8178148: Log more detailed information about scan rs phase

Add logging about the number of cards actually scanned, claimed and skipped during the Scan RS phase

Reviewed-by: ehelin, sangheki

hotspot/src/share/vm/gc/g1/g1CollectedHeap.cpp

@@ -3274,7 +3274,7 @@ G1CollectedHeap::do_collection_pause_at_safepoint(double target_pause_time_ms) {
         // investigate this in CR 7178365.
         double sample_end_time_sec = os::elapsedTime();
         double pause_time_ms = (sample_end_time_sec - sample_start_time_sec) * MILLIUNITS;
-        size_t total_cards_scanned = per_thread_states.total_cards_scanned();
+        size_t total_cards_scanned = g1_policy()->phase_times()->sum_thread_work_items(G1GCPhaseTimes::ScanRS, G1GCPhaseTimes::ScannedCards);
         g1_policy()->record_collection_pause_end(pause_time_ms, total_cards_scanned, heap_used_bytes_before_gc);
 
         evacuation_info.set_collectionset_used_before(collection_set()->bytes_used_before());
@@ -3464,12 +3464,10 @@ public:
       // treating the nmethods visited to act as roots for concurrent marking.
       // We only want to make sure that the oops in the nmethods are adjusted with regard to the
       // objects copied by the current evacuation.
-      size_t cards_scanned = _g1h->g1_rem_set()->oops_into_collection_set_do(&push_heap_rs_cl,
+      _g1h->g1_rem_set()->oops_into_collection_set_do(&push_heap_rs_cl,
                                                       pss->closures()->weak_codeblobs(),
                                                       worker_id);
 
-      _pss->add_cards_scanned(worker_id, cards_scanned);
-
       double strong_roots_sec = os::elapsedTime() - start_strong_roots_sec;
 
       double term_sec = 0.0;

hotspot/src/share/vm/gc/g1/g1GCPhaseTimes.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013, 2016, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2013, 2017, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -78,6 +78,13 @@ G1GCPhaseTimes::G1GCPhaseTimes(uint max_gc_threads) :
   _gc_par_phases[GCWorkerEnd] = new WorkerDataArray<double>(max_gc_threads, "GC Worker End (ms):");
   _gc_par_phases[Other] = new WorkerDataArray<double>(max_gc_threads, "GC Worker Other (ms):");
 
+  _scan_rs_scanned_cards = new WorkerDataArray<size_t>(max_gc_threads, "Scanned Cards:");
+  _gc_par_phases[ScanRS]->link_thread_work_items(_scan_rs_scanned_cards, ScannedCards);
+  _scan_rs_claimed_cards = new WorkerDataArray<size_t>(max_gc_threads, "Claimed Cards:");
+  _gc_par_phases[ScanRS]->link_thread_work_items(_scan_rs_claimed_cards, ClaimedCards);
+  _scan_rs_skipped_cards = new WorkerDataArray<size_t>(max_gc_threads, "Skipped Cards:");
+  _gc_par_phases[ScanRS]->link_thread_work_items(_scan_rs_skipped_cards, SkippedCards);
+
   _update_rs_processed_buffers = new WorkerDataArray<size_t>(max_gc_threads, "Processed Buffers:");
   _gc_par_phases[UpdateRS]->link_thread_work_items(_update_rs_processed_buffers);
 
@@ -210,8 +217,8 @@ void G1GCPhaseTimes::add_time_secs(GCParPhases phase, uint worker_i, double secs
   _gc_par_phases[phase]->add(worker_i, secs);
 }
 
-void G1GCPhaseTimes::record_thread_work_item(GCParPhases phase, uint worker_i, size_t count) {
+void G1GCPhaseTimes::record_thread_work_item(GCParPhases phase, uint worker_i, size_t count, uint index) {
-  _gc_par_phases[phase]->set_thread_work_item(worker_i, count);
+  _gc_par_phases[phase]->set_thread_work_item(worker_i, count, index);
 }
 
 // return the average time for a phase in milliseconds
@@ -219,9 +226,9 @@ double G1GCPhaseTimes::average_time_ms(GCParPhases phase) {
   return _gc_par_phases[phase]->average() * 1000.0;
 }
 
