Regenerate bindings: !524 Support full-length docstrings

cuda_bindings/cuda/bindings/_v2/nvrtc.pyx

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 #
 # This code was automatically generated across versions from 12.9.0 to 13.4.1. Do not modify it directly.
-# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=3b41a6e66b9064266d04521a6cc9b9a4f715c453d21323ab9741f9e34e6f2d80
+# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=27a65e25f2ddca0b89c0e3017c34f26b42c64fc9b221475ab0b8198207d44647
 
 
 # <<<< PREAMBLE CONTENT >>>>
@@ -657,6 +657,8 @@ cpdef tuple version():
 cpdef int get_num_supported_archs() except? -1:
     """nvrtcGetNumSupportedArchs sets the output parameter ``num_archs`` with the number of architectures supported by NVRTC. This can then be used to pass an array to ``nvrtcGetSupportedArchs`` to get the supported architectures.
 
+    see ``nvrtcGetSupportedArchs``.
+
     Returns:
         int: number of supported architectures.
 
@@ -672,6 +674,8 @@ cpdef int get_num_supported_archs() except? -1:
 cpdef object get_supported_archs():
     """nvrtcGetSupportedArchs populates the array passed via the output parameter ``supported_archs`` with the architectures supported by NVRTC. The array is sorted in the ascending order. The size of the array to be passed can be determined using ``nvrtcGetNumSupportedArchs``.
 
+    see ``nvrtcGetNumSupportedArchs``.
+
     Returns:
         int: sorted array of supported architectures.
 
@@ -881,6 +885,9 @@ cpdef bytes get_optix_ir(intptr_t prog):
 cpdef size_t get_program_log_size(intptr_t prog) except? 0:
     """nvrtcGetProgramLogSize sets ``log_size_ret`` with the size of the log generated by the previous compilation of ``prog`` (including the trailing ``NULL``).
 
+    Note that compilation log may be generated with warnings and informative
+    messages, even when the compilation of ``prog`` succeeds.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
 
@@ -925,6 +932,9 @@ cpdef bytes get_program_log(intptr_t prog):
 cpdef add_name_expression(intptr_t prog, name_expression):
     """nvrtcAddNameExpression notes the given name expression denoting the address of a global function or device/__constant__ variable.
 
+    The identical name expression string must be provided on a subsequent call
+    to nvrtcGetLoweredName to extract the lowered name.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
         name_expression (str): constant expression denoting the
@@ -961,6 +971,10 @@ cpdef size_t get_pch_heap_size() except? 0:
 cpdef set_pch_heap_size(size_t size):
     """set the size of the PCH Heap.
 
+    The requested size may be rounded up to a platform dependent alignment
+    (e.g. page size). If the PCH Heap has already been allocated, the heap
+    memory will be freed and a new PCH Heap will be allocated.
+
     Args:
         size (size_t): requested size of the PCH Heap, in bytes.
 
@@ -974,6 +988,20 @@ cpdef set_pch_heap_size(size_t size):
 cpdef int get_pch_create_status(intptr_t prog) except? -1:
     """returns the PCH creation status.
 
+    NVRTC_SUCCESS indicates that the PCH was successfully created.
+    NVRTC_ERROR_NO_PCH_CREATE_ATTEMPTED indicates that no PCH creation was
+    attempted, either because PCH functionality was not requested during the
+    preceding nvrtcCompileProgram call, or automatic PCH processing was
+    requested, and compiler chose not to create a PCH file.
+    NVRTC_ERROR_PCH_CREATE_HEAP_EXHAUSTED indicates that a PCH file could
+    potentially have been created, but the compiler ran out space in the PCH
+    heap. In this scenario, the :func:`get_pch_heap_size_required` can be used
+    to query the required heap size, the heap can be reallocated for this size
+    with :func:`set_pch_heap_size` and PCH creation may be reattempted again
+    invoking :func:`compile_program` with a new NVRTC program instance.
+    NVRTC_ERROR_PCH_CREATE indicates that an error condition prevented the PCH
+    file from being created.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
 

cuda_bindings/docs/source/release/13.5.0-notes.rst

@@ -0,0 +1,27 @@
+.. SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+.. SPDX-License-Identifier: Apache-2.0
+
+.. module:: cuda.bindings
+
+``cuda-bindings`` 13.5.0 Release notes
+======================================
+
+
+Enhancements
+------------
+
+- ``cuda-bindings`` docstrings now include the full documentation from the upstream CTK, not just the first paragraph.
+
+Prerelease feature
+------------------
+
+A new version of the ``nvrtc`` API is available as ``cuda.bindings._v2.nvrtc``.  The
+primary improvements are: (1) raising exceptions rather than returning error
+codes, (2) uses PEP8-compliant naming, and (3) more performance.  This API is
+still experimental and subject to change.
+
+Known issues
+------------
+
+* Updating from older versions (v12.6.2.post1 and below) via ``pip install -U cuda-python`` might not work. Please do a clean re-installation by uninstalling ``pip uninstall -y cuda-python`` followed by installing ``pip install cuda-python``.
+* ``nvml.system_get_process_name`` on WSL can return incorrect values.  To work around this, set the locale to "C" before calling ``nvml.device_get_compute_running_processes_v3`` (which sets the process names) and before calling ``nvml.system_get_process_name``. ``cuda_core`` does this automatically, but users of the raw NVML API will need to do this manually.