Examples of org.apache.commons.lang3.time.StopWatch

org.apache.commons.lang3.tuple.ImmutablePair
StopWatch provides a convenient API for timings.

To start the watch, call {@link #start()}. At this point you can:
- {@link #split()} the watch to get the time whilst the watch continues in the background. {@link #unsplit()} willremove the effect of the split. At this point, these three options are available again.
- {@link #suspend()} the watch to pause it. {@link #resume()} allows the watch to continue. Any time between thesuspend and resume will not be counted in the total. At this point, these three options are available again.
- {@link #stop()} the watch to complete the timing session.
It is intended that the output methods {@link #toString()} and {@link #getTime()} should only be called after stop,split or suspend, however a suitable result will be returned at other points.

NOTE: As from v2.1, the methods protect against inappropriate calls. Thus you cannot now call stop before start, resume before suspend or unsplit before split.

1. split(), suspend(), or stop() cannot be invoked twice
2. unsplit() may only be called if the watch has been split()
3. resume() may only be called if the watch has been suspend()
4. start() cannot be called twice without calling reset()

This class is not thread-safe
@author Apache Software Foundation @since 2.0 @version $Id: StopWatch.java 959577 2010-07-01 09:48:23Z sebb $

            context.put("lenskit.eval.command.name", getName());
            context.put("lenskit.eval.algorithm.name", algorithm.getName());


            // TODO Support serializing the recommender
            LenskitRecommender rec;
            StopWatch timer = new StopWatch();
            timer.start();
            try {
                logger.info("{}: building recommender {}", getName(), algorithm.getName());
                LenskitConfiguration config = new LenskitConfiguration();
                inputData.configure(config);
                rec = algorithm.buildRecommender(config);
            } catch (RecommenderBuildException e) {
                throw new TaskExecutionException(getName() + ": error building recommender", e);
            }
            timer.stop();
            logger.info("{}: trained in {}", getName(), timer);
            return action.apply(rec);
        } finally {
            context.finish();
        }

View Full Code Here

            outputs = task.getOutputs().getPrefixed(algorithmInfo, dataSet);
            TableWriter userResults = outputs.getUserWriter();
            List<Object> outputRow = Lists.newArrayList();


            logger.info("Building {} on {}", algorithmInfo, dataSet);
            StopWatch buildTimer = new StopWatch();
            buildTimer.start();
            buildRecommender();
            buildTimer.stop();
            logger.info("Built {} in {}", algorithmInfo.getName(), buildTimer);


            logger.info("Measuring {} on {}", algorithmInfo.getName(), dataSet.getName());


            StopWatch testTimer = new StopWatch();
            testTimer.start();
            List<Object> userRow = Lists.newArrayList();


            List<MetricWithAccumulator<?>> accumulators = Lists.newArrayList();


            for (Metric<?> eval: outputs.getMetrics()) {
                accumulators.add(makeMetricAccumulator(eval));
            }


            LongSet testUsers = dataSet.getTestData().getUserDAO().getUserIds();
            final NumberFormat pctFormat = NumberFormat.getPercentInstance();
            pctFormat.setMaximumFractionDigits(2);
            pctFormat.setMinimumFractionDigits(2);
            final int nusers = testUsers.size();
            logger.info("Testing {} on {} ({} users)", algorithmInfo, dataSet, nusers);
            int ndone = 0;
            for (LongIterator iter = testUsers.iterator(); iter.hasNext();) {
                if (Thread.interrupted()) {
                    throw new InterruptedException("eval job interrupted");
                }
                long uid = iter.nextLong();
                userRow.add(uid);
                userRow.add(null); // placeholder for the per-user time
                assert userRow.size() == 2;


