Mila::Dnn::Compute::CudaManagedMemoryResource Class Reference

CUDA managed memory resource for unified host/device accessible memory. More...

Inheritance diagram for Mila::Dnn::Compute::CudaManagedMemoryResource:

Public Member Functions
	CudaManagedMemoryResource (int device_id)
	Constructs CUDA managed memory resource with device ID.
Public Member Functions inherited from Mila::Dnn::Compute::MemoryResource
virtual	~MemoryResource ()=default
	Virtual destructor for proper cleanup of derived classes.

Static Public Attributes
static constexpr DeviceType	device_type = DeviceType::Cuda
static constexpr bool	is_device_accessible = DeviceAccessible::is_device_accessible
	Indicates managed memory is accessible from device code.
static constexpr bool	is_host_accessible = HostAccessible::is_host_accessible
	Indicates managed memory is accessible from host code.
Static Public Attributes inherited from Mila::Dnn::Compute::MemoryResource
static constexpr bool	is_device_accessible = false
	Checks if the memory is accessible from device code.
static constexpr bool	is_host_accessible = true
	Checks if the memory is accessible from host code.

Protected Member Functions
void *	do_allocate (std::size_t bytes, std::size_t=alignof(std::max_align_t)) override
	Allocates CUDA managed memory.
void	do_deallocate (void *ptr, std::size_t, std::size_t=alignof(std::max_align_t)) override
	Deallocates CUDA managed memory.
bool	do_is_equal (const std::pmr::memory_resource &other) const noexcept override
	Compares managed memory resources for equality.

Detailed Description

CUDA managed memory resource for unified host/device accessible memory.

Provides CUDA managed (unified) memory allocation that is accessible from both host and device code with automatic migration. Focuses purely on memory allocation responsibilities without tensor-specific operations or type conversions.

Constructor & Destructor Documentation

CudaManagedMemoryResource()

Mila::Dnn::Compute::CudaManagedMemoryResource::CudaManagedMemoryResource ( int device_id )

inlineexplicit

Constructs CUDA managed memory resource with device ID.

Parameters

device_id

CUDA device ID (0, 1, 2, etc.)

Exceptions

std::invalid_argument

If device_id is invalid

Member Function Documentation

do_allocate()

void * Mila::Dnn::Compute::CudaManagedMemoryResource::do_allocate ( std::size_t bytes, std::size_t = alignof(std::max_align_t) )

inlineoverrideprotectedvirtual

Allocates CUDA managed memory.

Uses cudaMallocManaged to allocate unified memory that can be accessed from both host and device code with automatic migration.

Parameters

bytes

Number of bytes to allocate The alignment argument is ignored: CUDA allocations are already suitably aligned.

Returns

Pointer to allocated managed memory

Exceptions

std::bad_alloc

If allocation fails

Implements Mila::Dnn::Compute::MemoryResource.

do_deallocate()

void Mila::Dnn::Compute::CudaManagedMemoryResource::do_deallocate ( void * ptr, std::size_t , std::size_t = alignof(std::max_align_t) )

inlineoverrideprotectedvirtual

Deallocates CUDA managed memory.

Uses cudaFree to release managed memory. Ensures operation occurs on the correct device.

Parameters

ptr	Pointer to managed memory to deallocate The alignment argument is unused, and kept for interface compatibility.

The size argument is unused and therefore intentionally unnamed.

Implements Mila::Dnn::Compute::MemoryResource.

do_is_equal()

bool Mila::Dnn::Compute::CudaManagedMemoryResource::do_is_equal ( const std::pmr::memory_resource & other ) const

inlineoverrideprotectedvirtualnoexcept

Compares managed memory resources for equality.

Managed memory resources are equal if they are both CudaManagedMemoryResource instances with the same device ID.

Parameters

other

The other memory resource to compare with

Returns

true if both are CudaManagedMemoryResource with same device ID

Implements Mila::Dnn::Compute::MemoryResource.

---

The documentation for this class was generated from the following file:

• Mila/Src/Dnn/Compute/Devices/Cuda/CudaManagedMemoryResource.ixx