Package eu.stratosphere.nephele.jobgraph

Examples of eu.stratosphere.nephele.jobgraph.JobInputVertex

110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
    JobGraph jobGraph = new JobGraph("CompensatableDanglingPageRank");
   
    // --------------- the inputs ---------------------

    // page rank input
    JobInputVertex pageWithRankInput = JobGraphUtils.createInput(new ImprovedDanglingPageRankInputFormat(),
      pageWithRankInputPath, "DanglingPageWithRankInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig pageWithRankInputConfig = new TaskConfig(pageWithRankInput.getConfiguration());
    pageWithRankInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    pageWithRankInputConfig.setOutputComparator(fieldZeroComparator, 0);
    pageWithRankInputConfig.setOutputSerializer(recSerializer);
    pageWithRankInputConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));

    // edges as adjacency list
    JobInputVertex adjacencyListInput = JobGraphUtils.createInput(new ImprovedAdjacencyListInputFormat(),
      adjacencyListInputPath, "AdjancencyListInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig adjacencyListInputConfig = new TaskConfig(adjacencyListInput.getConfiguration());
    adjacencyListInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    adjacencyListInputConfig.setOutputSerializer(recSerializer);
    adjacencyListInputConfig.setOutputComparator(fieldZeroComparator, 0);

    // --------------- the head ---------------------
    JobTaskVertex head = JobGraphUtils.createTask(IterationHeadPactTask.class, "IterationHead", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig headConfig = new TaskConfig(head.getConfiguration());
    headConfig.setIterationId(ITERATION_ID);
   
    // initial input / partial solution
    headConfig.addInputToGroup(0);
    headConfig.setIterationHeadPartialSolutionOrWorksetInputIndex(0);
    headConfig.setInputSerializer(recSerializer, 0);
    headConfig.setInputComparator(fieldZeroComparator, 0);
    headConfig.setInputLocalStrategy(0, LocalStrategy.SORT);
    headConfig.setMemoryInput(0, minorConsumer * JobGraphUtils.MEGABYTE);
    headConfig.setFilehandlesInput(0, NUM_FILE_HANDLES_PER_SORT);
    headConfig.setSpillingThresholdInput(0, SORT_SPILL_THRESHOLD);
   
    // back channel / iterations
    headConfig.setBackChannelMemory(minorConsumer * JobGraphUtils.MEGABYTE);
   
    // output into iteration
    headConfig.setOutputSerializer(recSerializer);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
   
    // final output
    TaskConfig headFinalOutConfig = new TaskConfig(new Configuration());
    headFinalOutConfig.setOutputSerializer(recSerializer);
    headFinalOutConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.setIterationHeadFinalOutputConfig(headFinalOutConfig);
   
    // the sync
    headConfig.setIterationHeadIndexOfSyncOutput(3);
    headConfig.setNumberOfIterations(numIterations);
   
    // the driver
    headConfig.setDriver(CollectorMapDriver.class);
    headConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
    headConfig.setStubWrapper(new UserCodeClassWrapper<CompensatingMap>(CompensatingMap.class));
    headConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    headConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    headConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    headConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
    headConfig.addIterationAggregator(CompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);

    // --------------- the join ---------------------
   
    JobTaskVertex intermediate = JobGraphUtils.createTask(IterationIntermediatePactTask.class,
      "IterationIntermediate", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig intermediateConfig = new TaskConfig(intermediate.getConfiguration());
    intermediateConfig.setIterationId(ITERATION_ID);
//    intermediateConfig.setDriver(RepeatableHashjoinMatchDriverWithCachedBuildside.class);
    intermediateConfig.setDriver(BuildSecondCachedMatchDriver.class);
    intermediateConfig.setDriverStrategy(DriverStrategy.HYBRIDHASH_BUILD_SECOND);
    intermediateConfig.setMemoryDriver(matchMemory * JobGraphUtils.MEGABYTE);
    intermediateConfig.addInputToGroup(0);
    intermediateConfig.addInputToGroup(1);
    intermediateConfig.setInputSerializer(recSerializer, 0);
    intermediateConfig.setInputSerializer(recSerializer, 1);
    intermediateConfig.setDriverComparator(fieldZeroComparator, 0);
    intermediateConfig.setDriverComparator(fieldZeroComparator, 1);
    intermediateConfig.setDriverPairComparator(pairComparatorFactory);
   
    intermediateConfig.setOutputSerializer(recSerializer);
    intermediateConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    intermediateConfig.setOutputComparator(fieldZeroComparator, 0);
   
    intermediateConfig.setStubWrapper(new UserCodeClassWrapper<CompensatableDotProductMatch>(CompensatableDotProductMatch.class));
    intermediateConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    intermediateConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    intermediateConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    intermediateConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));

    // ---------------- the tail (co group) --------------------
   
    JobTaskVertex tail = JobGraphUtils.createTask(IterationTailPactTask.class, "IterationTail", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig tailConfig = new TaskConfig(tail.getConfiguration());
    tailConfig.setIterationId(ITERATION_ID);
        tailConfig.setIsWorksetUpdate();
    // TODO we need to combine!
   