                Stopwatch userTimer = Stopwatch.createStarted();
                TestUser test = getUserResults(uid);


                userRow.add(test.getTrainHistory().size());
                userRow.add(test.getTestHistory().size());


                for (MetricWithAccumulator<?> accum : accumulators) {
                    List<Object> ures = accum.measureUser(test);
                    if (ures != null) {
                        userRow.addAll(ures);
                    }
                }
                userTimer.stop();
                userRow.set(1, userTimer.elapsed(TimeUnit.MILLISECONDS) * 0.001);
                if (userResults != null) {
                    try {
                        userResults.writeRow(userRow);
                    } catch (IOException e) {
                        throw new RuntimeException("error writing user row", e);
                    }
                }
                userRow.clear();


                ndone += 1;
                if (ndone % 100 == 0) {
                    testTimer.split();
                    double time = testTimer.getSplitTime();
                    double tpu = time / ndone;
                    double tleft = (nusers - ndone) * tpu;
                    logger.info("tested {} of {} users ({}), ETA {}",
                                ndone, nusers, pctFormat.format(((double) ndone) / nusers),
                                DurationFormatUtils.formatDurationHMS((long) tleft));
                }
            }
            testTimer.stop();
            logger.info("Tested {} in {}", algorithmInfo.getName(), testTimer);


            writeMetricValues(buildTimer, testTimer, outputRow, accumulators);
            bus.post(JobEvents.finished(this));
        } catch (Throwable th) {

View Full Code Here

        Vector uvec = Vector.createLength(userCount);
        Vector ivec = Vector.createLength(itemCount);


        for (int f = 0; f < featureCount; f++) {
            logger.debug("Training feature {}", f);
            StopWatch timer = new StopWatch();
            timer.start();


            uvec.fill(initialValue);
            ivec.fill(initialValue);


            FeatureInfo.Builder fib = new FeatureInfo.Builder(f);
            trainFeature(f, estimates, uvec, ivec, fib);
            summarizeFeature(uvec, ivec, fib);
            featureInfo.add(fib.build());


            // Update each rating's cached value to accommodate the feature values.
            estimates.update(uvec, ivec);


            // And store the data into the matrix
            userFeatures.setColumn(f, uvec);
            assert Math.abs(userFeatures.getColumnView(f).elementSum() - uvec.elementSum()) < 1.0e-4 : "user column sum matches";
            itemFeatures.setColumn(f, ivec);
            assert Math.abs(itemFeatures.getColumnView(f).elementSum() - ivec.elementSum()) < 1.0e-4 : "item column sum matches";


            timer.stop();
            logger.info("Finished feature {} in {}", f, timer);
        }


        // Wrap the user/item matrices because we won't use or modify them again
        return new FunkSVDModel(ImmutableMatrix.wrap(userFeatures),

View Full Code Here

    // warm up the mapper
    mapper.map(src, destClass);


    // perform x number of additional mappings
    log.info("Begin timings for " + testName);
    StopWatch timer = new StopWatch();
    timer.start();
    for (int i = 0; i < numIters; i++) {
      mapper.map(src, destClass);
    }
    timer.stop();
    log.info("Total time for additional " + numIters + " mappings: " + timer.getTime() + " milliseconds");
    log.info("avg time for " + numIters + " mappings: " + (timer.getTime() / numIters) + " milliseconds");
  }

View Full Code Here

    dbn.setMomentum(0);
    
    int recordsProcessed = 0;
    
    
    StopWatch watch = new StopWatch();
    watch.start();
    
    StopWatch batchWatch = new StopWatch();
    
    
//    do  {
      
      recordsProcessed += batchSize;
      
      System.out.println( "PreTrain: Batch Mode, Processed Total " + recordsProcessed + ", Elapsed Time " + watch.toString() );
      
      batchWatch.reset();
      batchWatch.start();
      dbn.preTrain( recordBatch.getFirst(), 1, learningRate, preTrainEpochs);
      batchWatch.stop();
      
      System.out.println( "Batch Training Elapsed Time " + batchWatch.toString() );


      System.out.println( "DBN Network Stats:\n" + dbn.generateNetworkSizeReport() );


