[VL] Daily Update Velox Version (2024_02_29) (#4802)

321a051 by xiaoxmeng, Improve spiller constructors (#8896)
b07d497 by Jacob Wujciak-Jens, Move Documentation Job to GHA (#8718)
fb48774 by Lu Niu, Set required simdjson version to be 3.2.0 (#8837)
6e18c05 by Yedidya Feldblum, migrate from FOLLY_MAYBE_UNUSED to [[maybe_unused]]
ec1e5cd by duanmeng, Rename addRootPool maxCapacity parameter (#8673)
c155ce9 by Deepak Majeti, Combine GCSReadBenchmark and S3ReadBenchmark with ReadBenchmark and move to benchmarks folder (#8813)
17457a8 by lingbin, Fix outdated comments in MemoryPool (#8604)
6a68ddd by duanmeng, Use ArbitratorTestUtil::newQueryCtx in SharedArbitrationTest (#8664)
490aa4c by Kevin Wilfong, Delete createSerializer from the VectorSerde class
d581e87 by Kevin Wilfong, PrestoSerializer should check the checksum before anything else
e6a986c by Jialiang Tan, Add shrinkPool implementation (#8865)
2078f23 by Ankita Victor, Add minute, second Spark function (#8831)

cpp/velox/operators/serializer/VeloxColumnarBatchSerializer.cc

@@ -60,7 +60,7 @@ std::shared_ptr<arrow::Buffer> VeloxColumnarBatchSerializer::serializeColumnarBa
   auto numRows = firstRowVector->size();
   auto arena = std::make_unique<StreamArena>(veloxPool_.get());
   auto rowType = asRowType(firstRowVector->type());
-  auto serializer = serde_->createSerializer(rowType, numRows, arena.get(), /* serdeOptions */ nullptr);
+  auto serializer = serde_->createIterativeSerializer(rowType, numRows, arena.get(), /* serdeOptions */ nullptr);
   for (auto& batch : batches) {
     auto rowVector = VeloxColumnarBatch::from(veloxPool_.get(), batch)->getRowVector();
     numRows = rowVector->size();

cpp/velox/shuffle/VeloxShuffleWriter.cc

@@ -272,8 +272,8 @@ void VeloxShuffleWriter::setPartitionBufferSize(uint16_t newSize) {
 arrow::Result<std::shared_ptr<arrow::Buffer>> VeloxShuffleWriter::generateComplexTypeBuffers(
     facebook::velox::RowVectorPtr vector) {
   auto arena = std::make_unique<facebook::velox::StreamArena>(veloxPool_.get());
-  auto serializer =
-      serde_.createSerializer(asRowType(vector->type()), vector->size(), arena.get(), /* serdeOptions */ nullptr);
+  auto serializer = serde_.createIterativeSerializer(
+      asRowType(vector->type()), vector->size(), arena.get(), /* serdeOptions */ nullptr);
   const facebook::velox::IndexRange allRows{0, vector->size()};
   serializer->append(vector, folly::Range(&allRows, 1));
   auto serializedSize = serializer->maxSerializedSize();
@@ -728,7 +728,7 @@ arrow::Status VeloxShuffleWriter::splitComplexType(const facebook::velox::RowVec
       if (arenas_[partition] == nullptr) {
         arenas_[partition] = std::make_unique<facebook::velox::StreamArena>(veloxPool_.get());
       }
-      complexTypeData_[partition] = serde_.createSerializer(
+      complexTypeData_[partition] = serde_.createIterativeSerializer(
           complexWriteType_, partition2RowCount_[partition], arenas_[partition].get(), /* serdeOptions */ nullptr);
     }
     rowIndexs[partition].emplace_back(facebook::velox::IndexRange{row, 1});

ep/build-velox/src/get_velox.sh

@@ -17,7 +17,7 @@
 set -exu
 
 VELOX_REPO=https://github.com/oap-project/velox.git
-VELOX_BRANCH=2024_02_28
+VELOX_BRANCH=2024_02_29
 VELOX_HOME=""
 
 #Set on run gluten on HDFS