Package org.antlr.v4.test.runtime.java.api.perf

Class ATNHotPathMicroBenchmark

java.lang.Object

org.antlr.v4.test.runtime.java.api.perf.ATNHotPathMicroBenchmark

public final class ATNHotPathMicroBenchmark
extends Object

Standalone micro-benchmark for ATN hot-path optimizations. Lives in the performance testsuite (not the runtime unit-test tree). Run with:

   mvn -pl perf-testsuite -DskipTests exec:java \
     -Dexec.mainClass=org.antlr.v4.test.runtime.java.api.perf.ATNHotPathMicroBenchmark
 

or directly from the IDE / java -cp ....

Reports nanosecond/op and relative speedups for:

• ATNConfig.hashCode() cache hit vs full MurmurHash recompute
• Lexer ATN reach with retained vs fresh OrderedATNConfigSet
• SLL config iteration without intermediate ArrayList

Field Summary

Fields

Modifier and Type	Field	Description
private static final int	ITERS
private static final int	WARMUP

Constructor Summary

Constructors

Constructor	Description
ATNHotPathMicroBenchmark()

Method Summary

Modifier and Type	Method	Description
private static void	benchHashCodeCache()	Compare cached ATNConfig.hashCode() against an equivalent uncached MurmurHash of the same fields.
private static void	benchLexerMatchThroughput()	End-to-end lexer match throughput on a small ATN with DFA cleared each outer batch so ATN simulation / reach remains on the hot path.
private static void	benchLexerReachReuse()	Measure cost of clearing+reusing an OrderedATNConfigSet vs allocating a new one each time (models LexerATNSimulator.computeTargetState).
private static void	benchSllConfigIteration()	Models SLL computeTargetState: index walk over ATNConfigSet vs copying into a new ArrayList first.
private static ATN	buildTinyLexerAtn()
private static void	fillReach(OrderedATNConfigSet set, ATNConfig[] pool)
private static void	lexAll(LexerATNSimulator sim, String text)
static void	main(String[] args)
private static int	uncachedHash(ATNConfig config)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

WARMUP

private static final int WARMUP

See Also:

• Constant Field Values

ITERS

private static final int ITERS

See Also:

• Constant Field Values

Constructor Details

ATNHotPathMicroBenchmark

public ATNHotPathMicroBenchmark()

Method Details

main

public static void main(String[] args)
                 throws Exception

Throws:

Exception

benchHashCodeCache

private static void benchHashCodeCache()

Compare cached ATNConfig.hashCode() against an equivalent uncached MurmurHash of the same fields.

uncachedHash

private static int uncachedHash(ATNConfig config)

benchLexerReachReuse

private static void benchLexerReachReuse()

Measure cost of clearing+reusing an OrderedATNConfigSet vs allocating a new one each time (models LexerATNSimulator.computeTargetState). Config objects are prebuilt so the measurement isolates set alloc/clear.

fillReach

private static void fillReach(OrderedATNConfigSet set,
 ATNConfig[] pool)

benchSllConfigIteration

private static void benchSllConfigIteration()

Models SLL computeTargetState: index walk over ATNConfigSet vs copying into a new ArrayList first.

benchLexerMatchThroughput

private static void benchLexerMatchThroughput()

End-to-end lexer match throughput on a small ATN with DFA cleared each outer batch so ATN simulation / reach remains on the hot path.

lexAll

private static void lexAll(LexerATNSimulator sim,
 String text)

buildTinyLexerAtn

private static ATN buildTinyLexerAtn()