Package ai.rapids.cudf

Class RmmAllocationMode

java.lang.Object

ai.rapids.cudf.RmmAllocationMode

public class RmmAllocationMode
extends Object

Field Summary

Fields

Modifier and Type	Field	Description
static final int	ARENA	Use arena suballocation strategy
static final int	CUDA_ASYNC	Use CUDA async suballocation strategy
static final int	CUDA_ASYNC_FABRIC	Use CUDA async suballocation strategy with fabric handles that are peer accessible with read-write access
static final int	CUDA_DEFAULT	Use cudaMalloc for allocation
static final int	CUDA_MANAGED_MEMORY	Use cudaMallocManaged rather than cudaMalloc
static final int	POOL	Use pool suballocation strategy

Constructor Summary

Constructors

Constructor	Description
RmmAllocationMode()

Method Summary

Methods inherited from class java.lang.Object

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

Field Details

CUDA_DEFAULT

public static final int CUDA_DEFAULT

Use cudaMalloc for allocation

See Also:

• Constant Field Values

POOL

public static final int POOL

Use pool suballocation strategy

See Also:

• Constant Field Values

CUDA_MANAGED_MEMORY

public static final int CUDA_MANAGED_MEMORY

Use cudaMallocManaged rather than cudaMalloc

See Also:

• Constant Field Values

ARENA

public static final int ARENA

Use arena suballocation strategy

See Also:

• Constant Field Values

CUDA_ASYNC

public static final int CUDA_ASYNC

Use CUDA async suballocation strategy

See Also:

• Constant Field Values

CUDA_ASYNC_FABRIC

public static final int CUDA_ASYNC_FABRIC

Use CUDA async suballocation strategy with fabric handles that are peer accessible with read-write access

See Also:

• Constant Field Values

Constructor Details

RmmAllocationMode

public RmmAllocationMode()