lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Cass Costello" <cas...@gmail.com>
Subject OpenBitSet performance question
Date Sat, 22 Jul 2006 19:20:01 GMT
Hello all,

I'm newish to both Lucene and Solr, but I'm loving learning both.  I was
intrigued by the comments in the OpenBitSet javadocs...

<snip>
OpenBitSet is faster than java.util.BitSet in most operations and *much*
faster at calculating cardinality of sets and results of set operations.
</snip>

...so I decided to see for myself and concocted a couple tests, which
consistently result in...

Took 421ms to get 5000 bitset intersection counts
Took 1465ms to get 5000 openbitset intersection counts

...and I'm wondering what I've done wrong.  The results are consistent
across differenct jvms and different hardware setups.  I'm using the 7/22
nightly of Solr.  See my test code below.

FYI, my interest in Solr/Lucene stems from a need to create a facetted
browse experience with 10s of thousands of facets derived from millions of
documents.

Thanks for your time,
Cass Costello


package playground.cass;

import java.util.BitSet;
import java.util.HashMap;
import java.util.Map;
import java.util.Set;
import junit.framework.TestCase;
import org.apache.solr.util.OpenBitSet;

public class TestsBitSetOperations extends TestCase {
    public TestsBitSetOperations(String arg0) {
        super(arg0);
    }

    protected void setUp() throws Exception {
        super.setUp();
    }

    protected void tearDown() throws Exception {
        super.tearDown();
    }

    public void testMultipleOpenBitSetAndsSpeed() {


        int i = 0;
        try {
            //create 5000 openbitsets, put them all in a single map
            Map<String,OpenBitSet> bsMap = new HashMap<String,OpenBitSet>();
            OpenBitSet bs;
            for ( ; i < 5000; i++ ) {
                bs = new OpenBitSet( 500000 );
                bs.fastSet( 5000 );
                bsMap.put( String.valueOf(i), bs );
            }

            //our comparsion openbitset
            OpenBitSet bs2 = new OpenBitSet( 500000 );
            bs2.fastSet( 5000 );

//            ensure jvm warmup
            Set<String> set = bsMap.keySet();
            for ( String key : set ) {
                OpenBitSet.intersectionCount( bs2, bsMap.get( key ) );
            }

            // meat of the test
            long start = System.currentTimeMillis();
            for ( String key : set ) {
                long count = OpenBitSet.intersectionCount( bs2, bsMap.get(
key ) );
            }
            long finish = System.currentTimeMillis();

            System.out.println( "Took " + (finish-start) + "ms to get 5000
openbitset intersection counts" );

        } catch ( Throwable t ) {
            t.printStackTrace( System.out );
            System.out.println( i );
        }
    }

    public void testMultipleBitSetAndsSpeed() {


        int i = 0;
        try {
            //create 5000 bitsets, put them all in a single map
            Map<String,BitSet> bsMap = new HashMap<String,BitSet>();
            BitSet bs;
            for ( ; i < 5000; i++ ) {
                bs = new BitSet( 500000 );
                bs.set( 5000 );
                bsMap.put( String.valueOf(i), bs );
            }

            BitSet bs2 = new BitSet( 500000 );
            bs2.set( 5000 );

            //ensure jvm warmup
            Set<String> set = bsMap.keySet();
            for ( String key : set ) {
                BitSet copy = (BitSet)bs2.clone();
                copy.and( bsMap.get( key ) );
            }

            BitSet copy;

//            meat of the test
            long start = System.currentTimeMillis();
            for ( String key : set ) {
                copy = (BitSet)bs2.clone();
                copy.and( bsMap.get( key ) );
                long count = copy.cardinality();
            }
            long finish = System.currentTimeMillis();

            System.out.println( "Took " + (finish-start) + "ms to get 5000
bitset intersection counts" );

        } catch ( Throwable t ) {
            t.printStackTrace( System.out );
            System.out.println( i );
        }

    }


}

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message