Common

Python module: cuvs.common

auto_sync_resources

1
def auto_sync_resources(f)

Decorator to automatically call sync on a cuVS Resources object when it isn’t passed to a function.

When a resources=None is passed to the wrapped function, this decorator will automatically create a default resources for the function, and call sync on that resources when the function exits.

This will also insert the appropriate docstring for the resources parameter

Resources

1
cdef class Resources

Resources is a lightweight python wrapper around the corresponding C++ class of resources exposed by RAFT’s C++ interface. Refer to the header file raft/core/resources.hpp for interface level details of this struct.

Parameters

Examples

Basic usage:

1
>>> from cuvs.common import Resources
2
>>> handle = Resources()
3
>>>
4
>>> # call algos here
5
>>>
6
>>> # final sync of all work launched in the stream of this handle
7
>>> handle.sync()

Using a cuPy stream with cuVS Resources:

1
>>> import cupy
2
>>> from cuvs.common import Resources
3
>>>
4
>>> cupy_stream = cupy.cuda.Stream()
5
>>> handle = Resources(stream=cupy_stream.ptr)

Members

sync

1
def sync(self)

get_c_obj

1
def get_c_obj(self)

Return the pointer to the underlying c_obj as a size_t

MultiGpuResources

1
cdef class MultiGpuResources

Multi-GPU Resources is a lightweight python wrapper around the corresponding C++ class of multi-GPU resources exposed by RAFT’s C++ interface. This class provides a handle for multi-GPU operations across all available GPUs.

Parameters

Examples

Basic usage:

1
>>> from cuvs.common import MultiGpuResources
2
>>> handle = MultiGpuResources()
3
>>>
4
>>> # call multi-GPU algos here
5
>>>
6
>>> # final sync of all work launched in the stream of this handle
7
>>> handle.sync()

Using a cuPy stream with cuVS Multi-GPU Resources:

1
>>> import cupy
2
>>> from cuvs.common import MultiGpuResources
3
>>>
4
>>> cupy_stream = cupy.cuda.Stream()
5
>>> handle = MultiGpuResources(stream=cupy_stream.ptr)

Using specific device IDs:

1
>>> from cuvs.common import MultiGpuResources
2
>>> handle = MultiGpuResources(device_ids=[0])
3
>>>
4
>>> # call multi-GPU algos here
5
>>>
6
>>> handle.sync()

Members

sync

1
def sync(self)

set_memory_pool

1
def set_memory_pool(self, percent_of_free_memory)

Set a memory pool on all devices managed by these resources.

Parameters

Examples

1
>>> from cuvs.common import MultiGpuResources
2
>>> handle = MultiGpuResources()
3
>>> handle.set_memory_pool(80)  # Use 80% of free memory

get_c_obj

1
def get_c_obj(self)

Return the pointer to the underlying c_obj as a size_t

auto_sync_multi_gpu_resources

1
def auto_sync_multi_gpu_resources(f)

Decorator to automatically call sync on a cuVS Multi-GPU Resources object when it isn’t passed to a function.

When a resources=None is passed to the wrapped function, this decorator will automatically create a default multi-GPU resources for the function, and call sync on that resources when the function exits.

This will also insert the appropriate docstring for the resources parameter