/*      
      if (fetcher.hasNext()) {

View Full Code Here

    
    int recordsProcessed = 0;
    int batchesProcessed = 0;
    long totalBatchProcessingTime = 0;
    
    StopWatch watch = new StopWatch();
    watch.start();
    
    StopWatch batchWatch = new StopWatch();
    
    
    do  {
      
      recordsProcessed += batchSize;
      batchesProcessed++;
      
      System.out.println( "PreTrain: Batch Mode, Processed Total " + recordsProcessed + ", Elapsed Time " + watch.toString() );
      
      batchWatch.reset();
      batchWatch.start();
      dbn.preTrain( first.getFirst(), 1, learningRate, preTrainEpochs);
      batchWatch.stop();
      
      totalBatchProcessingTime += batchWatch.getTime();
      
      System.out.println( "Batch Training Elapsed Time " + batchWatch.toString() );


      //System.out.println( "DBN Network Stats:\n" + dbn.generateNetworkSizeReport() );


      
      if (fetcher.hasNext()) {

View Full Code Here

    System.out.println("Worker > Compute() -------------------------- ");


    
    int recordsProcessed = 0;
    
    StopWatch batchWatch = new StopWatch();
    
    DataSet hdfs_recordBatch = null; //this.hdfs_fetcher.next();
    
    System.out.println("Iteration: " + this.currentIteration );
    
//    if (hdfs_recordBatch.getFirst().numRows() > 0) {
//    do  {
    
    if ( TrainingState.PRE_TRAIN == this.currentTrainingState ) {
    
      System.out.println("Worker > PRE TRAIN! " );
      
       if ( this.hdfs_fetcher.hasNext() ) {
        
         
         
        hdfs_recordBatch = this.hdfs_fetcher.next();


        System.out.println("Worker > Has Next! > Recs: " + hdfs_recordBatch.getFirst().numRows() );
        
        // check for the straggler batch condition
        if (0 == this.currentIteration && hdfs_recordBatch.getFirst().numRows() > 0 && hdfs_recordBatch.getFirst().numRows() < this.batchSize) {
          
        //  System.out.println( "Worker > Straggler Batch Condition!" );
          
          // ok, only in this situation do we lower the batch size
          this.batchSize = hdfs_recordBatch.getFirst().numRows();


          // re-setup the dataset iterator
          try {
            this.hdfs_fetcher = new MnistHDFSDataSetIterator( this.batchSize, this.totalTrainingDatasetSize, (TextRecordParser)lineParser );
          } catch (IOException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
          }


        //  System.out.println( "Worker > PreTrain: Setting up for a straggler split... (sub batch size)" );          
        //  System.out.println( "New batch size: " + this.batchSize );
        } else {
          
        //  System.out.println( "Worker > NO Straggler Batch Condition!" );
          
        }
        
        if (hdfs_recordBatch.getFirst().numRows() > 0) {
          
          if (hdfs_recordBatch.getFirst().numRows() < this.batchSize) {
            
          //  System.out.println( "Worker > PreTrain: [Jagged End of Split: Skipped] Processed Total " + recordsProcessed + " Total Time " + watch.toString() );
            
            
          } else {
            
          //  System.out.println( "Worker > Normal Processing!" );
            
            // calc stats on number records processed
            recordsProcessed += hdfs_recordBatch.getFirst().numRows();
            
            //System.out.println( "PreTrain: Batch Size: " + hdfs_recordBatch.getFirst().numRows() );
            
            batchWatch.reset();
            
            batchWatch.start();
        
            this.dbn.preTrain( hdfs_recordBatch.getFirst(), 1, this.learningRate, this.preTrainEpochs);
            
            batchWatch.stop();
    
            System.out.println( "Worker > PreTrain: Batch Mode, Processed Total " + recordsProcessed + ", Batch Time " + batchWatch.toString() + " Total Time " + watch.toString() );
    
          } // if
          
          
        } else {
        
          // in case we get a blank line
          System.out.println( "Worker > PreTrain > Idle pass, no records left to process in phase" );
          