    // inputs and driver
    tailConfig.setDriver(CoGroupDriver.class);
    tailConfig.setDriverStrategy(DriverStrategy.CO_GROUP);
    tailConfig.addInputToGroup(0);
    tailConfig.addInputToGroup(1);
    tailConfig.setInputSerializer(recSerializer, 0);
    tailConfig.setInputSerializer(recSerializer, 1);
    tailConfig.setDriverComparator(fieldZeroComparator, 0);
    tailConfig.setDriverComparator(fieldZeroComparator, 1);
    tailConfig.setDriverPairComparator(pairComparatorFactory);
    tailConfig.setInputAsynchronouslyMaterialized(0, true);
    tailConfig.setInputMaterializationMemory(0, minorConsumer * JobGraphUtils.MEGABYTE);
    tailConfig.setInputLocalStrategy(1, LocalStrategy.SORT);
    tailConfig.setInputComparator(fieldZeroComparator, 1);
    tailConfig.setMemoryInput(1, coGroupSortMemory * JobGraphUtils.MEGABYTE);
    tailConfig.setFilehandlesInput(1, NUM_FILE_HANDLES_PER_SORT);
    tailConfig.setSpillingThresholdInput(1, SORT_SPILL_THRESHOLD);
   
    // output
    tailConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    tailConfig.setOutputSerializer(recSerializer);
   
    // the stub
    tailConfig.setStubWrapper(new UserCodeClassWrapper<CompensatableDotProductCoGroup>(CompensatableDotProductCoGroup.class));
    tailConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    tailConfig.setStubParameter("pageRank.numDanglingVertices", String.valueOf(numDanglingVertices));
    tailConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    tailConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    tailConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
   
    // --------------- the output ---------------------

    JobOutputVertex output = JobGraphUtils.createFileOutput(jobGraph, "FinalOutput", degreeOfParallelism,
      numSubTasksPerInstance);
    TaskConfig outputConfig = new TaskConfig(output.getConfiguration());
    outputConfig.addInputToGroup(0);
    outputConfig.setInputSerializer(recSerializer, 0);
    outputConfig.setStubWrapper(new UserCodeClassWrapper<PageWithRankOutFormat>(PageWithRankOutFormat.class));
    outputConfig.setStubParameter(FileOutputFormat.FILE_PARAMETER_KEY, outputPath);
   
    // --------------- the auxiliaries ---------------------
   
    JobOutputVertex fakeTailOutput = JobGraphUtils.createFakeOutput(jobGraph, "FakeTailOutput",
      degreeOfParallelism, numSubTasksPerInstance);

    JobOutputVertex sync = JobGraphUtils.createSync(jobGraph, degreeOfParallelism);
    TaskConfig syncConfig = new TaskConfig(sync.getConfiguration());
    syncConfig.setNumberOfIterations(numIterations);
    syncConfig.addIterationAggregator(CompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);
    syncConfig.setConvergenceCriterion(CompensatableDotProductCoGroup.AGGREGATOR_NAME, DiffL1NormConvergenceCriterion.class);
    syncConfig.setIterationId(ITERATION_ID);
   
    // --------------- the wiring ---------------------

    JobGraphUtils.connect(pageWithRankInput, head, ChannelType.NETWORK, DistributionPattern.BIPARTITE);

    JobGraphUtils.connect(head, intermediate, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    intermediateConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
   
    JobGraphUtils.connect(adjacencyListInput, intermediate, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
   
    JobGraphUtils.connect(head, tail, ChannelType.NETWORK, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(intermediate, tail, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(1, degreeOfParallelism);

    JobGraphUtils.connect(head, output, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(tail, fakeTailOutput, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(head, sync, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    fakeTailOutput.setVertexToShareInstancesWith(tail);
    tail.setVertexToShareInstancesWith(head);
    pageWithRankInput.setVertexToShareInstancesWith(head);
    adjacencyListInput.setVertexToShareInstancesWith(head);
    intermediate.setVertexToShareInstancesWith(head);
    output.setVertexToShareInstancesWith(head);
    sync.setVertexToShareInstancesWith(head);

    return jobGraph;
View Full Code Here
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
    assignDriverResources(node, config);
    return vertex;
  }

  private JobInputVertex createDataSourceVertex(SourcePlanNode node) throws CompilerException {
    final JobInputVertex vertex = new JobInputVertex(node.getNodeName(), this.jobGraph);
    final TaskConfig config = new TaskConfig(vertex.getConfiguration());
   
    // set task class
    @SuppressWarnings("unchecked")
    final Class<AbstractInputTask<?>> clazz = (Class<AbstractInputTask<?>>) (Class<?>) DataSourceTask.class;
    vertex.setInputClass(clazz);

    // set user code
    config.setStubWrapper(node.getPactContract().getUserCodeWrapper());
    config.setStubParameters(node.getPactContract().getParameters());
   
View Full Code Here
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
    // --------------------------------------------------------------------------------------------------------------
    // 1. VERTICES
    // --------------------------------------------------------------------------------------------------------------

    // - input -----------------------------------------------------------------------------------------------------
    JobInputVertex input = JobGraphUtils.createInput(
      new PointInFormat(), inputPath, "Input", jobGraph, numSubTasks, numSubTasks);
    TaskConfig inputConfig = new TaskConfig(input.getConfiguration());
    {
      inputConfig.setOutputSerializer(serializer);
      inputConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    }

