nvinfer1::v_1_0::IGpuAllocator Class Reference

#include <NvInferRuntimeBase.h>

Inheritance diagram for nvinfer1::v_1_0::IGpuAllocator:

Public Member Functions
virtual TRT_DEPRECATED void *	allocate (uint64_t const size, uint64_t const alignment, AllocatorFlags const flags) noexcept=0
	A thread-safe callback implemented by the application to handle acquisition of GPU memory. More...
	~IGpuAllocator () override=default
	IGpuAllocator ()=default
virtual void *	reallocate (void *const, uint64_t, uint64_t) noexcept
	A thread-safe callback implemented by the application to resize an existing allocation. More...
virtual TRT_DEPRECATED bool	deallocate (void *const memory) noexcept=0
	A thread-safe callback implemented by the application to handle release of GPU memory. More...
virtual void *	allocateAsync (uint64_t const size, uint64_t const alignment, AllocatorFlags const flags, cudaStream_t) noexcept
	A thread-safe callback implemented by the application to handle stream-ordered acquisition of GPU memory. More...
virtual bool	deallocateAsync (void *const memory, cudaStream_t) noexcept
	A thread-safe callback implemented by the application to handle stream-ordered release of GPU memory. More...
InterfaceInfo	getInterfaceInfo () const noexcept override
	Return version information associated with this interface. Applications must not override this method. More...
Public Member Functions inherited from nvinfer1::IVersionedInterface
virtual APILanguage	getAPILanguage () const noexcept
	The language used to build the implementation of this Interface. More...
virtual	~IVersionedInterface () noexcept=default

Additional Inherited Members
Protected Member Functions inherited from nvinfer1::IVersionedInterface
	IVersionedInterface ()=default
	IVersionedInterface (IVersionedInterface const &)=default
	IVersionedInterface (IVersionedInterface &&)=default
IVersionedInterface &	operator= (IVersionedInterface const &) &=default
IVersionedInterface &	operator= (IVersionedInterface &&) &=default

Constructor & Destructor Documentation

~IGpuAllocator()

nvinfer1::v_1_0::IGpuAllocator::~IGpuAllocator ( )

overridedefault

IGpuAllocator()

nvinfer1::v_1_0::IGpuAllocator::IGpuAllocator ( )

default

Member Function Documentation

allocate()

virtual TRT_DEPRECATED void * nvinfer1::v_1_0::IGpuAllocator::allocate ( uint64_t const  size, uint64_t const  alignment, AllocatorFlags const  flags  )

pure virtualnoexcept

A thread-safe callback implemented by the application to handle acquisition of GPU memory.

Parameters

size	The size of the memory block required (in bytes).
alignment	The required alignment of memory. Alignment will be zero or a power of 2 not exceeding the alignment guaranteed by cudaMalloc. Thus this allocator can be safely implemented with cudaMalloc/cudaFree. An alignment value of zero indicates any alignment is acceptable.
flags	Reserved for future use. In the current release, 0 will be passed.

Returns

If the allocation was successful, the start address of a device memory block of the requested size. If an allocation request of size 0 is made, nullptr must be returned. If an allocation request cannot be satisfied, nullptr must be returned. If a non-null address is returned, it is guaranteed to have the specified alignment.

Note

The implementation must guarantee thread safety for concurrent allocate/reallocate/deallocate requests.

Usage considerations

• Allowed context for the API call

  • Thread-safe: Yes, this method is required to be thread-safe and may be called from multiple threads.

  Deprecated:

  Deprecated in TensorRT 10.0. Superseded by allocateAsync

Implemented in nvinfer1::v_1_0::IGpuAsyncAllocator.

allocateAsync()

virtual void * nvinfer1::v_1_0::IGpuAllocator::allocateAsync ( uint64_t const  size, uint64_t const  alignment, AllocatorFlags const  flags, cudaStream_t    )

inlinevirtualnoexcept

A thread-safe callback implemented by the application to handle stream-ordered acquisition of GPU memory.

The default behavior is to call method allocate(), which is synchronous and thus loses any performance benefits of asynchronous allocation. If you want the benefits of asynchronous allocation, see discussion of IGpuAsyncAllocator vs. IGpuAllocator in the documentation for nvinfer1::IGpuAllocator.

Parameters

size	The size of the memory block required (in bytes).
alignment	The required alignment of memory. Alignment will be zero or a power of 2 not exceeding the alignment guaranteed by cudaMalloc. Thus this allocator can be safely implemented with cudaMalloc/cudaFree. An alignment value of zero indicates any alignment is acceptable.
flags	Reserved for future use. In the current release, 0 will be passed.
stream	specifies the cudaStream for asynchronous usage.