-size_t G1GCPhaseTimes::sum_thread_work_items(GCParPhases phase) {
+size_t G1GCPhaseTimes::sum_thread_work_items(GCParPhases phase, uint index) {
-  assert(_gc_par_phases[phase]->thread_work_items() != NULL, "No sub count");
+  assert(_gc_par_phases[phase]->thread_work_items(index) != NULL, "No sub count");
-  return _gc_par_phases[phase]->thread_work_items()->sum();
+  return _gc_par_phases[phase]->thread_work_items(index)->sum();
 }
 
 template <class T>
@@ -239,12 +246,14 @@ void G1GCPhaseTimes::log_phase(WorkerDataArray<double>* phase, uint indent, outp
   phase->print_summary_on(out, print_sum);
   details(phase, Indents[indent]);
 
-  WorkerDataArray<size_t>* work_items = phase->thread_work_items();
+  for (uint i = 0; i < phase->MaxThreadWorkItems; i++) {
+    WorkerDataArray<size_t>* work_items = phase->thread_work_items(i);
     if (work_items != NULL) {
       out->print("%s", Indents[indent + 1]);
       work_items->print_summary_on(out, true);
       details(work_items, Indents[indent + 1]);
     }
+  }
 }
 
 void G1GCPhaseTimes::debug_phase(WorkerDataArray<double>* phase) const {

hotspot/src/share/vm/gc/g1/g1GCPhaseTimes.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013, 2016, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2013, 2017, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -76,6 +76,12 @@ class G1GCPhaseTimes : public CHeapObj<mtGC> {
     GCParPhasesSentinel
   };
 
+  enum GCScanRSWorkItems {
+    ScannedCards,
+    ClaimedCards,
+    SkippedCards
+  };
+
  private:
   // Markers for grouping the phases in the GCPhases enum above
   static const int GCMainParPhasesLast = GCWorkerEnd;
@@ -83,8 +89,15 @@ class G1GCPhaseTimes : public CHeapObj<mtGC> {
   static const int StringDedupPhasesLast = StringDedupTableFixup;
 
   WorkerDataArray<double>* _gc_par_phases[GCParPhasesSentinel];
+
   WorkerDataArray<size_t>* _update_rs_processed_buffers;
+
+  WorkerDataArray<size_t>* _scan_rs_scanned_cards;
+  WorkerDataArray<size_t>* _scan_rs_claimed_cards;
+  WorkerDataArray<size_t>* _scan_rs_skipped_cards;
+
   WorkerDataArray<size_t>* _termination_attempts;
+
   WorkerDataArray<size_t>* _redirtied_cards;
 
   double _cur_collection_par_time_ms;
@@ -170,12 +183,12 @@ class G1GCPhaseTimes : public CHeapObj<mtGC> {
   // add a number of seconds to a phase
   void add_time_secs(GCParPhases phase, uint worker_i, double secs);
 
-  void record_thread_work_item(GCParPhases phase, uint worker_i, size_t count);
+  void record_thread_work_item(GCParPhases phase, uint worker_i, size_t count, uint index = 0);
 
   // return the average time for a phase in milliseconds
   double average_time_ms(GCParPhases phase);
 
-  size_t sum_thread_work_items(GCParPhases phase);
+  size_t sum_thread_work_items(GCParPhases phase, uint index = 0);
 
  public:
 

hotspot/src/share/vm/gc/g1/g1ParScanThreadState.cpp

@@ -337,16 +337,6 @@ G1ParScanThreadState* G1ParScanThreadStateSet::state_for_worker(uint worker_id)
   return _states[worker_id];
 }
 
-void G1ParScanThreadStateSet::add_cards_scanned(uint worker_id, size_t cards_scanned) {
-  assert(worker_id < _n_workers, "out of bounds access");
-  _cards_scanned[worker_id] += cards_scanned;
-}
-
-size_t G1ParScanThreadStateSet::total_cards_scanned() const {
-  assert(_flushed, "thread local state from the per thread states should have been flushed");
-  return _total_cards_scanned;
-}
-
 const size_t* G1ParScanThreadStateSet::surviving_young_words() const {
   assert(_flushed, "thread local state from the per thread states should have been flushed");
   return _surviving_young_words_total;
@@ -354,7 +344,6 @@ const size_t* G1ParScanThreadStateSet::surviving_young_words() const {
 
 void G1ParScanThreadStateSet::flush() {
   assert(!_flushed, "thread local state from the per thread states should be flushed once");
-  assert(_total_cards_scanned == 0, "should have been cleared");
 
   for (uint worker_index = 0; worker_index < _n_workers; ++worker_index) {
     G1ParScanThreadState* pss = _states[worker_index];
@@ -363,8 +352,6 @@ void G1ParScanThreadStateSet::flush() {
       continue;
     }
 