    // - head ------------------------------------------------------------------------------------------------------
    JobTaskVertex head = JobGraphUtils.createTask(
      IterationHeadPactTask.class, "Iteration Head", jobGraph, numSubTasks, numSubTasks);
    TaskConfig headConfig = new TaskConfig(head.getConfiguration());
    {
      headConfig.setIterationId(ITERATION_ID);

      // input to iteration head
      headConfig.addInputToGroup(0);
      headConfig.setInputSerializer(serializer, 0);
      headConfig.setInputLocalStrategy(0, LocalStrategy.NONE);
      headConfig.setIterationHeadPartialSolutionOrWorksetInputIndex(0);

      // output into iteration
      headConfig.setOutputSerializer(serializer);
      headConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
      headConfig.setOutputComparator(comparator, 0);

      // final output
      TaskConfig headFinalOutConfig = new TaskConfig(new Configuration());
      headFinalOutConfig.setOutputSerializer(serializer);
      headFinalOutConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      headConfig.setIterationHeadFinalOutputConfig(headFinalOutConfig);

      // the sync
      headConfig.setIterationHeadIndexOfSyncOutput(2);

      // driver
      headConfig.setDriver(CollectorMapDriver.class);
      headConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
      headConfig.setStubWrapper(new UserCodeClassWrapper<DummyMapper>(DummyMapper.class));

      // back channel
      headConfig.setBackChannelMemory(MEM_PER_CONSUMER * JobGraphUtils.MEGABYTE);
    }

    // - tail ------------------------------------------------------------------------------------------------------
    JobTaskVertex tail = JobGraphUtils.createTask(
      IterationTailPactTask.class, "Chained Iteration Tail", jobGraph, numSubTasks, numSubTasks);
    TaskConfig tailConfig = new TaskConfig(tail.getConfiguration());
    {
      tailConfig.setIterationId(ITERATION_ID);

      // inputs and driver
      tailConfig.addInputToGroup(0);
      tailConfig.setInputSerializer(serializer, 0);

      // output
      tailConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      tailConfig.setOutputSerializer(serializer);

      // the driver
      tailConfig.setDriver(GroupReduceDriver.class);
      tailConfig.setDriverStrategy(DriverStrategy.SORTED_GROUP_REDUCE);
      tailConfig.setDriverComparator(comparator, 0);
      tailConfig.setStubWrapper(new UserCodeClassWrapper<DummyReducer>(DummyReducer.class));

      // chained mapper
      TaskConfig chainedMapperConfig = new TaskConfig(new Configuration());
      chainedMapperConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
      chainedMapperConfig.setStubWrapper(new UserCodeClassWrapper<IncrementCoordinatesMapper>(
        IncrementCoordinatesMapper.class));

      chainedMapperConfig.setInputLocalStrategy(0, LocalStrategy.NONE);
      chainedMapperConfig.setInputSerializer(serializer, 0);

      chainedMapperConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      chainedMapperConfig.setOutputSerializer(serializer);

      chainedMapperConfig.setIsWorksetUpdate();

      tailConfig.addChainedTask(ChainedCollectorMapDriver.class, chainedMapperConfig, "Chained ID Mapper");
    }

    // - output ----------------------------------------------------------------------------------------------------
    JobOutputVertex output = JobGraphUtils.createFileOutput(jobGraph, "Output", numSubTasks, numSubTasks);
    TaskConfig outputConfig = new TaskConfig(output.getConfiguration());
    {
      outputConfig.addInputToGroup(0);
      outputConfig.setInputSerializer(serializer, 0);

      outputConfig.setStubWrapper(new UserCodeClassWrapper<PointOutFormat>(PointOutFormat.class));
      outputConfig.setStubParameter(FileOutputFormat.FILE_PARAMETER_KEY, outputPath);
    }

    // - fake tail -------------------------------------------------------------------------------------------------
    JobOutputVertex fakeTail = JobGraphUtils.createFakeOutput(jobGraph, "Fake Tail", numSubTasks, numSubTasks);

    // - sync ------------------------------------------------------------------------------------------------------
    JobOutputVertex sync = JobGraphUtils.createSync(jobGraph, numSubTasks);
    TaskConfig syncConfig = new TaskConfig(sync.getConfiguration());
    syncConfig.setNumberOfIterations(maxIterations);
    syncConfig.setIterationId(ITERATION_ID);

    // --------------------------------------------------------------------------------------------------------------
    // 2. EDGES
    // --------------------------------------------------------------------------------------------------------------
    JobGraphUtils.connect(input, head, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(head, tail, ChannelType.IN_MEMORY, DistributionPattern.BIPARTITE);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, numSubTasks);

    JobGraphUtils.connect(head, output, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(head, sync, ChannelType.NETWORK, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(tail, fakeTail, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    // --------------------------------------------------------------------------------------------------------------
    // 3. INSTANCE SHARING
    // --------------------------------------------------------------------------------------------------------------
    input.setVertexToShareInstancesWith(head);

    tail.setVertexToShareInstancesWith(head);

    output.setVertexToShareInstancesWith(head);

View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
  }

  private static <T extends InputFormat<?,?>> JobInputVertex createInput(UserCodeWrapper<T> stub, String name, JobGraph graph,
      int degreeOfParallelism, int numSubTasksPerInstance)
  {
    JobInputVertex inputVertex = new JobInputVertex(name, graph);
   