Returns

If the allocation was successful, the start address of a device memory block of the requested size. If an allocation request of size 0 is made, nullptr must be returned. If an allocation request cannot be satisfied, nullptr must be returned. If a non-null address is returned, it is guaranteed to have the specified alignment.

Note

The implementation must guarantee thread safety for concurrent allocate/reallocate/deallocate requests.

Usage considerations

• Allowed context for the API call
  • Thread-safe: Yes, this method is required to be thread-safe and may be called from multiple threads.

Reimplemented in nvinfer1::v_1_0::IGpuAsyncAllocator.

deallocate()

virtual TRT_DEPRECATED bool nvinfer1::v_1_0::IGpuAllocator::deallocate ( void *const  memory )

pure virtualnoexcept

A thread-safe callback implemented by the application to handle release of GPU memory.

TensorRT may pass a nullptr to this function if it was previously returned by allocate().

Parameters

memory

A memory address that was previously returned by an allocate() or reallocate() call of the same allocator object.

Returns

True if the acquired memory is released successfully.

Note

The implementation must guarantee thread safety for concurrent allocate/reallocate/deallocate requests.

Usage considerations

• Allowed context for the API call
  • Thread-safe: Yes, this method is required to be thread-safe and may be called from multiple threads.

    Deprecated:

    Deprecated in TensorRT 10.0. Superseded by deallocateAsync

Implemented in nvinfer1::v_1_0::IGpuAsyncAllocator.

deallocateAsync()

virtual bool nvinfer1::v_1_0::IGpuAllocator::deallocateAsync ( void *const  memory, cudaStream_t    )

inlinevirtualnoexcept

A thread-safe callback implemented by the application to handle stream-ordered release of GPU memory.

The default behavior is to call method deallocate(), which is synchronous and thus loses any performance benefits of asynchronous deallocation. If you want the benefits of asynchronous deallocation, see discussion of IGpuAsyncAllocator vs. IGpuAllocator in the documentation for nvinfer1::IGpuAllocator.

TensorRT may pass a nullptr to this function if it was previously returned by allocate().

Parameters

memory	A memory address that was previously returned by an allocate() or reallocate() call of the same allocator object.
stream	specifies the cudaStream for asynchronous usage.

Returns

True if the acquired memory is released successfully.

Note

The implementation must guarantee thread safety for concurrent allocate/reallocate/deallocate requests.

The implementation is not required to be asynchronous. It is permitted to synchronize, albeit doing so will lose the performance advantage of asynchronous deallocation. Either way, it is critical that it not actually free the memory until the current stream position is reached.

Usage considerations

• Allowed context for the API call
  • Thread-safe: Yes, this method is required to be thread-safe and may be called from multiple threads.

Reimplemented in nvinfer1::v_1_0::IGpuAsyncAllocator.

getInterfaceInfo()

InterfaceInfo nvinfer1::v_1_0::IGpuAllocator::getInterfaceInfo ( ) const

inlineoverridevirtualnoexcept

Return version information associated with this interface. Applications must not override this method.

Implements nvinfer1::IVersionedInterface.

reallocate()

virtual void * nvinfer1::v_1_0::IGpuAllocator::reallocate ( void * const  , uint64_t  , uint64_t    )

inlinevirtualnoexcept

A thread-safe callback implemented by the application to resize an existing allocation.

Only allocations which were allocated with AllocatorFlag::kRESIZABLE will be resized.

Options are one of:

• resize in place leaving min(oldSize, newSize) bytes unchanged and return the original address
• move min(oldSize, newSize) bytes to a new location of sufficient size and return its address
• return nullptr, to indicate that the request could not be fulfilled.

If nullptr is returned, TensorRT will assume that resize() is not implemented, and that the allocation at baseAddr is still valid.

This method is made available for use cases where delegating the resize strategy to the application provides an opportunity to improve memory management. One possible implementation is to allocate a large virtual device buffer and progressively commit physical memory with cuMemMap. CU_MEM_ALLOC_GRANULARITY_RECOMMENDED is suggested in this case.

TensorRT may call realloc to increase the buffer by relatively small amounts.

Parameters

baseAddr	the address of the original allocation, which will have been returned by previously calling allocate() or reallocate() on the same object.
alignment	The alignment used by the original allocation. This will be the same value that was previously passed to the allocate() or reallocate() call that returned baseAddr.
newSize	The new memory size required (in bytes).

Returns

The address of the reallocated memory, or nullptr. If a non-null address is returned, it is guaranteed to have the specified alignment.

Note

The implementation must guarantee thread safety for concurrent allocate/reallocate/deallocate requests.

Usage considerations

• Allowed context for the API call
  • Thread-safe: Yes, this method is required to be thread-safe and may be called from multiple threads.

---

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

• NvInferRuntimeBase.h