-    _total_cards_scanned += _cards_scanned[worker_index];
-
     pss->flush(_surviving_young_words_total);
     delete pss;
     _states[worker_index] = NULL;

hotspot/src/share/vm/gc/g1/g1ParScanThreadState.hpp

@@ -198,8 +198,6 @@ class G1ParScanThreadStateSet : public StackObj {
   G1CollectedHeap* _g1h;
   G1ParScanThreadState** _states;
   size_t* _surviving_young_words_total;
-  size_t* _cards_scanned;
-  size_t _total_cards_scanned;
   size_t _young_cset_length;
   uint _n_workers;
   bool _flushed;
@@ -209,8 +207,6 @@ class G1ParScanThreadStateSet : public StackObj {
       _g1h(g1h),
       _states(NEW_C_HEAP_ARRAY(G1ParScanThreadState*, n_workers, mtGC)),
       _surviving_young_words_total(NEW_C_HEAP_ARRAY(size_t, young_cset_length, mtGC)),
-      _cards_scanned(NEW_C_HEAP_ARRAY(size_t, n_workers, mtGC)),
-      _total_cards_scanned(0),
       _young_cset_length(young_cset_length),
       _n_workers(n_workers),
       _flushed(false) {
@@ -218,22 +214,18 @@ class G1ParScanThreadStateSet : public StackObj {
       _states[i] = NULL;
     }
     memset(_surviving_young_words_total, 0, young_cset_length * sizeof(size_t));
-    memset(_cards_scanned, 0, n_workers * sizeof(size_t));
   }
 
   ~G1ParScanThreadStateSet() {
     assert(_flushed, "thread local state from the per thread states should have been flushed");
     FREE_C_HEAP_ARRAY(G1ParScanThreadState*, _states);
     FREE_C_HEAP_ARRAY(size_t, _surviving_young_words_total);
-    FREE_C_HEAP_ARRAY(size_t, _cards_scanned);
   }
 
   void flush();
 
   G1ParScanThreadState* state_for_worker(uint worker_id);
 
-  void add_cards_scanned(uint worker_id, size_t cards_scanned);
-  size_t total_cards_scanned() const;
   const size_t* surviving_young_words() const;
 
  private:

hotspot/src/share/vm/gc/g1/g1RemSet.cpp

@@ -334,8 +334,9 @@ G1ScanRSClosure::G1ScanRSClosure(G1RemSetScanState* scan_state,
   _push_heap_cl(push_heap_cl),
   _code_root_cl(code_root_cl),
   _strong_code_root_scan_time_sec(0.0),
-  _cards(0),
+  _cards_claimed(0),
-  _cards_done(0),
+  _cards_scanned(0),
+  _cards_skipped(0),
   _worker_i(worker_i) {
   _g1h = G1CollectedHeap::heap();
   _bot = _g1h->bot();
@@ -354,7 +355,7 @@ void G1ScanRSClosure::scan_card(size_t index, HeapWord* card_start, HeapRegion *
     _ct_bs->set_card_claimed(index);
     _push_heap_cl->set_region(r);
     r->oops_on_card_seq_iterate_careful<true>(mr, _push_heap_cl);
-    _cards_done++;
+    _cards_scanned++;
   }
 }
 
@@ -389,12 +390,13 @@ bool G1ScanRSClosure::doHeapRegion(HeapRegion* r) {
       claimed_card_block = _scan_state->iter_claimed_next(region_idx, _block_size);
     }
     if (current_card < claimed_card_block) {
+      _cards_skipped++;
       continue;
     }
     HeapWord* card_start = _g1h->bot()->address_for_index(card_index);
 
     HeapRegion* card_region = _g1h->heap_region_containing(card_start);
-    _cards++;
+    _cards_claimed++;
 
     _scan_state->add_dirty_region(card_region->hrm_index());
 
@@ -411,7 +413,7 @@ bool G1ScanRSClosure::doHeapRegion(HeapRegion* r) {
   return false;
 }
 
