Managed Memory

Functions

Managed Memory

HIP API » Memory Management

Collaboration diagram for Managed Memory:

Functions
hipError_t	hipMallocManaged (void **dev_ptr, size_t size, unsigned int flags __dparm(hipMemAttachGlobal))
	Allocates memory that will be automatically managed by HIP.
hipError_t	hipMemPrefetchAsync (const void *dev_ptr, size_t count, int device, hipStream_t stream __dparm(0))
	Prefetches memory to the specified destination device using HIP.
hipError_t	hipMemAdvise (const void *dev_ptr, size_t count, hipMemoryAdvise advice, int device)
	Advise about the usage of a given memory range to HIP.
hipError_t	hipMemRangeGetAttribute (void *data, size_t data_size, hipMemRangeAttribute attribute, const void *dev_ptr, size_t count)
	Query an attribute of a given memory range in HIP.
hipError_t	hipMemRangeGetAttributes (void **data, size_t *data_sizes, hipMemRangeAttribute *attributes, size_t num_attributes, const void *dev_ptr, size_t count)
	Query attributes of a given memory range in HIP.
hipError_t	hipStreamAttachMemAsync (hipStream_t stream, void *dev_ptr, size_t length __dparm(0), unsigned int flags __dparm(hipMemAttachSingle))
	Attach memory to a stream asynchronously in HIP.

Detailed Description

---

---

This section describes the managed memory management functions of HIP runtime API.

Note

The managed memory management APIs are implemented on Linux, under developement on Windows.

Function Documentation

◆ hipMallocManaged()

hipError_t hipMallocManaged	(	void **	dev_ptr,
		size_t	size,
		unsigned int flags	__dparmhipMemAttachGlobal
	)

Allocates memory that will be automatically managed by HIP.

This API is used for managed memory, allows data be shared and accessible to both CPU and GPU using a single pointer.

The API returns the allocation pointer, managed by HMM, can be used further to execute kernels on device and fetch data between the host and device as needed.

Note

It is recommend to do the capability check before call this API.

Parameters

[out]	dev_ptr	- pointer to allocated device memory
[in]	size	- requested allocation size in bytes, it should be granularity of 4KB
[in]	flags	- must be either hipMemAttachGlobal or hipMemAttachHost (defaults to hipMemAttachGlobal)

Returns

hipSuccess, hipErrorMemoryAllocation, hipErrorNotSupported, hipErrorInvalidValue

◆ hipMemAdvise()

hipError_t hipMemAdvise	(	const void *	dev_ptr,
		size_t	count,
		hipMemoryAdvise	advice,
		int	device
	)

Advise about the usage of a given memory range to HIP.

Parameters

[in]	dev_ptr	Pointer to memory to set the advice for
[in]	count	Size in bytes of the memory range, it should be CPU page size alligned.
[in]	advice	Advice to be applied for the specified memory range
[in]	device	Device to apply the advice for

Returns

hipSuccess, hipErrorInvalidValue

This HIP API advises about the usage to be applied on unified memory allocation in the range starting from the pointer address devPtr, with the size of count bytes. The advice can be one of enumeration value as defined in hipMemoryAdvise. The memory range must refer to managed memory allocated via the API hipMallocManaged. And the range will be handled with proper round down and round up respectively in the driver to be aligned to CPU page size, the same way as coresponding CUDA API behaves.

Note

This API is implemented on Linux, under development on Windows.

◆ hipMemPrefetchAsync()

hipError_t hipMemPrefetchAsync	(	const void *	dev_ptr,
		size_t	count,
		int	device,
		hipStream_t stream	__dparm0
	)

Prefetches memory to the specified destination device using HIP.

Parameters

[in]	dev_ptr	pointer to be prefetched
[in]	count	size in bytes for prefetching
[in]	device	destination device to prefetch to
[in]	stream	stream to enqueue prefetch operation

Returns

hipSuccess, hipErrorInvalidValue

Note

This API is implemented on Linux, under development on Windows.

◆ hipMemRangeGetAttribute()

hipError_t hipMemRangeGetAttribute	(	void *	data,
		size_t	data_size,
		hipMemRangeAttribute	attribute,
		const void *	dev_ptr,
		size_t	count
	)

Query an attribute of a given memory range in HIP.

Parameters

[in,out]	data	a pointer to a memory location where the result of each attribute query will be written to
[in]	data_size	the size of data
[in]	attribute	the attribute to query
[in]	dev_ptr	start of the range to query
[in]	count	size of the range to query

Returns

hipSuccess, hipErrorInvalidValue

Note

This API is implemented on Linux, under development on Windows.

◆ hipMemRangeGetAttributes()

hipError_t hipMemRangeGetAttributes	(	void **	data,
		size_t *	data_sizes,
		hipMemRangeAttribute *	attributes,
		size_t	num_attributes,
		const void *	dev_ptr,
		size_t	count
	)

Query attributes of a given memory range in HIP.

Parameters

[in,out]	data	a two-dimensional array containing pointers to memory locations where the result of each attribute query will be written to
[in]	data_sizes	an array, containing the sizes of each result
[in]	attributes	the attribute to query
[in]	num_attributes	an array of attributes to query (numAttributes and the number of attributes in this array should match)
[in]	dev_ptr	start of the range to query
[in]	count	size of the range to query

Returns

hipSuccess, hipErrorInvalidValue

Note

This API is implemented on Linux, under development on Windows.

◆ hipStreamAttachMemAsync()

hipError_t hipStreamAttachMemAsync	(	hipStream_t	stream,
		void *	dev_ptr,
		size_t length	__dparm0,
		unsigned int flags	__dparmhipMemAttachSingle
	)

Attach memory to a stream asynchronously in HIP.

Parameters

[in]	stream	- stream in which to enqueue the attach operation
[in]	dev_ptr	- pointer to memory (must be a pointer to managed memory or to a valid host-accessible region of system-allocated memory)
[in]	length	- length of memory (defaults to zero)
[in]	flags	- must be one of hipMemAttachGlobal, hipMemAttachHost or hipMemAttachSingle (defaults to hipMemAttachSingle)

Returns

hipSuccess, hipErrorInvalidValue

Note

This API is implemented on Linux, under development on Windows.