Package ai.rapids.cudf

Class BaseDeviceMemoryBuffer

java.lang.Object

ai.rapids.cudf.MemoryBuffer

ai.rapids.cudf.BaseDeviceMemoryBuffer

All Implemented Interfaces:

AutoCloseable

Direct Known Subclasses:

CudaMemoryBuffer, CuFileBuffer, DeviceMemoryBuffer, DeviceMemoryBufferView

public abstract class BaseDeviceMemoryBuffer
extends MemoryBuffer

Base class for all MemoryBuffers that are in device memory.

Nested Class Summary

Nested classes/interfaces inherited from class ai.rapids.cudf.MemoryBuffer

MemoryBuffer.EventHandler, MemoryBuffer.MemoryBufferCleaner

Field Summary

Fields inherited from class ai.rapids.cudf.MemoryBuffer

address, cleaner, closed, id, length, refCount

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	BaseDeviceMemoryBuffer(long address, long length, MemoryBuffer parent)
protected	BaseDeviceMemoryBuffer(long address, long length, MemoryBuffer.MemoryBufferCleaner cleaner)

Method Summary

Modifier and Type	Method	Description
final void	copyFromDeviceBufferAsync(long destOffset, BaseDeviceMemoryBuffer src, long srcOffset, long length, Cuda.Stream stream)	Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream.
final void	copyFromHostBuffer(long destOffset, HostMemoryBuffer src, long srcOffset, long length)	Copy a subset of src to this buffer starting at destOffset.
final void	copyFromHostBuffer(long destOffset, HostMemoryBuffer src, long srcOffset, long length, Cuda.Stream stream)	Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream.
final void	copyFromHostBuffer(HostMemoryBuffer src)	Copy everything from src to this buffer starting at the beginning of this buffer.
final void	copyFromHostBuffer(HostMemoryBuffer src, long srcOffset, long length)	Copy a subset of src to this buffer starting at the beginning of this.
final void	copyFromHostBuffer(HostMemoryBuffer src, Cuda.Stream stream)	Copy entire host buffer starting at the beginning of this buffer using a CUDA stream.
final void	copyFromHostBufferAsync(long destOffset, HostMemoryBuffer src, long srcOffset, long length, Cuda.Stream stream)	Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream.
final void	copyFromHostBufferAsync(HostMemoryBuffer src, Cuda.Stream stream)	Copy entire host buffer starting at the beginning of this buffer using a CUDA stream.
final DeviceMemoryBuffer	sliceWithCopy(long offset, long len)	Slice off a part of the device buffer, copying it instead of reference counting it.

Methods inherited from class ai.rapids.cudf.MemoryBuffer

addressOutOfBoundsCheck, close, copyFromMemoryBuffer, copyFromMemoryBufferAsync, getAddress, getEventHandler, getLength, getRefCount, incRefCount, noWarnLeakExpected, setEventHandler, slice, toString

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Details

BaseDeviceMemoryBuffer

protected BaseDeviceMemoryBuffer(long address,
 long length,
 MemoryBuffer parent)

BaseDeviceMemoryBuffer

protected BaseDeviceMemoryBuffer(long address,
 long length,
 MemoryBuffer.MemoryBufferCleaner cleaner)

Method Details

copyFromHostBuffer

public final void copyFromHostBuffer(long destOffset,
 HostMemoryBuffer src,
 long srcOffset,
 long length)

Copy a subset of src to this buffer starting at destOffset.

Parameters:

destOffset - the offset in this to start copying from.

src - what to copy from

srcOffset - offset into src to start out

length - how many bytes to copy

copyFromHostBuffer

public final void copyFromHostBuffer(long destOffset,
 HostMemoryBuffer src,
 long srcOffset,
 long length,
 Cuda.Stream stream)

Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream. The copy has completed when this returns, but the memory copy could overlap with operations occurring on other streams.

Parameters:

destOffset - the offset in this to start copying from.

src - what to copy from

srcOffset - offset into src to start out

length - how many bytes to copy

stream - CUDA stream to use

copyFromHostBufferAsync

public final void copyFromHostBufferAsync(long destOffset,
 HostMemoryBuffer src,
 long srcOffset,
 long length,
 Cuda.Stream stream)

Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream. The copy is async and may not have completed when this returns.

Parameters:

destOffset - the offset in this to start copying from.

src - what to copy from

srcOffset - offset into src to start out

length - how many bytes to copy

stream - CUDA stream to use

copyFromDeviceBufferAsync

public final void copyFromDeviceBufferAsync(long destOffset,
 BaseDeviceMemoryBuffer src,
 long srcOffset,
 long length,
 Cuda.Stream stream)

Copy a subset of src to this buffer starting at destOffset using the specified CUDA stream. The copy is async and may not have completed when this returns.

Parameters:

destOffset - the offset in this to start copying from.

src - what to copy from

srcOffset - offset into src to start out

length - how many bytes to copy

stream - CUDA stream to use

copyFromHostBuffer

public final void copyFromHostBuffer(HostMemoryBuffer src,
 long srcOffset,
 long length)

Copy a subset of src to this buffer starting at the beginning of this.

Parameters:

src - what to copy from

srcOffset - offset into src to start out

length - how many bytes to copy

copyFromHostBuffer

public final void copyFromHostBuffer(HostMemoryBuffer src)

Copy everything from src to this buffer starting at the beginning of this buffer.

Parameters:

src - - Buffer to copy data from

copyFromHostBuffer

public final void copyFromHostBuffer(HostMemoryBuffer src,
 Cuda.Stream stream)

Copy entire host buffer starting at the beginning of this buffer using a CUDA stream. The copy has completed when this returns, but the memory copy could overlap with operations occurring on other streams.

Parameters:

src - host buffer to copy from

stream - CUDA stream to use

copyFromHostBufferAsync

public final void copyFromHostBufferAsync(HostMemoryBuffer src,
 Cuda.Stream stream)

Copy entire host buffer starting at the beginning of this buffer using a CUDA stream. The copy is async and may not have completed when this returns.

Parameters:

src - host buffer to copy from

stream - CUDA stream to use

sliceWithCopy

public final DeviceMemoryBuffer sliceWithCopy(long offset,
 long len)

Slice off a part of the device buffer, copying it instead of reference counting it.

Parameters:

offset - where to start the slice at.

len - how many bytes to slice

Returns:

a device buffer that will need to be closed independently from this buffer.