-size_t G1RemSet::scan_rem_set(G1ParPushHeapRSClosure* oops_in_heap_closure,
+void G1RemSet::scan_rem_set(G1ParPushHeapRSClosure* oops_in_heap_closure,
                             CodeBlobClosure* heap_region_codeblobs,
                             uint worker_i) {
   double rs_time_start = os::elapsedTime();
@@ -422,10 +424,14 @@ size_t G1RemSet::scan_rem_set(G1ParPushHeapRSClosure* oops_in_heap_closure,
   double scan_rs_time_sec = (os::elapsedTime() - rs_time_start) -
                              cl.strong_code_root_scan_time_sec();
 
-  _g1p->phase_times()->record_time_secs(G1GCPhaseTimes::ScanRS, worker_i, scan_rs_time_sec);
+  G1GCPhaseTimes* p = _g1p->phase_times();
-  _g1p->phase_times()->record_time_secs(G1GCPhaseTimes::CodeRoots, worker_i, cl.strong_code_root_scan_time_sec());
 
-  return cl.cards_done();
+  p->record_time_secs(G1GCPhaseTimes::ScanRS, worker_i, scan_rs_time_sec);
+  p->record_thread_work_item(G1GCPhaseTimes::ScanRS, worker_i, cl.cards_scanned(), G1GCPhaseTimes::ScannedCards);
+  p->record_thread_work_item(G1GCPhaseTimes::ScanRS, worker_i, cl.cards_claimed(), G1GCPhaseTimes::ClaimedCards);
+  p->record_thread_work_item(G1GCPhaseTimes::ScanRS, worker_i, cl.cards_skipped(), G1GCPhaseTimes::SkippedCards);
+
+  p->record_time_secs(G1GCPhaseTimes::CodeRoots, worker_i, cl.strong_code_root_scan_time_sec());
 }
 
 // Closure used for updating RSets and recording references that
@@ -483,7 +489,7 @@ void G1RemSet::cleanupHRRS() {
   HeapRegionRemSet::cleanup();
 }
 
-size_t G1RemSet::oops_into_collection_set_do(G1ParPushHeapRSClosure* cl,
+void G1RemSet::oops_into_collection_set_do(G1ParPushHeapRSClosure* cl,
                                            CodeBlobClosure* heap_region_codeblobs,
                                            uint worker_i) {
   // A DirtyCardQueue that is used to hold cards containing references
@@ -498,7 +504,7 @@ size_t G1RemSet::oops_into_collection_set_do(G1ParPushHeapRSClosure* cl,
   DirtyCardQueue into_cset_dcq(&_into_cset_dirty_card_queue_set);
 
   update_rem_set(&into_cset_dcq, cl, worker_i);
-  return scan_rem_set(cl, heap_region_codeblobs, worker_i);;
+  scan_rem_set(cl, heap_region_codeblobs, worker_i);;
 }
 
 void G1RemSet::prepare_for_oops_into_collection_set_do() {

hotspot/src/share/vm/gc/g1/g1RemSet.hpp

@@ -108,10 +108,7 @@ public:
   // partitioning the work to be done. It should be the same as
   // the "i" passed to the calling thread's work(i) function.
   // In the sequential case this param will be ignored.
-  //
+  void oops_into_collection_set_do(G1ParPushHeapRSClosure* cl,
-  // Returns the number of cards scanned while looking for pointers
-  // into the collection set.
-  size_t oops_into_collection_set_do(G1ParPushHeapRSClosure* cl,
                                    CodeBlobClosure* heap_region_codeblobs,
                                    uint worker_i);
 
@@ -123,7 +120,7 @@ public:
   void prepare_for_oops_into_collection_set_do();
   void cleanup_after_oops_into_collection_set_do();
 
-  size_t scan_rem_set(G1ParPushHeapRSClosure* oops_in_heap_closure,
+  void scan_rem_set(G1ParPushHeapRSClosure* oops_in_heap_closure,
                     CodeBlobClosure* heap_region_codeblobs,
                     uint worker_i);
 
@@ -185,8 +182,10 @@ public:
 class G1ScanRSClosure : public HeapRegionClosure {
   G1RemSetScanState* _scan_state;
 
-  size_t _cards_done;
+  size_t _cards_scanned;
-  size_t _cards;
+  size_t _cards_claimed;
+  size_t _cards_skipped;
+
   G1CollectedHeap* _g1h;
 
   G1ParPushHeapRSClosure* _push_heap_cl;
@@ -213,8 +212,9 @@ public:
     return _strong_code_root_scan_time_sec;
   }
 
