d9fe5f7b67
Arena can replace MemPool in most scenarios. Except for memory reuse, MemPool supports reuse of previous memory chunks after clear, but Arena does not.
Some comparisons between MemPool and Arena:
1. Expansion
Arena is less than 128M index 2 alloc chunk; more than 128M memory, allocate 128M * n > `size`, n is equal to the minimum value that satisfies the expression;
MemPool less than 512K index 2 alloc chunk, greater than 512K memory, separately apply for a `size` length chunk
After Arena applied for a chunk larger than 128M last time, the minimum chunk applied for after that is 128M. Does this seem to be a waste of memory? MemPool is also similar. After the chunk of 512K was applied for last time, the minimum chunk of subsequent applications is 512K.
2. Alignment
MemPool defaults to 16 alignment, because memtable and other places that use int128 require 16 alignment;
Arena has no default alignment;
3. Memory reuse
Arena only supports `rollback`, which reuses the memory of the current chunk, usually the memory requested last time.
MemPool supports clear(), all chunks can be reused; or call ReturnPartialAllocation() to roll back the last requested memory; if the last chunk has no memory, search for the most free chunk for allocation
4. Realloc
Arena supports realloc contiguous memory; it also supports realloc contiguous memory from any position at the time of the last allocation. The difference between `alloc_continue` and `realloc` is:
1. Alloc_continue does not need to specify the old size, but the default old size = head->pos - range_start
2. alloc_continue supports expansion from range_start when additional_bytes is between head and pos, which is equivalent to reusing a part of memory, while realloc completely allocates a new memory
MemPool does not support realloc, but supports transferring or absorbing chunks between two MemPools
5. check mem limit
MemPool checks the mem limit, and Arena checks at the Allocator layer.
6. Support for ASAN
Arena does something extra
7. Error handling
MemPool supports returning the error message of application failure directly through `Status`, and Arena throws Exception.
Tests that Arena can consider
1. After the last applied chunk is larger than 128M, the minimum applied chunk is 128M, which seems to waste memory;
2. Support clear, memory multiplexing;
3. Increase the large list, alloc the memory larger than 128M, and the size is equal to `size`, so as to avoid the current chunk not being fully used, which is wasteful.
4. In some cases, it may be possible to allocate backwards to find chunks t
229 lines
9.7 KiB
C++
229 lines
9.7 KiB
C++
// Licensed to the Apache Software Foundation (ASF) under one
|
|
// or more contributor license agreements. See the NOTICE file
|
|
// distributed with this work for additional information
|
|
// regarding copyright ownership. The ASF licenses this file
|
|
// to you under the Apache License, Version 2.0 (the
|
|
// "License"); you may not use this file except in compliance
|
|
// with the License. You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing,
|
|
// software distributed under the License is distributed on an
|
|
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
|
// KIND, either express or implied. See the License for the
|
|
// specific language governing permissions and limitations
|
|
// under the License.
|
|
|
|
#include <gtest/gtest.h>
|
|
|
|
#include "vec/aggregate_functions/aggregate_function.h"
|
|
#include "vec/aggregate_functions/aggregate_function_simple_factory.h"
|
|
#include "vec/common/arena.h"
|
|
#include "vec/data_types/data_type.h"
|
|
#include "vec/data_types/data_type_date.h"
|
|
#include "vec/data_types/data_type_date_time.h"
|
|
#include "vec/data_types/data_type_decimal.h"
|
|
#include "vec/data_types/data_type_number.h"
|
|
