blob: 97aa57fa5cd8dce44de6bec40cea27fbab7db146 [file] [log] [blame]
/*
* 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.sysds.hops.estim;
import org.apache.commons.lang.NotImplementedException;
import org.apache.sysds.hops.OptimizerUtils;
import org.apache.sysds.runtime.matrix.data.MatrixBlock;
import org.apache.sysds.runtime.meta.DataCharacteristics;
import org.apache.sysds.runtime.meta.MatrixCharacteristics;
/**
* Basic average case estimator for matrix sparsity:
* sp = Math.min(1, sp1 * k) * Math.min(1, sp2 * k).
*
* Note: for outer-products (i.e., k=1) this worst-case
* estimate is equivalent to the average case estimate and
* the exact output sparsity.
*/
public class EstimatorBasicWorst extends SparsityEstimator
{
@Override
public DataCharacteristics estim(MMNode root) {
if (!root.getLeft().isLeaf())
estim(root.getLeft()); // obtain synopsis
if (root.getRight()!=null && !root.getRight().isLeaf())
estim(root.getRight()); // obtain synopsis
DataCharacteristics mc1 = !root.getLeft().isLeaf() ?
estim(root.getLeft()) : root.getLeft().getDataCharacteristics();
DataCharacteristics mc2 = root.getRight()==null ? null :
!root.getRight().isLeaf() ? estim(root.getRight()) :
root.getRight().getDataCharacteristics();
return root.setDataCharacteristics(
estimIntern(mc1, mc2, root.getOp()));
}
@Override
public double estim(MatrixBlock m1, MatrixBlock m2) {
return estim(m1, m2, OpCode.MM);
}
@Override
public double estim(MatrixBlock m1, MatrixBlock m2, OpCode op) {
return estimIntern(m1.getDataCharacteristics(), m2.getDataCharacteristics(), op).getSparsity();
}
@Override
public double estim(MatrixBlock m, OpCode op) {
return estimIntern(m.getDataCharacteristics(), null, op).getSparsity();
}
private DataCharacteristics estimIntern(DataCharacteristics mc1, DataCharacteristics mc2, OpCode op) {
switch (op) {
case MM:
return new MatrixCharacteristics(mc1.getRows(), mc2.getCols(),
OptimizerUtils.getMatMultNnz(mc1.getSparsity(), mc2.getSparsity(),
mc1.getRows(), mc1.getCols(), mc2.getCols(), true));
case MULT:
return new MatrixCharacteristics(mc1.getRows(), mc1.getCols(),
OptimizerUtils.getNnz(mc1.getRows(), mc1.getCols(),
Math.min(mc1.getSparsity(), mc2.getSparsity())));
case PLUS:
return new MatrixCharacteristics(mc1.getRows(), mc1.getCols(),
OptimizerUtils.getNnz(mc1.getRows(), mc1.getCols(),
Math.min(mc1.getSparsity() + mc2.getSparsity(), 1)));
case EQZERO:
case DIAG:
case CBIND:
case RBIND:
case NEQZERO:
case TRANS:
case RESHAPE:
return estimExactMetaData(mc1, mc2, op);
default:
throw new NotImplementedException();
}
}
}