    @SuppressWarnings("unchecked")
    Class<AbstractInputTask<?>> clazz = (Class<AbstractInputTask<?>>) (Class<?>) DataSourceTask.class;
    inputVertex.setInputClass(clazz);
   
    inputVertex.setNumberOfSubtasks(degreeOfParallelism);
    inputVertex.setNumberOfSubtasksPerInstance(numSubTasksPerInstance);
   
    TaskConfig inputConfig = new TaskConfig(inputVertex.getConfiguration());
    inputConfig.setStubWrapper(stub);
   
    return inputVertex;
  }
View Full Code Here
91
92
93
94
95
96
97
98
99
100
101
102
103
  // -------------------------------------------------------------------------------------------------------------

  private static JobInputVertex createPointsInput(JobGraph jobGraph, String pointsPath, int numSubTasks, TypeSerializerFactory<?> serializer) {
    @SuppressWarnings("unchecked")
    CsvInputFormat pointsInFormat = new CsvInputFormat('|', IntValue.class, DoubleValue.class, DoubleValue.class, DoubleValue.class);
    JobInputVertex pointsInput = JobGraphUtils.createInput(pointsInFormat, pointsPath, "[Points]", jobGraph, numSubTasks, numSubTasks);
    {
      TaskConfig taskConfig = new TaskConfig(pointsInput.getConfiguration());
      taskConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      taskConfig.setOutputSerializer(serializer);
     
      TaskConfig chainedMapper = new TaskConfig(new Configuration());
      chainedMapper.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
View Full Code Here
112
113
114
115
116
117
118
119
120
121
122
123
124
125
  }

  private static JobInputVertex createCentersInput(JobGraph jobGraph, String centersPath, int numSubTasks, TypeSerializerFactory<?> serializer) {
    @SuppressWarnings("unchecked")
    CsvInputFormat modelsInFormat = new CsvInputFormat('|', IntValue.class, DoubleValue.class, DoubleValue.class, DoubleValue.class);
    JobInputVertex modelsInput = JobGraphUtils.createInput(modelsInFormat, centersPath, "[Models]", jobGraph, numSubTasks, numSubTasks);

    {
      TaskConfig taskConfig = new TaskConfig(modelsInput.getConfiguration());
      taskConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      taskConfig.setOutputSerializer(serializer);

      TaskConfig chainedMapper = new TaskConfig(new Configuration());
      chainedMapper.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
View Full Code Here
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
    final TypeComparatorFactory<?> int0Comparator = new RecordComparatorFactory(new int[] { 0 }, new Class[] { IntValue.class });

    JobGraph jobGraph = new JobGraph("KMeans Iterative");

    // -- vertices ---------------------------------------------------------------------------------------------
    JobInputVertex points = createPointsInput(jobGraph, pointsPath, numSubTasks, serializer);
    JobInputVertex centers = createCentersInput(jobGraph, centersPath, numSubTasks, serializer);
   
    JobTaskVertex head = createIterationHead(jobGraph, numSubTasks, serializer);
    JobTaskVertex mapper = createMapper(jobGraph, numSubTasks, serializer, serializer, serializer, int0Comparator);
   
    JobTaskVertex reducer = createReducer(jobGraph, numSubTasks, serializer, int0Comparator, serializer);
   
    JobOutputVertex fakeTailOutput = JobGraphUtils.createFakeOutput(jobGraph, "FakeTailOutput", numSubTasks, numSubTasks);
   
    JobOutputVertex sync = createSync(jobGraph, numIterations, numSubTasks);
   
    JobOutputVertex output = createOutput(jobGraph, resultPath, numSubTasks, serializer);

    // -- edges ------------------------------------------------------------------------------------------------
    JobGraphUtils.connect(points, mapper, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    JobGraphUtils.connect(centers, head, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    JobGraphUtils.connect(head, mapper, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
    new TaskConfig(mapper.getConfiguration()).setBroadcastGateIterativeWithNumberOfEventsUntilInterrupt(0, numSubTasks);
    new TaskConfig(mapper.getConfiguration()).setInputCached(0, true);
    new TaskConfig(mapper.getConfiguration()).setInputMaterializationMemory(0, MEMORY_PER_CONSUMER * JobGraphUtils.MEGABYTE);

    JobGraphUtils.connect(mapper, reducer, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
    new TaskConfig(reducer.getConfiguration()).setGateIterativeWithNumberOfEventsUntilInterrupt(0, numSubTasks);
   
    JobGraphUtils.connect(reducer, fakeTailOutput, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    JobGraphUtils.connect(head, output, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    JobGraphUtils.connect(head, sync, ChannelType.NETWORK, DistributionPattern.BIPARTITE);

    // -- instance sharing -------------------------------------------------------------------------------------
    points.setVertexToShareInstancesWith(output);
    centers.setVertexToShareInstancesWith(output);
    head.setVertexToShareInstancesWith(output);
    mapper.setVertexToShareInstancesWith(output);
    reducer.setVertexToShareInstancesWith(output);
    fakeTailOutput.setVertexToShareInstancesWith(output);
    sync.setVertexToShareInstancesWith(output);
View Full Code Here
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
    JobGraph jobGraph = new JobGraph("CompensatableDanglingPageRank");
   
