org.apache.sysml.lops.WeightedDivMM Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of systemml Show documentation
Show all versions of systemml Show documentation
Declarative Machine Learning
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.apache.sysml.lops;
import org.apache.sysml.lops.LopProperties.ExecLocation;
import org.apache.sysml.lops.LopProperties.ExecType;
import org.apache.sysml.lops.compile.JobType;
import org.apache.sysml.parser.Expression.DataType;
import org.apache.sysml.parser.Expression.ValueType;
import org.apache.sysml.runtime.matrix.MatrixCharacteristics;
public class WeightedDivMM extends Lop
{
public static final String OPCODE = "mapwdivmm";
public static final String OPCODE_CP = "wdivmm";
private int _numThreads = 1;
public enum WDivMMType {
DIV_LEFT, //t(t(U) %*% (W / U%*%t(V)))
DIV_RIGHT, //(W / U%*%t(V)) %*% V
DIV_LEFT_EPS, //t(t(U) %*% (W / (U%*%t(V) + x)))
DIV_RIGHT_EPS, //(W / (U%*%t(V) + x)) %*% V
MULT_BASIC, //(W * U%*%t(V))
MULT_LEFT, //t(t(U) %*% (W * U%*%t(V)))
MULT_RIGHT, //(W * U%*%t(V)) %*% V
MULT_MINUS_LEFT, //t(t(U) %*% ((X!=0) * (U%*%t(V) - X)))
MULT_MINUS_RIGHT, //((X!=0) * (U%*%t(V) - X)) %*% V
MULT_MINUS_4_LEFT, //t(t(U) %*% (W * (U%*%t(V) - X)))
MULT_MINUS_4_RIGHT; //(W * (U%*%t(V) - X)) %*% V
public boolean isBasic(){
return (this == MULT_BASIC);
}
public boolean isLeft() {
return (this == DIV_LEFT || this == DIV_LEFT_EPS || this == MULT_LEFT
|| this == MULT_MINUS_LEFT || this == MULT_MINUS_4_LEFT);
}
public boolean isRight() {
return !(isLeft() || isBasic());
}
public boolean isMult() {
return (this == MULT_LEFT || this == MULT_RIGHT || this == MULT_MINUS_LEFT || this == MULT_MINUS_RIGHT
|| this == MULT_MINUS_4_LEFT || this == MULT_MINUS_4_RIGHT);
}
public boolean isMinus(){
return (this == MULT_MINUS_LEFT || this == MULT_MINUS_RIGHT
|| this == MULT_MINUS_4_LEFT || this == MULT_MINUS_4_RIGHT);
}
public boolean hasFourInputs() {
return (this == MULT_MINUS_4_LEFT || this == MULT_MINUS_4_RIGHT
|| this == DIV_LEFT_EPS || this == DIV_RIGHT_EPS);
}
public boolean hasScalar() {
return (this == DIV_LEFT_EPS || this == DIV_RIGHT_EPS);
}
public MatrixCharacteristics computeOutputCharacteristics(long Xrlen, long Xclen, long rank) {
if( isBasic() )
return new MatrixCharacteristics( Xrlen, Xclen, -1, -1);
else
return new MatrixCharacteristics(isLeft()?Xclen:Xrlen, rank, -1, -1);
}
}
private WDivMMType _weightsType = null;
public WeightedDivMM(Lop input1, Lop input2, Lop input3, Lop input4, DataType dt, ValueType vt, WDivMMType wt, ExecType et)
throws LopsException
{
super(Lop.Type.WeightedDivMM, dt, vt);
addInput(input1); //W
addInput(input2); //U
addInput(input3); //V
addInput(input4); //X (optional)
input1.addOutput(this);
input2.addOutput(this);
input3.addOutput(this);
input4.addOutput(this);
_weightsType = wt;
setupLopProperties(et);
}
private void setupLopProperties( ExecType et )
{
if( et == ExecType.MR )
{
//setup MR parameters
boolean breaksAlignment = true;
boolean aligner = false;
boolean definesMRJob = false;
lps.addCompatibility(JobType.GMR);
lps.addCompatibility(JobType.DATAGEN);
lps.setProperties( inputs, ExecType.MR, ExecLocation.Map, breaksAlignment, aligner, definesMRJob );
}
else //Spark/CP
{
//setup Spark parameters
boolean breaksAlignment = false;
boolean aligner = false;
boolean definesMRJob = false;
lps.addCompatibility(JobType.INVALID);
lps.setProperties( inputs, et, ExecLocation.ControlProgram, breaksAlignment, aligner, definesMRJob );
}
}
public String toString() {
return "Operation = WeightedDivMM";
}
/* MR instruction generation */
@Override
public String getInstructions(int input1, int input2, int input3, int input4, int output)
{
return getInstructions(
String.valueOf(input1),
String.valueOf(input2),
String.valueOf(input3),
String.valueOf(input4),
String.valueOf(output));
}
/* CP/SPARK instruction generation */
@Override
public String getInstructions(String input1, String input2, String input3, String input4, String output)
{
StringBuilder sb = new StringBuilder();
final ExecType et = getExecType();
sb.append(et);
sb.append(Lop.OPERAND_DELIMITOR);
if( et == ExecType.CP )
sb.append(OPCODE_CP);
else
sb.append(OPCODE);
sb.append(Lop.OPERAND_DELIMITOR);
sb.append( getInputs().get(0).prepInputOperand(input1));
sb.append(Lop.OPERAND_DELIMITOR);
sb.append( getInputs().get(1).prepInputOperand(input2));
sb.append(Lop.OPERAND_DELIMITOR);
sb.append( getInputs().get(2).prepInputOperand(input3));
sb.append(Lop.OPERAND_DELIMITOR);
if ( (et == ExecType.MR) && (getInputs().get(3).getDataType() == DataType.SCALAR) ) {
sb.append( getInputs().get(3).prepScalarInputOperand(et));
}
else {
sb.append( getInputs().get(3).prepInputOperand(input4));
}
sb.append(Lop.OPERAND_DELIMITOR);
sb.append( prepOutputOperand(output));
sb.append(Lop.OPERAND_DELIMITOR);
sb.append(_weightsType);
//append degree of parallelism
if( et == ExecType.CP ) {
sb.append( OPERAND_DELIMITOR );
sb.append( _numThreads );
}
return sb.toString();
}
@Override
public boolean usesDistributedCache()
{
if( getExecType()==ExecType.MR )
return true;
else
return false;
}
@Override
public int[] distributedCacheInputIndex()
{
if( getExecType()==ExecType.MR )
return new int[]{2,3};
else
return new int[]{-1};
}
public void setNumThreads(int k) {
_numThreads = k;
}
}