Package eu.stratosphere.pact.runtime.task.util

Examples of eu.stratosphere.pact.runtime.task.util.TaskConfig

249
250
251
252
253
254
255
256
257
258
    return tail;
  }
 
  private static JobOutputVertex createSync(JobGraph jobGraph, int numIterations, int dop) {
    JobOutputVertex sync = JobGraphUtils.createSync(jobGraph, dop);
    TaskConfig syncConfig = new TaskConfig(sync.getConfiguration());
    syncConfig.setNumberOfIterations(numIterations);
    syncConfig.setIterationId(ITERATION_ID);
    return sync;
  }
View Full Code Here
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
    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);
   
View Full Code Here
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
    // --------------- 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);

View Full Code Here
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
    // --------------- 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);

View Full Code Here
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
      TypeComparatorFactory<?> comparator) {
    @SuppressWarnings("unchecked")
    CsvInputFormat verticesInFormat = new CsvInputFormat(' ', LongValue.class);
    JobInputVertex verticesInput = JobGraphUtils.createInput(verticesInFormat, verticesPath, "VerticesInput",
      jobGraph, numSubTasks, numSubTasks);
    TaskConfig verticesInputConfig = new TaskConfig(verticesInput.getConfiguration());
    {
      verticesInputConfig.addOutputShipStrategy(ShipStrategyType.FORWARD);
      verticesInputConfig.setOutputSerializer(serializer);

      // chained mapper that duplicates the id
      TaskConfig chainedMapperConfig = new TaskConfig(new Configuration());
      chainedMapperConfig.setStubWrapper(new UserCodeClassWrapper<IdDuplicator>(IdDuplicator.class));
      chainedMapperConfig.setDriverStrategy(DriverStrategy.COLLECTOR_MAP);
      chainedMapperConfig.setInputLocalStrategy(0, LocalStrategy.NONE);
      chainedMapperConfig.setInputSerializer(serializer, 0);

      chainedMapperConfig.setOutputSerializer(serializer);
      chainedMapperConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
      chainedMapperConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
      chainedMapperConfig.setOutputComparator(comparator, 0);
      chainedMapperConfig.setOutputComparator(comparator, 1);

      verticesInputConfig.addChainedTask(ChainedCollectorMapDriver.class, chainedMapperConfig, "ID Duplicator");
    }

    return verticesInput;
View Full Code Here
198
199
200
201
202
203
204
205
206
207
208
209
210
211
    // edges
    @SuppressWarnings("unchecked")
    CsvInputFormat edgesInFormat = new CsvInputFormat(' ', LongValue.class, LongValue.class);
    JobInputVertex edgesInput = JobGraphUtils.createInput(edgesInFormat, edgesPath, "EdgesInput", jobGraph,
      numSubTasks, numSubTasks);
    TaskConfig edgesInputConfig = new TaskConfig(edgesInput.getConfiguration());
    {
      edgesInputConfig.setOutputSerializer(serializer);
      edgesInputConfig.addOutputShipStrategy(ShipStrategyType.PARTITION_HASH);
      edgesInputConfig.setOutputComparator(comparator, 0);
    }

    return edgesInput;
  }
View Full Code Here
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
      TypeComparatorFactory<?> comparator,
      TypePairComparatorFactory<?, ?> pairComparator) {

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

      // initial input / workset
      headConfig.addInputToGroup(0);
      headConfig.setInputSerializer(serializer, 0);
      headConfig.setInputComparator(comparator, 0);
      headConfig.setInputLocalStrategy(0, LocalStrategy.NONE);
      headConfig.setIterationHeadPartialSolutionOrWorksetInputIndex(0);

      // regular plan input (second input to the join)
      headConfig.addInputToGroup(1);
      headConfig.setInputSerializer(serializer, 1);
      headConfig.setInputComparator(comparator, 1);
      headConfig.setInputLocalStrategy(1, LocalStrategy.NONE);
      headConfig.setInputCached(1, true);
      headConfig.setInputMaterializationMemory(1, MEM_PER_CONSUMER * JobGraphUtils.MEGABYTE);

      // initial solution set input
      headConfig.addInputToGroup(2);
      headConfig.setInputSerializer(serializer, 2);
      headConfig.setInputComparator(comparator, 2);
      headConfig.setInputLocalStrategy(2, LocalStrategy.NONE);
      headConfig.setIterationHeadSolutionSetInputIndex(2);

      headConfig.setSolutionSetSerializer(serializer);
      headConfig.setSolutionSetComparator(comparator);