    // --------------- the inputs ---------------------

    // page rank input
    JobInputVertex pageWithRankInput = JobGraphUtils.createInput(new CustomImprovedDanglingPageRankInputFormat(),
      pageWithRankInputPath, "DanglingPageWithRankInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig pageWithRankInputConfig = new TaskConfig(pageWithRankInput.getConfiguration());
    pageWithRankInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    pageWithRankInputConfig.setOutputComparator(vertexWithRankAndDanglingComparator, 0);
    pageWithRankInputConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    pageWithRankInputConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));

    // edges as adjacency list
    JobInputVertex adjacencyListInput = JobGraphUtils.createInput(new CustomImprovedAdjacencyListInputFormat(),
      adjacencyListInputPath, "AdjancencyListInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig adjacencyListInputConfig = new TaskConfig(adjacencyListInput.getConfiguration());
    adjacencyListInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    adjacencyListInputConfig.setOutputSerializer(vertexWithAdjacencyListSerializer);
    adjacencyListInputConfig.setOutputComparator(vertexWithAdjacencyListComparator, 0);

    // --------------- the head ---------------------
    JobTaskVertex head = JobGraphUtils.createTask(IterationHeadPactTask.class, "IterationHead", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig headConfig = new TaskConfig(head.getConfiguration());
    headConfig.setIterationId(ITERATION_ID);
   
    // initial input / partial solution
    headConfig.addInputToGroup(0);
    headConfig.setIterationHeadPartialSolutionOrWorksetInputIndex(0);
    headConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    headConfig.setInputComparator(vertexWithRankAndDanglingComparator, 0);
    headConfig.setInputLocalStrategy(0, LocalStrategy.SORT);
    headConfig.setMemoryInput(0, minorConsumer * JobGraphUtils.MEGABYTE);
    headConfig.setFilehandlesInput(0, NUM_FILE_HANDLES_PER_SORT);
    headConfig.setSpillingThresholdInput(0, SORT_SPILL_THRESHOLD);
   
    // back channel / iterations
    headConfig.setBackChannelMemory(minorConsumer * JobGraphUtils.MEGABYTE);
   
    // output into iteration
    headConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
   
    // final output
    TaskConfig headFinalOutConfig = new TaskConfig(new Configuration());
    headFinalOutConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    headFinalOutConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.setIterationHeadFinalOutputConfig(headFinalOutConfig);
   
    // the sync
    headConfig.setIterationHeadIndexOfSyncOutput(3);
    headConfig.setNumberOfIterations(numIterations);
   
    // the driver
    headConfig.setDriver(CollectorMapDriver.class);
    headConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
    headConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatingMap>(CustomCompensatingMap.class));
    headConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    headConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    headConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    headConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
    headConfig.addIterationAggregator(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);

    // --------------- the join ---------------------
   
    JobTaskVertex intermediate = JobGraphUtils.createTask(IterationIntermediatePactTask.class,
      "IterationIntermediate", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig intermediateConfig = new TaskConfig(intermediate.getConfiguration());
    intermediateConfig.setIterationId(ITERATION_ID);
//    intermediateConfig.setDriver(RepeatableHashjoinMatchDriverWithCachedBuildside.class);
    intermediateConfig.setDriver(BuildSecondCachedMatchDriver.class);
    intermediateConfig.setDriverStrategy(DriverStrategy.HYBRIDHASH_BUILD_SECOND);
    intermediateConfig.setMemoryDriver(matchMemory * JobGraphUtils.MEGABYTE);
    intermediateConfig.addInputToGroup(0);
    intermediateConfig.addInputToGroup(1);
    intermediateConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    intermediateConfig.setInputSerializer(vertexWithAdjacencyListSerializer, 1);
    intermediateConfig.setDriverComparator(vertexWithRankAndDanglingComparator, 0);
    intermediateConfig.setDriverComparator(vertexWithAdjacencyListComparator, 1);
    intermediateConfig.setDriverPairComparator(matchComparator);
   
    intermediateConfig.setOutputSerializer(vertexWithRankSerializer);
    intermediateConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    intermediateConfig.setOutputComparator(vertexWithRankComparator, 0);
   
    intermediateConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatableDotProductMatch>(CustomCompensatableDotProductMatch.class));
    intermediateConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    intermediateConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    intermediateConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    intermediateConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));

    // ---------------- the tail (co group) --------------------
   
    JobTaskVertex tail = JobGraphUtils.createTask(IterationTailPactTask.class, "IterationTail", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig tailConfig = new TaskConfig(tail.getConfiguration());
    tailConfig.setIterationId(ITERATION_ID);
        tailConfig.setIsWorksetUpdate();
    // TODO we need to combine!
   
