接口的使用
ai.rapids.cudf.HostMemoryAllocator

使用 HostMemoryAllocator 的包

包	描述
ai.rapids.cudf

在 ai.rapids.cudf 中对 HostMemoryAllocator 的使用

在 ai.rapids.cudf 中实现 HostMemoryAllocator 的类

修饰符和类型	类和描述
类	DefaultHostMemoryAllocator

在 ai.rapids.cudf 中返回 HostMemoryAllocator 的方法

修饰符和类型	方法和描述
static HostMemoryAllocator	DefaultHostMemoryAllocator.get() 检索默认使用的当前主机内存分配器，如果未直接传递给 API

在 ai.rapids.cudf 中参数类型为 HostMemoryAllocator 的方法

修饰符和类型	方法和描述
static HostMemoryBuffer	PinnedMemoryPool.allocate(long bytes, HostMemoryAllocator hostMemoryAllocator) 创建主机缓冲区的工厂方法，但最好指向固定内存。
static JCudfSerialization.HostConcatResult	JCudfSerialization.concatToHostBuffer(JCudfSerialization.SerializedTableHeader[] headers, HostMemoryBuffer[] dataBuffers, HostMemoryAllocator hostMemoryAllocator) 将主机内存中的多个表连接成单个主机表缓冲区。
HostColumnVector	ColumnView.copyToHost(HostMemoryAllocator hostMemoryAllocator) 同步地将数据复制到主机。
HostColumnVector	ColumnView.copyToHostAsync(Cuda.Stream stream, HostMemoryAllocator hostMemoryAllocator) 异步地将数据复制到主机。
static TableWriter	Table.getCSVBufferWriter(CSVWriterOptions options, HostBufferConsumer bufferConsumer, HostMemoryAllocator hostMemoryAllocator)
static StreamedTableReader	Table.readArrowIPCChunked(ArrowIPCOptions options, HostBufferProvider provider, HostMemoryAllocator hostMemoryAllocator) 获取将返回表的读取器。
static Table	Table.readAvro(AvroOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator) 读取 Avro 格式的数据。
static Table	Table.readCSV(Schema schema, CSVOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator) 读取 CSV 格式的数据。
static Table	Table.readJSON(Schema schema, JSONOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator) 读取 JSON 格式的数据。
static Table	Table.readJSON(Schema schema, JSONOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator, int emptyRowCount) 已弃用。 此方法已弃用，因为 emptyRowCount 未使用。请改用没有 emptyRowCount 的方法。
static Table	Table.readORC(ORCOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator) 读取 ORC 格式的数据。
static Table	Table.readParquet(ParquetOptions opts, byte[] buffer, long offset, long len, HostMemoryAllocator hostMemoryAllocator) 读取 parquet 格式的数据。
static JCudfSerialization.TableAndRowCountPair	JCudfSerialization.readTableFrom(InputStream in, HostMemoryAllocator hostMemoryAllocator) 从给定的 InputStream 读取序列化表。
static void	DefaultHostMemoryAllocator.set(HostMemoryAllocator hostMemoryAllocator) 设置新的默认主机内存分配器实现。
static TableWriter	Table.writeArrowIPCChunked(ArrowIPCWriterOptions options, HostBufferConsumer consumer, HostMemoryAllocator hostMemoryAllocator) 获取一个表写入器，用于写入 arrow IPC 数据并使用回调处理每个块。
static void	Table.writeColumnViewsToParquet(ParquetWriterOptions options, HostBufferConsumer consumer, HostMemoryAllocator hostMemoryAllocator, ColumnView... columnViews) 这是一个不断演进的 API，很可能在未来版本中移除。
static TableWriter	Table.writeORCChunked(ORCWriterOptions options, HostBufferConsumer consumer, HostMemoryAllocator hostMemoryAllocator) 获取一个表写入器，用于写入 ORC 数据并使用回调处理每个块。
static TableWriter	Table.writeParquetChunked(ParquetWriterOptions options, HostBufferConsumer consumer, HostMemoryAllocator hostMemoryAllocator) 获取一个表写入器，用于写入 parquet 数据并使用回调处理每个块。

在 ai.rapids.cudf 中参数类型为 HostMemoryAllocator 的构造函数

构造函数和描述
MultiBufferDataSource(HostMemoryAllocator allocator, HostMemoryBuffer... buffers) 创建由多个缓冲区支持的新数据源。