-  size_t cards_done() { return _cards_done;}
+  size_t cards_scanned() const { return _cards_scanned; }
-  size_t cards_looked_up() { return _cards;}
+  size_t cards_claimed() const { return _cards_claimed; }
+  size_t cards_skipped() const { return _cards_skipped; }
 };
 
 class UpdateRSOopClosure: public ExtendedOopClosure {

hotspot/src/share/vm/gc/g1/workerDataArray.hpp

@@ -33,20 +33,25 @@ class outputStream;
 template <class T>
 class WorkerDataArray  : public CHeapObj<mtGC> {
   friend class WDAPrinter;
+public:
+  static const uint MaxThreadWorkItems = 3;
+private:
   T*          _data;
   uint        _length;
   const char* _title;
 
-  WorkerDataArray<size_t>* _thread_work_items;
+  WorkerDataArray<size_t>* _thread_work_items[MaxThreadWorkItems];
 
  public:
   WorkerDataArray(uint length, const char* title);
   ~WorkerDataArray();
 
-  void link_thread_work_items(WorkerDataArray<size_t>* thread_work_items);
+  void link_thread_work_items(WorkerDataArray<size_t>* thread_work_items, uint index = 0);
-  void set_thread_work_item(uint worker_i, size_t value);
+  void set_thread_work_item(uint worker_i, size_t value, uint index = 0);
-  WorkerDataArray<size_t>* thread_work_items() const {
+  void add_thread_work_item(uint worker_i, size_t value, uint index = 0);
-    return _thread_work_items;
+  WorkerDataArray<size_t>* thread_work_items(uint index = 0) const {
+    assert(index < MaxThreadWorkItems, "Tried to access thread work item %u max %u", index, MaxThreadWorkItems);
+    return _thread_work_items[index];
   }
 
   static T uninitialized();

hotspot/src/share/vm/gc/g1/workerDataArray.inline.hpp

@@ -32,11 +32,13 @@
 template <typename T>
 WorkerDataArray<T>::WorkerDataArray(uint length, const char* title) :
  _title(title),
- _length(0),
+ _length(0) {
- _thread_work_items(NULL) {
   assert(length > 0, "Must have some workers to store data for");
   _length = length;
   _data = NEW_C_HEAP_ARRAY(T, _length, mtGC);
+  for (uint i = 0; i < MaxThreadWorkItems; i++) {
+    _thread_work_items[i] = NULL;
+  }
   reset();
 }
 
@@ -59,14 +61,23 @@ WorkerDataArray<T>::~WorkerDataArray() {
 }
 
 template <typename T>
-void WorkerDataArray<T>::link_thread_work_items(WorkerDataArray<size_t>* thread_work_items) {
+void WorkerDataArray<T>::link_thread_work_items(WorkerDataArray<size_t>* thread_work_items, uint index) {
-  _thread_work_items = thread_work_items;
+  assert(index < MaxThreadWorkItems, "Tried to access thread work item %u (max %u)", index, MaxThreadWorkItems);
+  _thread_work_items[index] = thread_work_items;
 }
 
 template <typename T>
-void WorkerDataArray<T>::set_thread_work_item(uint worker_i, size_t value) {
+void WorkerDataArray<T>::set_thread_work_item(uint worker_i, size_t value, uint index) {
-  assert(_thread_work_items != NULL, "No sub count");
+  assert(index < MaxThreadWorkItems, "Tried to access thread work item %u (max %u)", index, MaxThreadWorkItems);
-  _thread_work_items->set(worker_i, value);
+  assert(_thread_work_items[index] != NULL, "No sub count");
+  _thread_work_items[index]->set(worker_i, value);
+}
+
+template <typename T>
+void WorkerDataArray<T>::add_thread_work_item(uint worker_i, size_t value, uint index) {
+  assert(index < MaxThreadWorkItems, "Tried to access thread work item %u (max %u)", index, MaxThreadWorkItems);
+  assert(_thread_work_items[index] != NULL, "No sub count");
+  _thread_work_items[index]->add(worker_i, value);
 }
 
 template <typename T>
@@ -148,8 +159,10 @@ void WorkerDataArray<T>::print_details_on(outputStream* out) const {
 template <typename T>
 void WorkerDataArray<T>::reset() {
   set_all(uninitialized());
-  if (_thread_work_items != NULL) {
+  for (uint i = 0; i < MaxThreadWorkItems; i++) {
-    _thread_work_items->reset();
+    if (_thread_work_items[i] != NULL) {
+      _thread_work_items[i]->reset();
+    }
   }
 }
 

hotspot/test/gc/g1/TestGCLogMessages.java

@@ -94,7 +94,14 @@ public class TestGCLogMessages {
         new LogMessageWithLevel("Other", Level.INFO),
 
