org.apache.sysml.runtime.compress.estim

Class CompressedSizeEstimator

java.lang.Object

org.apache.sysml.runtime.compress.estim.CompressedSizeEstimator

Direct Known Subclasses:

CompressedSizeEstimatorExact, CompressedSizeEstimatorSample

public abstract class CompressedSizeEstimator
extends Object

Base class for all compressed size estimators

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	CompressedSizeEstimator.SizeEstimationFactors

Field Summary

Fields

Modifier and Type	Field and Description
protected MatrixBlock	_data
protected int	_numRows

Constructor Summary

Constructors

Constructor and Description
CompressedSizeEstimator(MatrixBlock data)

Method Summary

Modifier and Type	Method and Description
protected CompressedSizeEstimator.SizeEstimationFactors	computeSizeEstimationFactors(UncompressedBitmap ubm, boolean inclRLE)
abstract CompressedSizeInfo	estimateCompressedColGroupSize(int[] colIndexes)
abstract CompressedSizeInfo	estimateCompressedColGroupSize(UncompressedBitmap ubm)
protected static long	getDDCSize(int numVals, int numRows, int numCols) Estimates the number of bytes needed to encode this column group in DDC1 or DDC2 format.
int	getNumRows()
protected static long	getOLESize(int numVals, float numOffs, int numSeqs, int numCols) Estimates the number of bytes needed to encode this column group in OLE format.
protected static long	getRLESize(int numVals, int numRuns, int numCols) Estimates the number of bytes needed to encode this column group in RLE encoding format.

Methods inherited from class java.lang.Object

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

Field Detail

_data

protected MatrixBlock _data

_numRows

protected final int _numRows

Constructor Detail

CompressedSizeEstimator

public CompressedSizeEstimator(MatrixBlock data)

Method Detail

getNumRows

public int getNumRows()

estimateCompressedColGroupSize

public abstract CompressedSizeInfo estimateCompressedColGroupSize(int[] colIndexes)

estimateCompressedColGroupSize

public abstract CompressedSizeInfo estimateCompressedColGroupSize(UncompressedBitmap ubm)

computeSizeEstimationFactors

protected CompressedSizeEstimator.SizeEstimationFactors computeSizeEstimationFactors(UncompressedBitmap ubm,
                                                                                     boolean inclRLE)

getRLESize

protected static long getRLESize(int numVals,
                                 int numRuns,
                                 int numCols)

Estimates the number of bytes needed to encode this column group in RLE encoding format.

Parameters:

numVals - number of value tuples

numRuns - number of runs

numCols - number of columns

Returns:

number of bytes to encode column group in RLE format

getOLESize

protected static long getOLESize(int numVals,
                                 float numOffs,
                                 int numSeqs,
                                 int numCols)

Estimates the number of bytes needed to encode this column group in OLE format.

Parameters:

numVals - number of value tuples

numOffs - number of offsets

numSeqs - number of segment headers

numCols - number of columns

Returns:

number of bytes to encode column group in RLE format

getDDCSize

protected static long getDDCSize(int numVals,
                                 int numRows,
                                 int numCols)

Estimates the number of bytes needed to encode this column group in DDC1 or DDC2 format.

Parameters:

numVals - number of value tuples

numRows - number of rows

numCols - number of columns

Returns:

number of bytes to encode column group in RLE format