Ginkgo  Generated from pipelines/1554403166 branch based on develop. Ginkgo version 1.9.0
A numerical linear algebra library targeting many-core architectures
Public Member Functions | List of all members
gko::CudaAsyncAllocator Class Reference
Inheritance diagram for gko::CudaAsyncAllocator:
[legend]
Collaboration diagram for gko::CudaAsyncAllocator:
[legend]

Public Member Functions

void * allocate (size_type num_bytes) override
 
void deallocate (void *ptr) override
 
 CudaAsyncAllocator (CUstream_st *stream)
 
bool check_environment (int device_id, CUstream_st *stream) const override
 Checks if the allocator can be used safely with the provided device ID and stream. More...
 

Member Function Documentation

◆ check_environment()

bool gko::CudaAsyncAllocator::check_environment ( int  device_id,
CUstream_st *  stream 
) const
overridevirtual

Checks if the allocator can be used safely with the provided device ID and stream.

The check is necessary to ensure safe usage of stream-ordered allocators and unified shared memory allocators.

Parameters
device_idthe device ID the allocator will be used in.
streamthe stream the allocator will be used with.
Returns
true if and only if the allocator can be used by CudaExecutor in the given environment.

Reimplemented from gko::CudaAllocatorBase.


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