         // Update RS
+        new LogMessageWithLevel("Update RS", Level.DEBUG),
+        new LogMessageWithLevel("Processed Buffers", Level.DEBUG),
         new LogMessageWithLevel("Scan HCC", Level.TRACE),
+        // Scan RS
+        new LogMessageWithLevel("Scan RS", Level.DEBUG),
+        new LogMessageWithLevel("Scanned Cards", Level.DEBUG),
+        new LogMessageWithLevel("Claimed Cards", Level.DEBUG),
+        new LogMessageWithLevel("Skipped Cards", Level.DEBUG),
         // Ext Root Scan
         new LogMessageWithLevel("Thread Roots", Level.TRACE),
         new LogMessageWithLevel("StringTable Roots", Level.TRACE),

hotspot/test/native/gc/g1/test_workerDataArray.cpp

@@ -34,7 +34,11 @@ class WorkerDataArrayTest : public ::testing::Test {
  protected:
   WorkerDataArrayTest() :
     title("Test array"),
-    array(3, title) {
+    array(3, title),
+    sub_item_title("Sub item array"),
+    sub_item(3, sub_item_title) {
+
+    array.link_thread_work_items(&sub_item);
   }
 
   const char* print_summary() {
@@ -65,6 +69,9 @@ class WorkerDataArrayTest : public ::testing::Test {
   const char* title;
   WorkerDataArray<T> array;
 
+  const char* sub_item_title;
+  WorkerDataArray<size_t> sub_item;
+
  private:
   virtual const char* expected_summary() = 0;
   virtual const char* expected_details() = 0;
@@ -111,6 +118,10 @@ class BasicWorkerDataArrayTest : public WorkerDataArrayTest<size_t> {
     array.set(0, 5);
     array.set(1, 3);
     array.set(2, 7);
+
+    array.set_thread_work_item(0, 1);
+    array.set_thread_work_item(1, 2);
+    array.set_thread_work_item(2, 3);
   }
 
  private:
@@ -125,10 +136,12 @@ class BasicWorkerDataArrayTest : public WorkerDataArrayTest<size_t> {
 
 TEST_VM_F(BasicWorkerDataArrayTest, sum_test) {
   ASSERT_EQ(15u, array.sum());
+  ASSERT_EQ(6u, array.thread_work_items(0)->sum());
 }
 
 TEST_VM_F(BasicWorkerDataArrayTest, average_test) {
   ASSERT_NEAR(5.0, array.average(), epsilon);
+  ASSERT_NEAR(2.0, array.thread_work_items(0)->average(), epsilon);
 }
 
 TEST_VM_F(BasicWorkerDataArrayTest, print_summary_on_test) {
@@ -149,6 +162,16 @@ class AddWorkerDataArrayTest : public WorkerDataArrayTest<size_t> {
     for (uint i = 0; i < 3; i++) {
       array.add(i, 1);
     }
+
+    WorkerDataArray<size_t>* sub_items = array.thread_work_items(0);
+
+    sub_items->set(0, 1);
+    sub_items->set(1, 2);
+    sub_items->set(2, 3);
+
+    for (uint i = 0; i < 3; i++) {
+      array.add_thread_work_item(i, 1);
+    }
   }
 
  private:
@@ -163,10 +186,12 @@ class AddWorkerDataArrayTest : public WorkerDataArrayTest<size_t> {
 
 TEST_VM_F(AddWorkerDataArrayTest, sum_test) {
   ASSERT_EQ(18u, array.sum());
+  ASSERT_EQ(9u, array.thread_work_items(0)->sum());
 }
 
 TEST_VM_F(AddWorkerDataArrayTest, average_test) {
   ASSERT_NEAR(6.0, array.average(), epsilon);
+  ASSERT_NEAR(3.0, array.thread_work_items(0)->average(), epsilon);
 }
 
 TEST_VM_F(AddWorkerDataArrayTest, print_summary_on_test) {