mapred
Class MapRedEpicReducerEvaluate
java.lang.Object
mapred.MapRedEpicReducerEvaluate
public class MapRedEpicReducerEvaluate
- extends java.lang.Object
Implementation of EPiC's approach with a slight difference.
The difference from MapRedEpic
is that the evaluation of the indicator polynomial is switched
from the Mappers to the Reducer. Therefore, instead of sending only the final results corresponding to
the subsets, all monomials are sent from Mappers to Reducer, which increases the communication among nodes.
This is a "naive" implementation of EPiC.
This implementation supports all kinds of counting (conjunctive, disjunctive, range, GF(2) arithmetized).
- Author:
- vohuudtr
Method Summary |
static void |
setup(org.apache.hadoop.mapred.JobConf conf)
Initializes the class based on provided configuration parameters. |
Methods inherited from class java.lang.Object |
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
MapRedEpicReducerEvaluate
public MapRedEpicReducerEvaluate()
setup
public static void setup(org.apache.hadoop.mapred.JobConf conf)
- Initializes the class based on provided configuration parameters.
The initialization comprises setting job's name, assigning Mapper
and Reducer class as well as their input and output key-value classes.
- Parameters:
conf
- configuration parameters.