Distributed Processing

This mode assumes that the data set is split into nblocks blocks across computation nodes.

Algorithm Parameters

The low order moments algorithm in the distributed processing mode has the following parameters:

Parameter	Default Value	Description
computeStep	Not applicable	The parameter required to initialize the algorithm. Can be: step1Local - the first step, performed on local nodes step2Master - the second step, performed on a master node
algorithmFPType	double	The floating-point type that the algorithm uses for intermediate computations. Can be float or double.
method	defaultDense	Available methods for computation of low order moments: defaultDense - default performance-oriented method singlePassDense - implementation of the single-pass algorithm proposed by D.H.D. West sumDense - implementation of the algorithm in the cases where the basic statistics associated with the numeric table are pre-computed sums; returns an error if pre-computed sums are not defined fastCSR - performance-oriented method for CSR numeric tables singlePassCSR - implementation of the single-pass algorithm proposed by D.H.D. West; optimized for CSR numeric tables sumCSR - implementation of the algorithm in the cases where the basic statistics associated with the numeric table are pre-computed sums; optimized for CSR numeric tables; returns an error if pre-computed sums are not defined

Parameter

Default Value

Description

computeStep

Not applicable

The parameter required to initialize the algorithm. Can be:

step1Local - the first step, performed on local nodes
step2Master - the second step, performed on a master node

algorithmFPType

double

The floating-point type that the algorithm uses for intermediate computations. Can be float or double.

method

defaultDense

Available methods for computation of low order moments:

defaultDense - default performance-oriented method
singlePassDense - implementation of the single-pass algorithm proposed by D.H.D. West
sumDense - implementation of the algorithm in the cases where the basic statistics associated with the numeric table are pre-computed sums; returns an error if pre-computed sums are not defined
fastCSR - performance-oriented method for CSR numeric tables
singlePassCSR - implementation of the single-pass algorithm proposed by D.H.D. West; optimized for CSR numeric tables
sumCSR - implementation of the algorithm in the cases where the basic statistics associated with the numeric table are pre-computed sums; optimized for CSR numeric tables; returns an error if pre-computed sums are not defined

Computation of low order moments follows the general schema described in Algorithms:

Step 1 - on Local Nodes

In this step, the low order moments algorithm accepts the input described below. Pass the Input ID as a parameter to the methods that provide input for your algorithm. For more details, see Algorithms.

Input ID	Input
data	Pointer to the `n`_i x `p` numeric table that represents the `i`-th data block on the local node. While the input for defaultDense, singlePassDense, or sumDense method can be an object of any class derived from NumericTable, the input for fastCSR, singlePassCSR, or sumCSR method can only be an object of the CSRNumericTable class.

Input ID

Input

data

Pointer to the n_i x p numeric table that represents the i-th data block on the local node.

While the input for defaultDense, singlePassDense, or sumDense method can be an object of any class derived from NumericTable, the input for fastCSR, singlePassCSR, or sumCSR method can only be an object of the CSRNumericTable class.

In this step, the low order moments algorithm calculates the results described below. Pass the Result ID as a parameter to the methods that access the results of your algorithm. For more details, see Algorithms.

Result ID	Result
nObservations	Pointer 1 x 1 numeric table that contains the number of observations processed so far on the local node. By default, this result is an object of the HomogenNumericTable class, but you can define the result as an object of any class derived from NumericTable except CSRNumericTable.
Partial characteristics computed so far on the local node, each in a 1 x `p` numeric table. By default, each table is an object of the HomogenNumericTable class, but you can define the tables as objects of any class derived from NumericTable except PackedSymmetricMatrix, PackedTriangularMatrix, and CSRNumericTable.
partialMinimum	Partial minimums.
partialMaximum	Partial maximums.
partialSum	Partial sums.
partialSumSquares	Partial sums of squares.
partialSumSquaresCentered	Partial sums of squared differences from the means.

Step 2 - on Master Node

Input ID	Input
partialResults	A collection that contains numeric tables with partial results computed in Step 1 on local nodes (six numeric tables from each local node). These numeric tables can be objects of any class derived from the NumericTable class except PackedSymmetricMatrix and PackedTriangularMatrix.

In this step, the low order moments algorithm calculates the results described in the following table. Pass the Result ID as a parameter to the methods that access the results of your algorithm. For more details, see Algorithms.

Each result is a pointer to the 1 x p numeric table that contains characteristics for each feature in the data set. By default, the tables are objects of the HomogenNumericTable class, but you can define each table as an object of any class derived from NumericTable except PackedSymmetricMatrix, PackedTriangularMatrix, and CSRNumericTable.

Result ID	Characteristic
minimum	Minimums.
maximum	Maximums.
sum	Sums.
sumSquares	Sums of squares.
sumSquaresCentered	Sums of squared differences from the means.
mean	Estimates for the means.
secondOrderRawMoment	Estimates for the second order raw moments.
variance	Estimates for the variances.
standardDeviation	Estimates for the standard deviations.
variation	Estimates for the variations.

Optimization Notice
Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice. Notice revision #20110804

Optimization Notice

Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804

Examples

C++:

Java*:

Distributed Processing

Algorithm Parameters

Step 1 - on Local Nodes

Step 2 - on Master Node

Examples

See Also