
avgStreamSize,
streamSize,
n,
size,
basename(resourceName));
@ -185,8 +301,6 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
for (int i = 0; i<n; i++) {
streams[i] = loader.getResourceAsStream(resourceName);
}
System.gc();
long beforeFreeMem = Runtime.getRuntime().freeMemory();
long start = System.nanoTime(); // track only time to suck data out of stream
for (int i = 0; i<n; i++) {
try (InputStream is = streams[i];
@ -198,17 +312,16 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
long stop = System.nanoTime();
long tus = (stop-start)/1000;
int size = input[0].size();
System.gc();
long afterFreeMem = Runtime.getRuntime().freeMemory();
int avgStreamSize = (int)((beforeFreeMem-afterFreeMem) / (float)n);
long streamSize = GraphLayout.parseInstance((Object)input[0]).totalSize();
streamFootprints.add(basename(resourceName)+" ("+size+" char): "+GraphLayout.parseInstance((Object)input[0]).toFootprint());
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
currentMethodName,
tus/n,
avgStreamSize,
n,
size,
basename(resourceName));
if ( output ) System.out.printf("%27s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
currentMethodName,
tus/n,
streamSize,
n,
size,
basename(resourceName));
}
public void load_new_utf8(String resourceName, int n) throws Exception {
@ -218,8 +331,6 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
for (int i = 0; i<n; i++) {
streams[i] = loader.getResourceAsStream(resourceName);
}
System.gc();
long beforeFreeMem = Runtime.getRuntime().freeMemory();
long start = System.nanoTime(); // track only time to suck data out of stream
for (int i = 0; i<n; i++) {
try (InputStream is = streams[i]) {
@ -229,14 +340,13 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
long stop = System.nanoTime();
long tus = (stop-start)/1000;
int size = input[0].size();
System.gc();
long afterFreeMem = Runtime.getRuntime().freeMemory();
int avgStreamSize = (int)((beforeFreeMem-afterFreeMem) / (float)n);
long streamSize = GraphLayout.parseInstance((Object)input[0]).totalSize();
streamFootprints.add(basename(resourceName)+" ("+size+" char): "+GraphLayout.parseInstance((Object)input[0]).toFootprint());
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
if ( output ) System.out.printf("%27s average time %5dus size %6db over %4d loads of %5d symbols from %s\n",
currentMethodName,
tus/n,
avgStreamSize,
streamSize,
n,
size,
basename(resourceName));
@ -250,7 +360,7 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
JavaLexer lexer = new JavaLexer(input);
double avg = tokenize(lexer, n, clearLexerDFACache);
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus over %4d runs of %5d symbols%s\n",
if ( output ) System.out.printf("%27s average time %5dus over %4d runs of %5d symbols%s\n",
currentMethodName,
(int)avg,
n,
@ -265,7 +375,7 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
JavaLexer lexer = new JavaLexer(input);
double avg = tokenize(lexer, n, clearLexerDFACache);
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus over %4d runs of %5d symbols%s\n",
if ( output ) System.out.printf("%27s average time %5dus over %4d runs of %5d symbols%s\n",
currentMethodName,
(int)avg,
n,
@ -282,7 +392,7 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
graphemesLexer lexer = new graphemesLexer(input);
double avg = tokenize(lexer, n, clearLexerDFACache);
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus over %4d runs of %5d symbols from %s%s\n",
if ( output ) System.out.printf("%27s average time %5dus over %4d runs of %5d symbols from %s%s\n",
currentMethodName,
(int)avg,
n,
@ -298,7 +408,7 @@ public class TimeLexerSpeed { // don't call it Test else it'll run during "mvn t
graphemesLexer lexer = new graphemesLexer(input);
double avg = tokenize(lexer, n, clearLexerDFACache);
String currentMethodName = new Exception().getStackTrace()[0].getMethodName();
if ( output ) System.out.printf("%25s average time %5dus over %4d runs of %5d symbols from %s%s\n",
if ( output ) System.out.printf("%27s average time %5dus over %4d runs of %5d symbols from %s%s\n",
currentMethodName,
(int)avg,
n,