[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
courses:mapreduce-tutorial:step-31 [2012/02/06 08:40]
straka
courses:mapreduce-tutorial:step-31 [2012/02/06 14:52] (current)
dusek
Line 32: Line 32:
 It is crucial that all the mappers run simultaneously. This can be achieved using the ''/net/projects/hadoop/bin/compute-splitsize'' script: for given Hadoop input and requested number of mappers, it computes the appropriate splitsize. It is crucial that all the mappers run simultaneously. This can be achieved using the ''/net/projects/hadoop/bin/compute-splitsize'' script: for given Hadoop input and requested number of mappers, it computes the appropriate splitsize.
  
-When the computation finishes, only one of the mappers should print the results, as all of them have the same results. For simplicity, the ''cooperate'' method has ''boolean shouldWrite'' argument, which is set in exactly one mapper.+When the computation finishes, only one of the mappers should print the results, as all of them have the same results. For simplicity, the ''cooperate'' method has ''boolean writeResults'' argument, which is set in exactly one mapper.
  
 ===== Example ===== ===== Example =====
-This example reads the keys of ''/net/projects/hadoop/examples/inputs/numbers-small'', computes the sum of all the keys and print it:+This example reads the keys of ''/net/projects/hadoop/examples/inputs/numbers-small'', computes the sum of all the keys and prints it:
 <code java Sum.java> <code java Sum.java>
 +import java.io.IOException;
 +
 +import org.apache.hadoop.conf.*;
 +import org.apache.hadoop.fs.Path;
 +import org.apache.hadoop.io.*;
 import org.apache.hadoop.mapreduce.*; import org.apache.hadoop.mapreduce.*;
 import org.apache.hadoop.mapreduce.lib.allreduce.*; import org.apache.hadoop.mapreduce.lib.allreduce.*;
Line 102: Line 107:
 You can run the example locally using: You can run the example locally using:
   wget --no-check-certificate 'https://wiki.ufal.ms.mff.cuni.cz/_export/code/courses:mapreduce-tutorial:step-31?codeblock=0' -O Sum.java   wget --no-check-certificate 'https://wiki.ufal.ms.mff.cuni.cz/_export/code/courses:mapreduce-tutorial:step-31?codeblock=0' -O Sum.java
-  make -f /net/projects/hadoop/java/Makefile Sum.java+  make -f /net/projects/hadoop/java/Makefile Sum.jar
   rm -rf step-31-out; /net/projects/hadoop/bin/hadoop Sum.jar /net/projects/hadoop/examples/inputs/numbers-small step-31-out   rm -rf step-31-out; /net/projects/hadoop/bin/hadoop Sum.jar /net/projects/hadoop/examples/inputs/numbers-small step-31-out
   less step-31-out/part-*   less step-31-out/part-*
    
-To run on a cluster with //C// machines using //C// mappers+To run on a cluster using specified number of machines: 
-  rm -rf step-31-out; /net/projects/hadoop/bin/hadoop Sum.jar -c C `/net/projects/hadoop/bin/compute-splitsize /net/projects/hadoop/examples/inputs/numbers-small C` /net/projects/hadoop/examples/inputs/numbers-small step-31-out+  rm -rf step-31-out; M=#of_machines; INPUT=/net/projects/hadoop/examples/inputs/numbers-small; /net/projects/hadoop/bin/hadoop Sum.jar -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out
   less step-31-out/part-*   less step-31-out/part-*
  
Line 122: Line 127:
   # NOW VIEW THE FILE   # NOW VIEW THE FILE
   # $EDITOR Statistics.java   # $EDITOR Statistics.java
-  make -f /net/projects/hadoop/java/Makefile Statistics.java +  make -f /net/projects/hadoop/java/Makefile Statistics.jar 
-  rm -rf step-31-out; /net/projects/hadoop/bin/hadoop Statistics.jar -c C `/net/projects/hadoop/bin/compute-splitsize /net/projects/hadoop/examples/inputs/numbers-small C` /net/projects/hadoop/examples/inputs/numbers-small step-31-out+  rm -rf step-31-out; M=#of_machines; INPUT=/net/projects/hadoop/examples/inputs/numbers-small; /net/projects/hadoop/bin/hadoop Statistics.jar -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out
   less step-31-out/part-*   less step-31-out/part-*
  