        }
        
      } else {
        
        System.out.println( "Worker > PreTrain > Idle pass, no records left to process in phase" );
        
      }
      
    //  System.out.println( "Worker > Check PreTrain completion > completedEpochs: " + this.completedDatasetEpochs + ", preTrainDatasetPasses: " + this.preTrainDatasetPasses );
      
      // check for completion of split, to signal master on state change
      if (false == this.hdfs_fetcher.hasNext() && this.completedDatasetEpochs + 1 >= this.preTrainDatasetPasses ) {
        
        this.preTrainPhaseComplete = true;
      //  System.out.println( "Worker > Completion of pre-train phase" );
        
      }
      
          
    
    } else if ( TrainingState.FINE_TUNE == this.currentTrainingState) {
      
      //System.out.println( "DBN Network Stats:\n" + dbn.generateNetworkSizeReport() );


      if ( this.hdfs_fetcher.hasNext() ) {
        
        hdfs_recordBatch = this.hdfs_fetcher.next();
        
        if (hdfs_recordBatch.getFirst().numRows() > 0) {
          
          if (hdfs_recordBatch.getFirst().numRows() < this.batchSize) {
            
          //  System.out.println( "Worker > FineTune: [Jagged End of Split: Skipped] Processed Total " + recordsProcessed + " Total Time " + watch.toString() );


          } else {
            
            batchWatch.reset();
            
            batchWatch.start();
            
            this.dbn.finetune( hdfs_recordBatch.getSecond(), learningRate, fineTuneEpochs );
            
            batchWatch.stop();
            
            System.out.println( "Worker > FineTune > Batch Mode, Processed Total " + recordsProcessed + ", Batch Time " + batchWatch.toString() + " Total Time " + watch.toString() );
            
          }
          
        } else {

View Full Code Here

    public PreBuildOperation(final T owner, final AbstractBuild<?,?> build, final BuildListener listener) {
        super(owner, build, listener);
    }


    public boolean execute() {
        StopWatch watch = new StopWatch();
        watch.start();


        log.debug("Executing");
        try {
            boolean result = doExecute();
            log.debug("Finished in {}", watch);

View Full Code Here

        super(owner, build, listener);
        this.launcher = checkNotNull(launcher);
    }


    public boolean execute() throws InterruptedException, IOException {
        StopWatch watch = new StopWatch();
        watch.start();
        log.debug("Executing");
        try {
            boolean result = doExecute();
            log.debug("Finished in {}", watch);
            return result;

View Full Code Here

        String sql = "SELECT v.proposal_id as proposalId, avg(v.note) as moy, count(v.proposal_id) as nbVote, t.title as proposalTitle FROM vote v, proposal t where (v.proposal_id=t.id) group by v.proposal_id order by moy desc";
        List<SqlRow> rows = Ebean.createSqlQuery(sql).findList();
        
        Map<Long, Pair<Double, Integer>> moyennes = new HashMap<Long, Pair<Double, Integer>>();
        for (SqlRow row : rows) {
            Pair<Double, Integer> moyProposal = new ImmutablePair(row.getDouble("moy"), row.getInteger("nbVote"));
            moyennes.put(row.getLong("proposalId"), moyProposal);
        }
        
        return moyennes;
    }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of org.apache.commons.lang3.time.StopWatch

au.org.intersect.samifier.domain.PeptideSearchResult

co.jirm.orm.writer.SqlWriterStrategy

com.crawljax.forms.FormInput

com.google.visualization.datasource.render.JsonRenderer

com.lassitercg.faces.components.event.SheetUpdate

com.lassitercg.faces.components.sheet.BadUpdate

com.lassitercg.faces.components.sheet.Sheet$RowColIndex

hudson.plugins.git.GitSCM

hudson.plugins.git.opt.PreBuildMergeOptions

jp.vmi.selenium.selenese.DriverIndependentTest

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.