#include "vec/data_types/data_type_string.h"
|
|
#include "vec/data_types/data_type_time_v2.h"
|
|
|
|
namespace doris::vectorized {
|
|
|
|
void register_aggregate_function_histogram(AggregateFunctionSimpleFactory& factory);
|
|
|
|
class VAggHistogramTest : public testing::Test {
|
|
public:
|
|
void SetUp() override {
|
|
AggregateFunctionSimpleFactory factory = AggregateFunctionSimpleFactory::instance();
|
|
register_aggregate_function_histogram(factory);
|
|
}
|
|
|
|
void TearDown() override {}
|
|
|
|
template <typename DataType>
|
|
void agg_histogram_add_elements(AggregateFunctionPtr agg_function, AggregateDataPtr place,
|
|
size_t input_rows, size_t max_num_buckets) {
|
|
using FieldType = typename DataType::FieldType;
|
|
auto type = std::make_shared<DataType>();
|
|
|
|
if (max_num_buckets == 0) {
|
|
auto input_col = type->create_column();
|
|
for (size_t i = 0; i < input_rows; ++i) {
|
|
if constexpr (std::is_same_v<DataType, DataTypeString>) {
|
|
auto item = std::string("item") + std::to_string(i);
|
|
input_col->insert_data(item.c_str(), item.size());
|
|
} else {
|
|
auto item = FieldType(static_cast<uint64_t>(i));
|
|
input_col->insert_data(reinterpret_cast<const char*>(&item), 0);
|
|
}
|
|
}
|
|
|
|
EXPECT_EQ(input_col->size(), input_rows);
|
|
|
|
const IColumn* column[1] = {input_col.get()};
|
|
for (int i = 0; i < input_col->size(); i++) {
|
|
agg_function->add(place, column, i, &_agg_arena_pool);
|
|
}
|
|
|
|
return;
|
|
}
|
|
|
|
MutableColumns columns(2);
|
|
columns[0] = type->create_column();
|
|
columns[1] = ColumnInt32::create();
|
|
|
|
for (size_t i = 0; i < input_rows; ++i) {
|
|
if constexpr (std::is_same_v<DataType, DataTypeString>) {
|
|
auto item = std::string("item") + std::to_string(i);
|
|
columns[0]->insert_data(item.c_str(), item.size());
|
|
} else {
|
|
auto item = FieldType(static_cast<uint64_t>(i));
|
|
columns[0]->insert_data(reinterpret_cast<const char*>(&item), 0);
|
|
}
|
|
columns[1]->insert_data(reinterpret_cast<char*>(&max_num_buckets),
|
|
sizeof(max_num_buckets));
|
|
}
|
|
|
|
EXPECT_EQ(columns[0]->size(), input_rows);
|
|
|
|
const IColumn* column[2] = {columns[0].get(), columns[1].get()};
|
|
for (int i = 0; i < input_rows; i++) {
|
|
agg_function->add(place, column, i, &_agg_arena_pool);
|
|
}
|
|
}
|
|
|
|
template <typename DataType>
|
|
void test_agg_histogram(size_t input_rows = 0, size_t max_num_buckets = 0) {
|
|
DataTypes data_types1 = {(DataTypePtr)std::make_shared<DataType>()};
|
|
DataTypes data_types2 = {(DataTypePtr)std::make_shared<DataType>(),
|
|
std::make_shared<DataTypeInt32>()};
|
|
|
|
auto data_types = (max_num_buckets == 0) ? data_types1 : data_types2;
|
|
LOG(INFO) << "test_agg_histogram for type"
|
|
<< "(" << data_types[0]->get_name() << ")";
|
|
|
|
AggregateFunctionSimpleFactory factory = AggregateFunctionSimpleFactory::instance();
|
|
auto agg_function = factory.get("histogram", data_types);
|
|
EXPECT_NE(agg_function, nullptr);
|
|
|
|
std::unique_ptr<char[]> memory(new char[agg_function->size_of_data()]);
|
|
AggregateDataPtr place = memory.get();
|
|
agg_function->create(place);
|
|
agg_histogram_add_elements<DataType>(agg_function, place, input_rows, max_num_buckets);
|
|
|
|
ColumnString buf;
|
|
VectorBufferWriter buf_writer(buf);
|
|
agg_function->serialize(place, buf_writer);
|
|
buf_writer.commit();
|
|
VectorBufferReader buf_reader(buf.get_data_at(0));
|
|
agg_function->deserialize(place, buf_reader, &_agg_arena_pool);
|
|
|
|
std::unique_ptr<char[]> memory2(new char[agg_function->size_of_data()]);
|
|
AggregateDataPtr place2 = memory2.get();
|
|
agg_function->create(place2);
|
|
agg_histogram_add_elements<DataType>(agg_function, place2, input_rows, max_num_buckets);
|
|
agg_function->merge(place, place2, &_agg_arena_pool);
|
|
|
|
auto column_result1 = ColumnString::create();
