io.druid.query.topn

Class Generic2AggPooledTopNScannerPrototype

java.lang.Object

io.druid.query.topn.Generic2AggPooledTopNScannerPrototype

All Implemented Interfaces:

Generic2AggPooledTopNScanner

public final class Generic2AggPooledTopNScannerPrototype
extends Object
implements Generic2AggPooledTopNScanner

Constructor Summary

Constructors

Constructor and Description
Generic2AggPooledTopNScannerPrototype()

Method Summary

Modifier and Type	Method and Description
long	scanAndAggregate(DimensionSelector dimensionSelector, BufferAggregator aggregator1, int aggregator1Size, BufferAggregator aggregator2, int aggregator2Size, Cursor cursor, int[] positions, ByteBuffer resultsBuffer) Any changes to this method should be coordinated with TopNUtils, PooledTopNAlgorithm.computeSpecializedScanAndAggregateImplementations() and downstream methods.

Methods inherited from class java.lang.Object

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

Constructor Detail

Generic2AggPooledTopNScannerPrototype

public Generic2AggPooledTopNScannerPrototype()

Method Detail

scanAndAggregate

public long scanAndAggregate(DimensionSelector dimensionSelector,
                             BufferAggregator aggregator1,
                             int aggregator1Size,
                             BufferAggregator aggregator2,
                             int aggregator2Size,
                             Cursor cursor,
                             int[] positions,
                             ByteBuffer resultsBuffer)

Any changes to this method should be coordinated with TopNUtils, PooledTopNAlgorithm.computeSpecializedScanAndAggregateImplementations() and downstream methods. It should be checked with a tool like https://github.com/AdoptOpenJDK/jitwatch that C2 compiler output for this method doesn't have any method calls in the while loop, i. e. all method calls are inlined. To be able to see assembly of this method in JITWatch and other similar tools, PooledTopNAlgorithm.specializeGeneric2AggPooledTopN should be turned off. Note that in this case the benchmark should be "naturally monomorphic", i. e. execute this method always with the same runtime shape. If the while loop contains not inlined method calls, it should be considered as a performance bug.

Specified by:

scanAndAggregate in interface Generic2AggPooledTopNScanner

aggregator1Size - number of bytes required by aggregator1 for a single aggregation

aggregator2Size - number of bytes required by aggregator2 for a single aggregation

positions - a cache for positions in resultsBuffer, where specific (indexed) dimension values are aggregated

Returns:

number of scanned rows, i. e. number of steps made with the given cursor