Line 141: Line 146:
   # NOW VIEW THE FILE   # NOW VIEW THE FILE
   # $EDITOR Median.java   # $EDITOR Median.java
-  make -f /net/projects/hadoop/java/Makefile Median.java +  make -f /net/projects/hadoop/java/Makefile Median.jar 
-  rm -rf step-31-out; /net/projects/hadoop/bin/hadoop Median.jar -c C `/net/projects/hadoop/bin/compute-splitsize /net/projects/hadoop/examples/inputs/numbers-small C` /net/projects/hadoop/examples/inputs/numbers-small step-31-out+  rm -rf step-31-out; M=#of_machines; INPUT=/net/projects/hadoop/examples/inputs/numbers-small; /net/projects/hadoop/bin/hadoop Median.jar -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out
   less step-31-out/part-*   less step-31-out/part-*
  
Line 149: Line 154:
 ===== Exercise 3 ===== ===== Exercise 3 =====
  
-Implement an AllReduce job on ''/net/projects/hadoop/examples/inputs/numbers-small'', which computes+Implement an AllReduce job on ''/net/projects/hadoop/examples/inputs/points-small'', which implements the [[http://en.wikipedia.org/wiki/K-means_clustering#Standard_algorithm|K-means clustering algorithm]]. See [[.:step-15|K-means clustering exercise]] for description of input data.
  
-You can download the template {{:courses:mapreduce-tutorial:step-31-exercise3.txt|Median.java}} and execute it using: +You can download the template {{:courses:mapreduce-tutorial:step-31-exercise3.txt|KMeans.java}}. This template uses two Hadoop properties: 
-  wget --no-check-certificate 'https://wiki.ufal.ms.mff.cuni.cz/_media/courses:mapreduce-tutorial:step-31-exercise3.txt' -O KMeans.java.java+  * ''clusters.num'' -- number of clusters 
 +  * ''clusters.file'' -- file where to read the initial clusters from 
 +You can download and compile it using: 
 +  wget --no-check-certificate 'https://wiki.ufal.ms.mff.cuni.cz/_media/courses:mapreduce-tutorial:step-31-exercise3.txt' -O KMeans.java
   # NOW VIEW THE FILE   # NOW VIEW THE FILE
-  # $EDITOR KMeans.java.java +  # $EDITOR KMeans.java 
-  make -f /net/projects/hadoop/java/Makefile KMeans.java.java +  make -f /net/projects/hadoop/java/Makefile KMeans.jar 
-  rm -rf step-31-out; /net/projects/hadoop/bin/hadoop KMeans.java.jar -c `/net/projects/hadoop/bin/compute-splitsize /net/projects/hadoop/examples/inputs/numbers-small C` /net/projects/hadoop/examples/inputs/numbers-small step-31-out +You can run it using specified number of machines on the following input data: 
-  less step-31-out/part-*+  * ''/net/projects/hadoop/examples/inputs/points-small'': 
 +<code>M=#of_machines; K=50; INPUT=/net/projects/hadoop/examples/inputs/points-small/points.txt 
 +rm -rf step-31-out; /net/projects/hadoop/bin/hadoop KMeans.jar -Dclusters.num=$K -Dclusters.file=$INPUT -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out</code> 
 +  * ''/net/projects/hadoop/examples/inputs/points-medium'': 
 +<code>M=#of_machines; K=100; INPUT=/net/projects/hadoop/examples/inputs/points-medium/points.txt 
 +rm -rf step-31-out; /net/projects/hadoop/bin/hadoop KMeans.jar -Dclusters.num=$K -Dclusters.file=$INPUT -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out</code> 
 +  * ''/net/projects/hadoop/examples/inputs/points-large'': 
 +<code>M=#of_machines; K=200; INPUT=/net/projects/hadoop/examples/inputs/points-large/points.txt 
 +rm -rf step-31-out/net/projects/hadoop/bin/hadoop KMeans.jar -Dclusters.num=$K -Dclusters.file=$INPUT -c $M `/net/projects/hadoop/bin/compute-splitsize $INPUT $M` $INPUT step-31-out</code>
  
 Solution: {{:courses:mapreduce-tutorial:step-31-solution3.txt|KMeans.java}}. Solution: {{:courses:mapreduce-tutorial:step-31-solution3.txt|KMeans.java}}.
  

[ Back to the navigation ] [ Back to the content ]