|
|
agg_function->insert_result_into(place, *column_result1);
|
|
EXPECT_EQ(column_result1->size(), 1);
|
|
EXPECT_TRUE(column_result1->get_offsets()[0] >= 1);
|
|
|
|
auto column_result2 = ColumnString::create();
|
|
agg_function->insert_result_into(place2, *column_result2);
|
|
EXPECT_EQ(column_result2->size(), 1);
|
|
EXPECT_TRUE(column_result2->get_offsets()[0] >= 1);
|
|
|
|
LOG(INFO) << column_result1->get_data_at(0).to_string();
|
|
LOG(INFO) << column_result2->get_data_at(0).to_string();
|
|
|
|
// test empty data
|
|
if (input_rows == 0 && max_num_buckets == 0) {
|
|
std::string expect_empty_result = "{\"num_buckets\":0,\"buckets\":[]}";
|
|
std::string empty_result1 = column_result1->get_data_at(0).to_string();
|
|
std::string empty_result2 = column_result2->get_data_at(0).to_string();
|
|
EXPECT_EQ(empty_result1, expect_empty_result);
|
|
EXPECT_EQ(empty_result2, expect_empty_result);
|
|
}
|
|
|
|
// test with data
|
|
if (input_rows == 1000 && max_num_buckets == 5) {
|
|
if constexpr (std::is_same_v<DataType, DataTypeInt32>) {
|
|
std::string expect_result1 =
|
|
"{\"num_buckets\":5,\"buckets\":["
|
|
"{\"lower\":\"0\",\"upper\":\"189\",\"count\":200,\"pre_sum\":0,\"ndv\":"
|
|
"151},"
|
|
"{\"lower\":\"190\",\"upper\":\"380\",\"count\":200,\"pre_sum\":200,"
|
|
"\"ndv\":149},"
|
|
"{\"lower\":\"382\",\"upper\":\"582\",\"count\":200,\"pre_sum\":400,"
|
|
"\"ndv\":150},"
|
|
"{\"lower\":\"586\",\"upper\":\"796\",\"count\":200,\"pre_sum\":600,"
|
|
"\"ndv\":157},"
|
|
"{\"lower\":\"797\",\"upper\":\"999\",\"count\":200,\"pre_sum\":800,"
|
|
"\"ndv\":147}]}";
|
|
std::string expect_result2 =
|
|
"{\"num_buckets\":5,\"buckets\":["
|
|
"{\"lower\":\"0\",\"upper\":\"207\",\"count\":100,\"pre_sum\":0,\"ndv\":"
|
|
"100},"
|
|
"{\"lower\":\"209\",\"upper\":\"410\",\"count\":100,\"pre_sum\":100,"
|
|
"\"ndv\":100},"
|
|
"{\"lower\":\"412\",\"upper\":\"599\",\"count\":100,\"pre_sum\":200,"
|
|
"\"ndv\":100},"
|
|
"{\"lower\":\"600\",\"upper\":\"797\",\"count\":100,\"pre_sum\":300,"
|
|
"\"ndv\":100},"
|
|
"{\"lower\":\"799\",\"upper\":\"998\",\"count\":100,\"pre_sum\":400,"
|
|
"\"ndv\":100}]}";
|
|
std::string result1 = column_result1->get_data_at(0).to_string();
|
|
std::string result2 = column_result2->get_data_at(0).to_string();
|
|
EXPECT_EQ(result1, expect_result1);
|
|
EXPECT_EQ(result2, expect_result2);
|
|
}
|
|
}
|
|
|
|
agg_function->destroy(place);
|
|
agg_function->destroy(place2);
|
|
}
|
|
|
|
private:
|
|
vectorized::Arena _agg_arena_pool;
|
|
};
|
|
|
|
TEST_F(VAggHistogramTest, test_empty) {
|
|
test_agg_histogram<DataTypeInt8>();
|
|
test_agg_histogram<DataTypeInt16>();
|
|
test_agg_histogram<DataTypeInt32>();
|
|
test_agg_histogram<DataTypeInt64>();
|
|
test_agg_histogram<DataTypeInt128>();
|
|
|
|
test_agg_histogram<DataTypeFloat32>();
|
|
test_agg_histogram<DataTypeFloat64>();
|
|
|
|
test_agg_histogram<DataTypeDate>();
|
|
test_agg_histogram<DataTypeDateTime>();
|
|
test_agg_histogram<DataTypeString>();
|
|
test_agg_histogram<DataTypeDecimal<Decimal128>>();
|
|
}
|
|
|
|
TEST_F(VAggHistogramTest, test_with_data) {
|
|
// rows 1000, max bucket size 5
|
|
test_agg_histogram<DataTypeString>(1000, 5);
|
|
test_agg_histogram<DataTypeInt8>(100, 5);
|
|
test_agg_histogram<DataTypeInt16>(100, 5);
|
|
test_agg_histogram<DataTypeInt32>(100, 5);
|
|
test_agg_histogram<DataTypeInt64>(100, 5);
|
|
test_agg_histogram<DataTypeInt128>(100, 5);
|
|
test_agg_histogram<DataTypeFloat32>(100, 5);
|
|
test_agg_histogram<DataTypeFloat64>(100, 5);
|
|
|
|
test_agg_histogram<DataTypeDate>(100, 5);
|
|
test_agg_histogram<DataTypeDateV2>(100, 5);
|
|
|
|
test_agg_histogram<DataTypeDateTime>(100, 5);
|
|
test_agg_histogram<DataTypeDateTimeV2>(100, 5);
|
|
|
|
test_agg_histogram<DataTypeDecimal<Decimal128>>(100, 5);
|
|
}
|
|
|
|
} // namespace doris::vectorized
|