Fix/default stream singletons

cuda_core/cuda/core/__init__.py

@@ -61,4 +61,9 @@
 )
 from cuda.core._module import Kernel, ObjectCode  # noqa: E402
 from cuda.core._program import Program, ProgramOptions  # noqa: E402
-from cuda.core._stream import Stream, StreamOptions  # noqa: E402
+from cuda.core._stream import (  # noqa: E402
+    LEGACY_DEFAULT_STREAM,
+    PER_THREAD_DEFAULT_STREAM,
+    Stream,
+    StreamOptions,
+)

cuda_core/cuda/core/_stream.pyx

@@ -107,49 +107,13 @@ cdef class Stream:
         return s
 
     @classmethod
-    def legacy_default(cls):
-        """Return the legacy default stream.
-
-        The legacy default stream is an implicit stream which synchronizes
-        with all other streams in the same CUDA context except for non-blocking
-        streams. When any operation is launched on the legacy default stream,
-        it waits for all previously launched operations in blocking streams to
-        complete, and all subsequent operations in blocking streams wait for
-        the legacy default stream operation to complete.
-
-        Returns
-        -------
-        Stream
-            The legacy default stream instance for the current context.
-
-        See Also
-        --------
-        per_thread_default : Per-thread default stream alternative.
-
-        """
+    def _legacy_default(cls):
+        """Return the legacy default stream (supports subclassing)."""
         return Stream._from_handle(cls, get_legacy_stream())
 
     @classmethod
-    def per_thread_default(cls):
-        """Return the per-thread default stream.
-
-        The per-thread default stream is local to both the calling thread and
-        the CUDA context. Unlike the legacy default stream, it does not
-        synchronize with other streams and behaves like an explicitly created
-        non-blocking stream. This allows for better concurrency in multi-threaded
-        applications.
-
-        Returns
-        -------
-        Stream
-            The per-thread default stream instance for the current thread
-            and context.
-
-        See Also
-        --------
-        legacy_default : Legacy default stream alternative.
-
-        """
+    def _per_thread_default(cls):
+        """Return the per-thread default stream (supports subclassing)."""
         return Stream._from_handle(cls, get_per_thread_stream())
 
     @classmethod
@@ -404,8 +368,8 @@ cdef class Stream:
 
 
 # c-only python objects, not public
-cdef Stream C_LEGACY_DEFAULT_STREAM = Stream.legacy_default()
-cdef Stream C_PER_THREAD_DEFAULT_STREAM = Stream.per_thread_default()
+cdef Stream C_LEGACY_DEFAULT_STREAM = Stream._legacy_default()
+cdef Stream C_PER_THREAD_DEFAULT_STREAM = Stream._per_thread_default()
 
 # standard python objects, public
 LEGACY_DEFAULT_STREAM = C_LEGACY_DEFAULT_STREAM

cuda_core/docs/source/release/0.6.0-notes.rst

@@ -0,0 +1,33 @@
+.. SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+.. SPDX-License-Identifier: Apache-2.0
+
+.. currentmodule:: cuda.core
+
+``cuda.core`` 0.6.0 Release Notes
+==================================
+
+New features
+------------
+
+- Added public access to default CUDA streams via module-level constants ``LEGACY_DEFAULT_STREAM`` and ``PER_THREAD_DEFAULT_STREAM``
+
+  Users can now access default streams directly from the ``cuda.core`` namespace:
+
+  .. code-block:: python
+
+      from cuda.core import LEGACY_DEFAULT_STREAM, PER_THREAD_DEFAULT_STREAM
+
+      # Use legacy default stream (synchronizes with all blocking streams)
+      LEGACY_DEFAULT_STREAM.sync()
+
+      # Use per-thread default stream (non-blocking, thread-local)
+      PER_THREAD_DEFAULT_STREAM.sync()
+
+  The legacy default stream synchronizes with all blocking streams in the same CUDA context, ensuring strict ordering but potentially limiting concurrency. The per-thread default stream is local to the calling thread and does not synchronize with other streams, enabling concurrent execution in multi-threaded applications.
+
+  This replaces the previous undocumented workaround of using ``Stream.from_handle(0)`` to access the legacy default stream.
+
+Fixes and enhancements
+-----------------------
+
+None.

cuda_core/tests/test_stream.py

@@ -117,34 +117,18 @@ def test_stream_legacy_default_subclassing():
     class MyStream(Stream):
         pass
 
-    stream = MyStream.legacy_default()
+    stream = MyStream._legacy_default()
     assert isinstance(stream, MyStream)
 
 
 def test_stream_per_thread_default_subclassing():
     class MyStream(Stream):
         pass
 
-    stream = MyStream.per_thread_default()
+    stream = MyStream._per_thread_default()
     assert isinstance(stream, MyStream)
 
 
-def test_stream_legacy_default_public_api(init_cuda):
-    """Test public legacy_default() method."""
-    stream = Stream.legacy_default()
-    assert isinstance(stream, Stream)
-    # Verify it's the same as LEGACY_DEFAULT_STREAM
-    assert stream == LEGACY_DEFAULT_STREAM
-
-
-def test_stream_per_thread_default_public_api(init_cuda):
-    """Test public per_thread_default() method."""
-    stream = Stream.per_thread_default()
-    assert isinstance(stream, Stream)
-    # Verify it's the same as PER_THREAD_DEFAULT_STREAM
-    assert stream == PER_THREAD_DEFAULT_STREAM
-
-
 # ============================================================================
 # Stream Equality Tests
 # ============================================================================