perf(test): Split HashJoinTest between two files for balanced CI groups

Redistribute MultiThreadedHashJoinTest tests between HashJoinTest.cpp
(34 tests) and HashJoinTestExtra.cpp (20 tests + existing HashJoinTest
tests from HashJoinTestSpill.cpp). This balances the two heaviest hash
join test binaries (~575s and ~585s) instead of the previous 919s + 242s
split.

Rename HashJoinTestSpill.cpp to HashJoinTestExtra.cpp since it now
contains both MultiThreadedHashJoinTest and HashJoinTest test cases.

Rebalance all 8 exec test groups via greedy bin-packing with measured
EC2 timings. Max group drops from ~920s to ~784s (bounded by
IndexLookupJoinTest).

Author: kgpai

velox/exec/tests/CMakeLists.txt

@@ -36,112 +36,110 @@ target_link_libraries(
   GTest::gtest_main
 )
 
-# Sources are ordered so that heavy tests are spread across groups via greedy
-# bin-packing by actual CI execution time. With VELOX_TESTS_PER_GROUP=10,
-# CMake takes files positionally in batches of 10 (files 1-10 = group0,
-# 11-20 = group1, etc.), so each batch of 10 must be independently balanced.
+# Sources are ordered via greedy bin-packing by measured EC2 execution time.
+# With VELOX_TESTS_PER_GROUP=10, CMake takes files positionally in batches of
+# 10 (files 1-10 = group0, 11-20 = group1, etc.).
 #
-# Actual CI timings (run 23771243060): group0=882s, group1=1055s, group2=365s,
-# group3=408s, group4=333s, group5=768s, group6=108s, group7=810s.
+# Measured per-file EC2 timings (sequential, 8-core, 30GB):
+#   IndexLookupJoinTest 783s, MultiFragmentTest 599s,
+#   HashJoinTestExtra ~585s, SpillerTest 584s, HashJoinTest ~575s,
+#   TableWriterTest 225s, TableScanTest 189s, IndexLookupJoinTestExtra 188s,
+#   MergeJoinTest 111s, AggregationTest 105s, OutputBufferManagerTest 98s,
+#   ScaleWriterLocalPartitionTest 96s, OrderByTest 71s, TopNRowNumberTest 64s,
+#   HashTableTest 60s, StreamingAggregationTest 58s, ExchangeClientTest 54s,
+#   RowNumberTest 42s.
 #
-# Heavy files (estimated from group timings):
-#   HashJoinTest ~870s, IndexLookupJoinTest ~520s (split from ~1040s),
-#   IndexLookupJoinTestExtra ~520s, TableWriterTest ~400s, TableScanTest ~350s,
-#   HashJoinTestSpill ~340s, MultiFragmentTest ~300s, AggregationTest ~200s,
-#   SpillerTest ~200s, WindowTest ~150s.
-#
-# Each heavy file is the sole heavyweight in its group, paired with lightweight
-# tests. One heavy file per group ensures no group exceeds ~880s.
+# Max group ~784s (bounded by IndexLookupJoinTest.cpp).
 set(
   VELOX_EXEC_TEST_SOURCES
-  # group0: HashJoinTest (~870s) + 9 lightweight
-  HashJoinTest.cpp
-  PrintPlanWithStatsTest.cpp
-  ThreadDebugInfoTest.cpp
-  WriterFuzzerUtilTest.cpp
-  UnorderedStreamReaderTest.cpp
-  ProbeOperatorStateTest.cpp
-  PlanNodeStatsTest.cpp
-  AsyncConnectorTest.cpp
-  PlanNodeToSummaryStringTest.cpp
-  PlanNodeToStringTest.cpp
-  # group1: IndexLookupJoinTest (~520s) + 9 lightweight
+  # group0 (~784s): IndexLookupJoinTest 783 + 9 lightweight
   IndexLookupJoinTest.cpp
-  MergeTest.cpp
-  FilterProjectTest.cpp
+  ConcatFilesSpillMergeStreamTest.cpp
+  MixedUnionWithTableScanTest.cpp
+  MemoryReclaimerTest.cpp
+  EnforceDistinctTest.cpp
+  TraceUtilTest.cpp
+  HashPartitionFunctionTest.cpp
+  SpatialIndexTest.cpp
+  ValuesTest.cpp
+  ParallelProjectTest.cpp
+  # group1 (~599s): MultiFragmentTest 599 + 9 lightweight
+  MultiFragmentTest.cpp
+  EnforceSingleRowTest.cpp
+  FilterToExpressionTest.cpp
+  ScaledScanControllerTest.cpp
+  HilbertIndexTest.cpp
+  OperatorTraceTest.cpp
+  LimitTest.cpp
+  SplitListenerTest.cpp
+  AddressableNonNullValueListTest.cpp
+  ArrowStreamTest.cpp
+  # group2 (~585s): HashJoinTestExtra ~585 + 9 lightweight