    // inputs and driver
    tailConfig.setDriver(CoGroupDriver.class);
    tailConfig.setDriverStrategy(DriverStrategy.CO_GROUP);
    tailConfig.addInputToGroup(0);
    tailConfig.addInputToGroup(1);
    tailConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    tailConfig.setInputSerializer(vertexWithRankSerializer, 1);
    tailConfig.setDriverComparator(vertexWithRankAndDanglingComparator, 0);
    tailConfig.setDriverComparator(vertexWithRankComparator, 1);
    tailConfig.setDriverPairComparator(coGroupComparator);
    tailConfig.setInputAsynchronouslyMaterialized(0, true);
    tailConfig.setInputMaterializationMemory(0, minorConsumer * JobGraphUtils.MEGABYTE);
    tailConfig.setInputLocalStrategy(1, LocalStrategy.SORT);
    tailConfig.setInputComparator(vertexWithRankComparator, 1);
    tailConfig.setMemoryInput(1, coGroupSortMemory * JobGraphUtils.MEGABYTE);
    tailConfig.setFilehandlesInput(1, NUM_FILE_HANDLES_PER_SORT);
    tailConfig.setSpillingThresholdInput(1, SORT_SPILL_THRESHOLD);
   
    // output
    tailConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    tailConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
   
    // the stub
    tailConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatableDotProductCoGroup>(CustomCompensatableDotProductCoGroup.class));
    tailConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    tailConfig.setStubParameter("pageRank.numDanglingVertices", String.valueOf(numDanglingVertices));
    tailConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    tailConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    tailConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
   
    // --------------- the output ---------------------

    JobOutputVertex output = JobGraphUtils.createFileOutput(jobGraph, "FinalOutput", degreeOfParallelism,
      numSubTasksPerInstance);
    TaskConfig outputConfig = new TaskConfig(output.getConfiguration());
    outputConfig.addInputToGroup(0);
    outputConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    outputConfig.setStubWrapper(new UserCodeClassWrapper<CustomPageWithRankOutFormat>(CustomPageWithRankOutFormat.class));
    outputConfig.setStubParameter(FileOutputFormat.FILE_PARAMETER_KEY, outputPath);
   
    // --------------- the auxiliaries ---------------------
   
    JobOutputVertex fakeTailOutput = JobGraphUtils.createFakeOutput(jobGraph, "FakeTailOutput",
      degreeOfParallelism, numSubTasksPerInstance);

    JobOutputVertex sync = JobGraphUtils.createSync(jobGraph, degreeOfParallelism);
    TaskConfig syncConfig = new TaskConfig(sync.getConfiguration());
    syncConfig.setNumberOfIterations(numIterations);
    syncConfig.addIterationAggregator(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);
    syncConfig.setConvergenceCriterion(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, DiffL1NormConvergenceCriterion.class);
    syncConfig.setIterationId(ITERATION_ID);
   
    // --------------- the wiring ---------------------

    JobGraphUtils.connect(pageWithRankInput, head, ChannelType.NETWORK, DistributionPattern.BIPARTITE);

    JobGraphUtils.connect(head, intermediate, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    intermediateConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
   
    JobGraphUtils.connect(adjacencyListInput, intermediate, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
   
    JobGraphUtils.connect(head, tail, ChannelType.NETWORK, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(intermediate, tail, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(1, degreeOfParallelism);

    JobGraphUtils.connect(head, output, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(tail, fakeTailOutput, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(head, sync, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    fakeTailOutput.setVertexToShareInstancesWith(tail);
    tail.setVertexToShareInstancesWith(head);
    pageWithRankInput.setVertexToShareInstancesWith(head);
    adjacencyListInput.setVertexToShareInstancesWith(head);
    intermediate.setVertexToShareInstancesWith(head);
    output.setVertexToShareInstancesWith(head);
    sync.setVertexToShareInstancesWith(head);

    return jobGraph;
View Full Code Here
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
    JobGraph jobGraph = new JobGraph("CompensatableDanglingPageRank");
   
    // --------------- the inputs ---------------------

    // page rank input
    JobInputVertex pageWithRankInput = JobGraphUtils.createInput(new CustomImprovedDanglingPageRankInputFormat(),
      pageWithRankInputPath, "DanglingPageWithRankInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig pageWithRankInputConfig = new TaskConfig(pageWithRankInput.getConfiguration());
    pageWithRankInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    pageWithRankInputConfig.setOutputComparator(vertexWithRankAndDanglingComparator, 0);
    pageWithRankInputConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    pageWithRankInputConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));

    // edges as adjacency list
    JobInputVertex adjacencyListInput = JobGraphUtils.createInput(new CustomImprovedAdjacencyListInputFormat(),
      adjacencyListInputPath, "AdjancencyListInput", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig adjacencyListInputConfig = new TaskConfig(adjacencyListInput.getConfiguration());
    adjacencyListInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    adjacencyListInputConfig.setOutputSerializer(vertexWithAdjacencyListSerializer);
    adjacencyListInputConfig.setOutputComparator(vertexWithAdjacencyListComparator, 0);

    // --------------- the head ---------------------
    JobTaskVertex head = JobGraphUtils.createTask(IterationHeadPactTask.class, "IterationHead", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig headConfig = new TaskConfig(head.getConfiguration());
    headConfig.setIterationId(ITERATION_ID);
   
    // initial input / partial solution
    headConfig.addInputToGroup(0);
    headConfig.setIterationHeadPartialSolutionOrWorksetInputIndex(0);
    headConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    headConfig.setInputComparator(vertexWithRankAndDanglingComparator, 0);
    headConfig.setInputLocalStrategy(0, LocalStrategy.SORT);
    headConfig.setMemoryInput(0, minorConsumer * JobGraphUtils.MEGABYTE);
    headConfig.setFilehandlesInput(0, NUM_FILE_HANDLES_PER_SORT);
    headConfig.setSpillingThresholdInput(0, SORT_SPILL_THRESHOLD);
   
