Added some performance counters to HTKMLFReader

DataReader/HTKMLFReader/HTKMLFReader.cpp

@@ -25,6 +25,7 @@
 #include "DataReader.h"
 #include "commandArgUtil.h"
 #include "HTKMLFReader.h"
+#include "TimerUtility.h"
 #ifdef LEAKDETECT
 #include <vld.h> // for memory leak detection
 #endif
@@ -323,7 +324,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                 LogicError("nbrUttsInEachRecurrentIter cannot be more than 1 in frame mode reading.");
             }
 
-            int verbosity = readerConfig("verbosity","2");
+            m_verbosity = readerConfig("verbosity","2");
 
             // determine if we partial minibatches are desired
             std::string minibatchMode(readerConfig("minibatchMode","Partial"));
@@ -467,7 +468,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
 
                 // now get the frame source. This has better randomization and doesn't create temp files
                 m_frameSource.reset(new msra::dbn::minibatchutterancesourcemulti(infilesmulti, labelsmulti, m_featDims, m_labelDims, numContextLeft, numContextRight, randomize, *m_lattices, m_latticeMap, m_frameMode));
-                m_frameSource->setverbosity(verbosity);
+                m_frameSource->setverbosity(m_verbosity);
             }
             else if (!_stricmp(readMethod.c_str(),"rollingWindow"))
             {
@@ -540,7 +541,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                 int addEnergy = 0;
 
                 m_frameSource.reset(new msra::dbn::minibatchframesourcemulti(infilesmulti, labelsmulti, m_featDims, m_labelDims, numContextLeft, numContextRight, randomize, pagePaths, mayhavenoframe, addEnergy));
-                m_frameSource->setverbosity(verbosity);
+                m_frameSource->setverbosity(m_verbosity);
             }
             else
             {
@@ -915,11 +916,14 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                 {
                     if (m_nameToTypeMap.find(iter->first)==m_nameToTypeMap.end())
                         RuntimeError("minibatch requested for input node %ls not found in reader - cannot generate input\n", iter->first.c_str());
-
                 }
                 m_checkDictionaryKeys=false;
             }
 
+            Timer aggregateTimer;
+            if (m_verbosity > 2)
+                aggregateTimer.Start();
+
             do 
             {
                 if (!m_truncated)       // frame mode or whole utterances
@@ -1291,6 +1295,13 @@ namespace Microsoft { namespace MSR { namespace CNTK {
             }
             while(skip); // keep going if we didn't get the right size minibatch
 
+            if (m_verbosity > 2)
+            {
+                aggregateTimer.Stop();
+                double totalMBReadTime = aggregateTimer.ElapsedSeconds();
+                fprintf(stderr, "Total Minibatch read time = %.8g\n", totalMBReadTime);
+            }
+
             return true;
         }
 
@@ -1487,6 +1498,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
                     m_numFramesToProcess[i] = 0;
                 return false;
             }
+
             size_t numOfFea = m_featuresBufferMultiIO.size();
             size_t numOfLabel = m_labelsBufferMultiIO.size();
 
@@ -1614,7 +1626,19 @@ namespace Microsoft { namespace MSR { namespace CNTK {
 
             m_processedFrame[i] = 0;
 
+            Timer mbIterAdvancementTimer;
+            if (m_verbosity > 2)
+                mbIterAdvancementTimer.Start();
+
             (*m_mbiter)++;
+
+            if (m_verbosity > 2)
+            {
+                mbIterAdvancementTimer.Stop();
+                double advancementTime = mbIterAdvancementTimer.ElapsedSeconds();
+                fprintf(stderr, "Time to advance mbiter = %.8g\n", advancementTime);
+            }
+
             if (!(*m_mbiter))
                 m_noData = true;
 

DataReader/HTKMLFReader/HTKMLFReader.h

@@ -89,6 +89,8 @@ private:
 
     std::vector<std::vector<std::vector<ElemType>>>m_labelToTargetMapMultiIO;
     
+    int m_verbosity;
+
     void PrepareForTrainingOrTesting(const ConfigParameters& config);
     void PrepareForWriting(const ConfigParameters& config);
     

DataReader/HTKMLFReader/HTKMLFReader.vcxproj

@@ -118,6 +118,9 @@
     <ClInclude Include="utterancesourcemulti.h" />
   </ItemGroup>
   <ItemGroup>
+    <ClCompile Include="..\..\Common\TimerUtility.cpp">
+      <PrecompiledHeader>NotUsing</PrecompiledHeader>
+    </ClCompile>
     <ClCompile Include="DataReader.cpp" />
     <ClCompile Include="DataWriter.cpp" />
     <ClCompile Include="dllmain.cpp">

DataReader/HTKMLFReader/HTKMLFReader.vcxproj.filters

@@ -14,6 +14,9 @@
       <Filter>Duplicates to remove</Filter>
     </ClCompile>
     <ClCompile Include="DataReader.cpp" />
+    <ClCompile Include="..\..\Common\TimerUtility.cpp">
+      <Filter>Common</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <ClInclude Include="biggrowablevectors.h" />