+  HashJoinTestExtra.cpp
+  AggregateFunctionRegistryTest.cpp
+  RoundRobinPartitionFunctionTest.cpp
+  ColumnStatsCollectorTest.cpp
+  MixedUnionTest.cpp
+  ExpandTest.cpp
+  FunctionResolutionTest.cpp
+  SpillStatsTest.cpp
+  MarkSortedTest.cpp
+  PartitionedOutputTest.cpp
+  # group3 (~584s): SpillerTest 584 + 9 lightweight
+  SpillerTest.cpp
+  PlanNodeToSummaryStringTest.cpp
   CountingJoinTest.cpp
   CustomJoinTest.cpp
-  AssignUniqueIdTest.cpp
   TaskListenerTest.cpp
   SplitTest.cpp
   SqlTest.cpp
   WindowFunctionRegistryTest.cpp
-  # group2: IndexLookupJoinTestExtra (~520s) + 9 lightweight
-  IndexLookupJoinTestExtra.cpp
-  OrderByTest.cpp
-  PrefixSortTest.cpp
-  MarkDistinctTest.cpp
-  MergerTest.cpp
   StreamingEnforceDistinctTest.cpp
   HashBitRangeTest.cpp
-  AggregateFunctionRegistryTest.cpp
-  RoundRobinPartitionFunctionTest.cpp
-  ColumnStatsCollectorTest.cpp
-  # group3: TableWriterTest (~400s) + 9 lightweight
+  # group4 (~575s): HashJoinTest ~575 + 9 lightweight
+  HashJoinTest.cpp
+  SpillTest.cpp
+  WindowTest.cpp
+  PrefixSortTest.cpp
+  MergerTest.cpp
+  LocalPartitionTest.cpp
+  PrintPlanWithStatsTest.cpp
+  ProbeOperatorStateTest.cpp
+  MarkDistinctTest.cpp
+  MergeTest.cpp
+  # group5 (~445s): TableWriterTest 225 + OutputBufMgr 98 + TopNRowNumber 64 + StreamingAgg 58
   TableWriterTest.cpp
   OutputBufferManagerTest.cpp
-  NestedLoopJoinTest.cpp
-  MixedUnionTest.cpp
-  VectorHasherTest.cpp
+  TopNRowNumberTest.cpp
+  StreamingAggregationTest.cpp
   ContainerRowSerdeTest.cpp
-  ExpandTest.cpp
-  FunctionResolutionTest.cpp
-  SpillStatsTest.cpp
-  MarkSortedTest.cpp
-  # group4: TableScanTest (~350s) + 9 lightweight
+  RowContainerTest.cpp
+  TopNTest.cpp
+  WriterFuzzerUtilTest.cpp
+  PlanNodeStatsTest.cpp
+  ThreadDebugInfoTest.cpp
+  # group6 (~467s): TableScanTest 189 + AggregationTest 105 + OrderBy 71 + HashTable 60 + RowNumber 42
   TableScanTest.cpp
+  AggregationTest.cpp
+  OrderByTest.cpp
   HashTableTest.cpp
-  StreamingAggregationTest.cpp
-  PartitionedOutputTest.cpp
-  TopNRowNumberTest.cpp
-  TopNTest.cpp
-  EnforceSingleRowTest.cpp
-  FilterToExpressionTest.cpp
-  ScaledScanControllerTest.cpp
-  HilbertIndexTest.cpp
-  # group5: HashJoinTestSpill (~340s) + OperatorTraceTest + 8 lightweight
-  HashJoinTestSpill.cpp
-  OperatorTraceTest.cpp
-  ScaleWriterLocalPartitionTest.cpp
-  LimitTest.cpp
+  RowNumberTest.cpp
   UnnestTest.cpp
-  SplitListenerTest.cpp
-  AddressableNonNullValueListTest.cpp
-  ArrowStreamTest.cpp
-  ConcatFilesSpillMergeStreamTest.cpp
-  MixedUnionWithTableScanTest.cpp
-  # group6: AggregationTest (~200s) + ExchangeClientTest + MemoryReclaimerTest + 7 lightweight
-  AggregationTest.cpp
-  ExchangeClientTest.cpp
-  MemoryReclaimerTest.cpp
-  RowContainerTest.cpp
-  SpillTest.cpp
-  LocalPartitionTest.cpp
+  PlanNodeSerdeTest.cpp
   HashJoinBridgeTest.cpp
-  RowNumberTest.cpp
-  EnforceDistinctTest.cpp
-  TraceUtilTest.cpp
-  # group7: SpillerTest (~200s) + WindowTest (~150s) + MultiFragmentTest (~300s) + 7 lightweight
-  MultiFragmentTest.cpp
-  SpillerTest.cpp
-  WindowTest.cpp
-  MergeJoinTest.cpp
   SortBufferTest.cpp
-  HashPartitionFunctionTest.cpp
-  SpatialIndexTest.cpp
-  ValuesTest.cpp
-  ParallelProjectTest.cpp
-  PlanNodeSerdeTest.cpp
+  VectorHasherTest.cpp
+  # group7 (~449s): IndexLookupJoinTestExtra 188 + MergeJoin 111 + ScaleWriter 96 + Exchange 54
+  IndexLookupJoinTestExtra.cpp
+  MergeJoinTest.cpp
+  ScaleWriterLocalPartitionTest.cpp
+  ExchangeClientTest.cpp
+  NestedLoopJoinTest.cpp
+  UnorderedStreamReaderTest.cpp
+  PlanNodeToStringTest.cpp
+  AssignUniqueIdTest.cpp
+  FilterProjectTest.cpp
+  AsyncConnectorTest.cpp
 )
 
 set(