      // back channel / iterations
      headConfig.setIsWorksetIteration();
      headConfig.setBackChannelMemory(MEM_PER_CONSUMER * JobGraphUtils.MEGABYTE);
      headConfig.setSolutionSetMemory(MEM_PER_CONSUMER * JobGraphUtils.MEGABYTE);

      // 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);

View Full Code Here
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
      TypeComparatorFactory<?> comparator) {

    // --------------- the intermediate (reduce to min id) ---------------
    JobTaskVertex intermediate = JobGraphUtils.createTask(IterationIntermediatePactTask.class,
      "Find Min Component-ID", jobGraph, numSubTasks, numSubTasks);
    TaskConfig intermediateConfig = new TaskConfig(intermediate.getConfiguration());
    {
      intermediateConfig.setIterationId(ITERATION_ID);

      intermediateConfig.addInputToGroup(0);
      intermediateConfig.setInputSerializer(serializer, 0);
      intermediateConfig.setInputComparator(comparator, 0);
      intermediateConfig.setInputLocalStrategy(0, LocalStrategy.SORT);
      intermediateConfig.setMemoryInput(0, MEM_PER_CONSUMER * JobGraphUtils.MEGABYTE);
      intermediateConfig.setFilehandlesInput(0, 64);
      intermediateConfig.setSpillingThresholdInput(0, 0.85f);

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

      intermediateConfig.setDriver(GroupReduceDriver.class);
      intermediateConfig.setDriverStrategy(DriverStrategy.SORTED_GROUP_REDUCE);
      intermediateConfig.setDriverComparator(comparator, 0);
      intermediateConfig.setStubWrapper(
        new UserCodeClassWrapper<MinimumComponentIDReduce>(MinimumComponentIDReduce.class));
    }

    return intermediate;
  }
View Full Code Here
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
  }

  private static JobOutputVertex createOutput(JobGraph jobGraph, String resultPath, int numSubTasks,
      TypeSerializerFactory<?> serializer) {
    JobOutputVertex output = JobGraphUtils.createFileOutput(jobGraph, "Final Output", numSubTasks, numSubTasks);
    TaskConfig outputConfig = new TaskConfig(output.getConfiguration());
    {

      outputConfig.addInputToGroup(0);
      outputConfig.setInputSerializer(serializer, 0);

      outputConfig.setStubWrapper(new UserCodeClassWrapper<CsvOutputFormat>(CsvOutputFormat.class));
      outputConfig.setStubParameter(FileOutputFormat.FILE_PARAMETER_KEY, resultPath);

      Configuration outputUserConfig = outputConfig.getStubParameters();
      outputUserConfig.setString(CsvOutputFormat.RECORD_DELIMITER_PARAMETER, "\n");
      outputUserConfig.setString(CsvOutputFormat.FIELD_DELIMITER_PARAMETER, " ");
      outputUserConfig.setClass(CsvOutputFormat.FIELD_TYPE_PARAMETER_PREFIX + 0, LongValue.class);
      outputUserConfig.setInteger(CsvOutputFormat.RECORD_POSITION_PARAMETER_PREFIX + 0, 0);
      outputUserConfig.setClass(CsvOutputFormat.FIELD_TYPE_PARAMETER_PREFIX + 1, LongValue.class);
View Full Code Here
223
224
225
226
227
228
229
230
231
232
233
234
  private static JobInputVertex createPointsInput(JobGraph jobGraph, String pointsPath, int numSubTasks, TypeSerializerFactory<?> serializer) {
    CsvInputFormat pointsInFormat = new CsvInputFormat(' ', LongValue.class, LongValue.class, LongValue.class, LongValue.class);
    JobInputVertex pointsInput = JobGraphUtils.createInput(pointsInFormat, pointsPath, "Input[Points]", jobGraph, numSubTasks, numSubTasks);

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

    return pointsInput;
  }
View Full Code Here
TOP

Related Classes of eu.stratosphere.pact.runtime.task.util.TaskConfig

  • eu.stratosphere.api.common.distributions.DataDistribution
  • eu.stratosphere.compiler.plantranslate.NepheleJobGraphGenerator
  • eu.stratosphere.compiler.plantranslate.NepheleJobGraphGenerator$IterationDescriptor
  • eu.stratosphere.configuration.Configuration
  • eu.stratosphere.pact.runtime.iterative.task.AbstractIterativePactTask
  • eu.stratosphere.pact.runtime.iterative.task.IterationHeadPactTask
  • eu.stratosphere.pact.runtime.iterative.task.IterationSynchronizationSinkTask
  • eu.stratosphere.pact.runtime.task.AbstractCachedBuildSideMatchDriver
  • eu.stratosphere.pact.runtime.task.AllGroupReduceDriver
  • eu.stratosphere.pact.runtime.task.AllReduceDriver
    • 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.