Implements IPC-enabled memory pools for Linux in DeviceMemoryResource (#930)

* Updates DeviceMemoryResource to accept Device objects for the device_id argument.

* Adds IPC support to DeviceMemoryResource. Creates a staticmethod `current` to access the default memory pool.

* Added a missing import statement.

* Restores DeviceMemoryResource default behavior to return the current memory pool. Removes the `current` staticmethod. Adds an option dataclass for constructor options.

* Adjusts docstrings for consistency.

* Adjusts dataclass member defs in *Options classes for Cython.

* Adds __bool__ to the MemoryResource interface.

* Minor docstring update.

* Changes verbiage from "shared" or "shareable" handle to "allocation handle" to better align with the driver API.

* Significantly reworks the tests for IPC-enabled memory pools.
Introduces `IPCAllocationHandle` to manage pool-sharing resources.
Introduces `IPCChannel` to for sharing allocation handles in a
platform-independent way (though currently only Linux is supported).

* Various fixes including: groups cimport statements; removes __bool__ method from MemoryResource; Cythonizes helper classes; disables __init__ for non-API classes; removes abstract factory constructor from IPCChannel; removes the _requires_ipc decorator (checks are now inlined)

* Formatting changes.

* Creates an attributes suite to bundle the DeviceMemoryResource attributes.

* Format changes.

* Test updates. Parameterizes tests (rather than use internal loops). Adds stream synchronization. Eliminates unnecessary clean-up. Removes unnecessary check for CUDA 12 or later.

* Reworks mempool attribute implementation.

* Adds testing for errors when allocating from an imported memory pool.

* Move IPC import/export methods into the Buffer class.

* Updated release notes.

* Reworks DeviceMemoryResourceAttributes to use a descriptor.

* Remove use of deprecated abstractproperty.


Co-authored-by: Keenan Simpson <ksimpson@nvidia.com>

Author: Andy-Jost

cuda_core/cuda/core/experimental/__init__.py

@@ -14,7 +14,13 @@
 from cuda.core.experimental._launch_config import LaunchConfig
 from cuda.core.experimental._launcher import launch
 from cuda.core.experimental._linker import Linker, LinkerOptions
-from cuda.core.experimental._memory import Buffer, DeviceMemoryResource, LegacyPinnedMemoryResource, MemoryResource
+from cuda.core.experimental._memory import (
+    Buffer,
+    DeviceMemoryResource,
+    IPCChannel,
+    LegacyPinnedMemoryResource,
+    MemoryResource,
+)
 from cuda.core.experimental._module import Kernel, ObjectCode
 from cuda.core.experimental._program import Program, ProgramOptions
 from cuda.core.experimental._stream import Stream, StreamOptions