    // back channel / iterations
    headConfig.setBackChannelMemory(minorConsumer * JobGraphUtils.MEGABYTE);
   
    // output into iteration
    headConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
   
    // final output
    TaskConfig headFinalOutConfig = new TaskConfig(new Configuration());
    headFinalOutConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
    headFinalOutConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    headConfig.setIterationHeadFinalOutputConfig(headFinalOutConfig);
   
    // the sync
    headConfig.setIterationHeadIndexOfSyncOutput(3);
    headConfig.setNumberOfIterations(numIterations);
   
    // the driver
    headConfig.setDriver(CollectorMapDriver.class);
    headConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
    headConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatingMap>(CustomCompensatingMap.class));
    headConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    headConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    headConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    headConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
    headConfig.addIterationAggregator(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);

    // --------------- the join ---------------------
   
    JobTaskVertex intermediate = JobGraphUtils.createTask(IterationIntermediatePactTask.class,
      "IterationIntermediate", jobGraph, degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig intermediateConfig = new TaskConfig(intermediate.getConfiguration());
    intermediateConfig.setIterationId(ITERATION_ID);
//    intermediateConfig.setDriver(RepeatableHashjoinMatchDriverWithCachedBuildside.class);
    intermediateConfig.setDriver(BuildSecondCachedMatchDriver.class);
    intermediateConfig.setDriverStrategy(DriverStrategy.HYBRIDHASH_BUILD_SECOND);
    intermediateConfig.setMemoryDriver(matchMemory * JobGraphUtils.MEGABYTE);
    intermediateConfig.addInputToGroup(0);
    intermediateConfig.addInputToGroup(1);
    intermediateConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    intermediateConfig.setInputSerializer(vertexWithAdjacencyListSerializer, 1);
    intermediateConfig.setDriverComparator(vertexWithRankAndDanglingComparator, 0);
    intermediateConfig.setDriverComparator(vertexWithAdjacencyListComparator, 1);
    intermediateConfig.setDriverPairComparator(matchComparator);
   
    intermediateConfig.setOutputSerializer(vertexWithRankSerializer);
    intermediateConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
   
    intermediateConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatableDotProductMatch>(CustomCompensatableDotProductMatch.class));
    intermediateConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    intermediateConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    intermediateConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    intermediateConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
   
    // the combiner and the output
    TaskConfig combinerConfig = new TaskConfig(new Configuration());
    combinerConfig.addInputToGroup(0);
    combinerConfig.setInputSerializer(vertexWithRankSerializer, 0);
    combinerConfig.setDriverStrategy(DriverStrategy.SORTED_GROUP_COMBINE);
    combinerConfig.setDriverComparator(vertexWithRankComparator, 0);
    combinerConfig.setMemoryDriver(coGroupSortMemory * JobGraphUtils.MEGABYTE);
    combinerConfig.setOutputSerializer(vertexWithRankSerializer);
    combinerConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
    combinerConfig.setOutputComparator(vertexWithRankComparator, 0);
    combinerConfig.setStubWrapper(new UserCodeClassWrapper<CustomRankCombiner>(CustomRankCombiner.class));
    intermediateConfig.addChainedTask(SynchronousChainedCombineDriver.class, combinerConfig, "Combiner");

    // ---------------- the tail (co group) --------------------
   
    JobTaskVertex tail = JobGraphUtils.createTask(IterationTailPactTask.class, "IterationTail", jobGraph,
      degreeOfParallelism, numSubTasksPerInstance);
    TaskConfig tailConfig = new TaskConfig(tail.getConfiguration());
    tailConfig.setIterationId(ITERATION_ID);
        tailConfig.setIsWorksetUpdate();
   
    // inputs and driver
    tailConfig.setDriver(CoGroupDriver.class);
    tailConfig.setDriverStrategy(DriverStrategy.CO_GROUP);
    tailConfig.addInputToGroup(0);
    tailConfig.addInputToGroup(1);
    tailConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    tailConfig.setInputSerializer(vertexWithRankSerializer, 1);
    tailConfig.setDriverComparator(vertexWithRankAndDanglingComparator, 0);
    tailConfig.setDriverComparator(vertexWithRankComparator, 1);
    tailConfig.setDriverPairComparator(coGroupComparator);
    tailConfig.setInputAsynchronouslyMaterialized(0, true);
    tailConfig.setInputMaterializationMemory(0, minorConsumer * JobGraphUtils.MEGABYTE);
    tailConfig.setInputLocalStrategy(1, LocalStrategy.SORT);
    tailConfig.setInputComparator(vertexWithRankComparator, 1);
    tailConfig.setMemoryInput(1, coGroupSortMemory * JobGraphUtils.MEGABYTE);
    tailConfig.setFilehandlesInput(1, NUM_FILE_HANDLES_PER_SORT);
    tailConfig.setSpillingThresholdInput(1, SORT_SPILL_THRESHOLD);
    tailConfig.addIterationAggregator(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);
   
    // output
    tailConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
    tailConfig.setOutputSerializer(vertexWithRankAndDanglingSerializer);
   
    // the stub
    tailConfig.setStubWrapper(new UserCodeClassWrapper<CustomCompensatableDotProductCoGroup>(CustomCompensatableDotProductCoGroup.class));
    tailConfig.setStubParameter("pageRank.numVertices", String.valueOf(numVertices));
    tailConfig.setStubParameter("pageRank.numDanglingVertices", String.valueOf(numDanglingVertices));
    tailConfig.setStubParameter("compensation.failingWorker", failingWorkers);
    tailConfig.setStubParameter("compensation.failingIteration", String.valueOf(failingIteration));
    tailConfig.setStubParameter("compensation.messageLoss", String.valueOf(messageLoss));
   
    // --------------- the output ---------------------

    JobOutputVertex output = JobGraphUtils.createFileOutput(jobGraph, "FinalOutput", degreeOfParallelism,
      numSubTasksPerInstance);
    TaskConfig outputConfig = new TaskConfig(output.getConfiguration());
    outputConfig.addInputToGroup(0);
    outputConfig.setInputSerializer(vertexWithRankAndDanglingSerializer, 0);
    outputConfig.setStubWrapper(new UserCodeClassWrapper<CustomPageWithRankOutFormat>(CustomPageWithRankOutFormat.class));
    outputConfig.setStubParameter(FileOutputFormat.FILE_PARAMETER_KEY, outputPath);
   
    // --------------- the auxiliaries ---------------------
   
    JobOutputVertex fakeTailOutput = JobGraphUtils.createFakeOutput(jobGraph, "FakeTailOutput",
      degreeOfParallelism, numSubTasksPerInstance);

    JobOutputVertex sync = JobGraphUtils.createSync(jobGraph, degreeOfParallelism);
    TaskConfig syncConfig = new TaskConfig(sync.getConfiguration());
    syncConfig.setNumberOfIterations(numIterations);
    syncConfig.addIterationAggregator(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, PageRankStatsAggregator.class);
    syncConfig.setConvergenceCriterion(CustomCompensatableDotProductCoGroup.AGGREGATOR_NAME, DiffL1NormConvergenceCriterion.class);
    syncConfig.setIterationId(ITERATION_ID);
   
    // --------------- the wiring ---------------------

    JobGraphUtils.connect(pageWithRankInput, head, ChannelType.NETWORK, DistributionPattern.BIPARTITE);

    JobGraphUtils.connect(head, intermediate, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    intermediateConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
   
    JobGraphUtils.connect(adjacencyListInput, intermediate, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
   
    JobGraphUtils.connect(head, tail, ChannelType.NETWORK, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(intermediate, tail, ChannelType.NETWORK, DistributionPattern.BIPARTITE);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(0, 1);
    tailConfig.setGateIterativeWithNumberOfEventsUntilInterrupt(1, degreeOfParallelism);

    JobGraphUtils.connect(head, output, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);
    JobGraphUtils.connect(tail, fakeTailOutput, ChannelType.IN_MEMORY, DistributionPattern.POINTWISE);

    JobGraphUtils.connect(head, sync, ChannelType.NETWORK, DistributionPattern.POINTWISE);
   
    fakeTailOutput.setVertexToShareInstancesWith(tail);
    tail.setVertexToShareInstancesWith(head);
    pageWithRankInput.setVertexToShareInstancesWith(head);
    adjacencyListInput.setVertexToShareInstancesWith(head);
    intermediate.setVertexToShareInstancesWith(head);
    output.setVertexToShareInstancesWith(head);
    sync.setVertexToShareInstancesWith(head);

    return jobGraph;
View Full Code Here
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
   */
  private ExecutionGraph createExecutionGraph(final ChannelType channelType, final InstanceManager instanceManager) {

    final JobGraph jobGraph = new JobGraph("Job Graph");

    final JobInputVertex inputVertex = new JobInputVertex("Input 1", jobGraph);
    inputVertex.setInputClass(InputTask.class);
    inputVertex.setNumberOfSubtasks(1);

    final JobOutputVertex outputVertex = new JobOutputVertex("Output 1", jobGraph);
    outputVertex.setOutputClass(OutputTask.class);
    outputVertex.setNumberOfSubtasks(1);

    try {
      inputVertex.connectTo(outputVertex, channelType);
    } catch (JobGraphDefinitionException e) {
      fail(StringUtils.stringifyException(e));
    }

    try {
View Full Code Here
TOP

Related Classes of eu.stratosphere.nephele.jobgraph.JobInputVertex

  • eu.stratosphere.compiler.plantranslate.NepheleJobGraphGenerator
  • eu.stratosphere.nephele.jobmanager.scheduler.queue.QueueSchedulerTest
  • eu.stratosphere.test.broadcastvars.BroadcastVarsNepheleITCase
  • eu.stratosphere.test.broadcastvars.KMeansIterativeNepheleITCase
  • eu.stratosphere.test.iterative.nephele.ConnectedComponentsNepheleITCase
  • eu.stratosphere.test.iterative.nephele.customdanglingpagerank.CustomCompensatableDanglingPageRank
  • eu.stratosphere.test.iterative.nephele.customdanglingpagerank.CustomCompensatableDanglingPageRankWithCombiner
  • eu.stratosphere.test.iterative.nephele.danglingpagerank.CompensatableDanglingPageRank
  • eu.stratosphere.test.iterative.nephele.IterationWithChainingNepheleITCase
  • eu.stratosphere.test.iterative.nephele.JobGraphUtils
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.