nvmath-python 0.5.0__cp313-cp313-win_amd64.whl

nvmath/bindings/cublasLt.pxd

@@ -0,0 +1,99 @@
+# Copyright (c) 2024-2025, NVIDIA CORPORATION & AFFILIATES. ALL RIGHTS RESERVED.
+#
+# SPDX-License-Identifier: Apache-2.0
+#
+# This code was automatically generated across versions from 11.0.3 to 12.8.0. Do not modify it directly.
+
+from libc.stdint cimport intptr_t
+
+from .cycublasLt cimport *
+
+
+###############################################################################
+# Types
+###############################################################################
+
+ctypedef cublasLtHandle_t Handle
+ctypedef cublasLtMatrixLayout_t MatrixLayout
+ctypedef cublasLtMatmulDesc_t MatmulDesc
+ctypedef cublasLtMatrixTransformDesc_t MatrixTransformDesc
+ctypedef cublasLtMatmulPreference_t MatmulPreference
+ctypedef cublasLtLoggerCallback_t LoggerCallback
+
+ctypedef cudaStream_t Stream
+ctypedef cudaDataType DataType
+ctypedef libraryPropertyType_t LibraryPropertyType
+
+
+###############################################################################
+# Enum
+###############################################################################
+
+ctypedef cublasLtMatmulTile_t _MatmulTile
+ctypedef cublasLtMatmulStages_t _MatmulStages
+ctypedef cublasLtPointerMode_t _PointerMode
+ctypedef cublasLtPointerModeMask_t _PointerModeMask
+ctypedef cublasLtOrder_t _Order
+ctypedef cublasLtMatrixLayoutAttribute_t _MatrixLayoutAttribute
+ctypedef cublasLtMatmulDescAttributes_t _MatmulDescAttribute
+ctypedef cublasLtMatrixTransformDescAttributes_t _MatrixTransformDescAttribute
+ctypedef cublasLtReductionScheme_t _ReductionScheme
+ctypedef cublasLtEpilogue_t _Epilogue
+ctypedef cublasLtMatmulSearch_t _MatmulSearch
+ctypedef cublasLtMatmulPreferenceAttributes_t _MatmulPreferenceAttribute
+ctypedef cublasLtMatmulAlgoCapAttributes_t _MatmulAlgoCapAttribute
+ctypedef cublasLtMatmulAlgoConfigAttributes_t _MatmulAlgoConfigAttribute
+ctypedef cublasLtClusterShape_t _ClusterShape
+ctypedef cublasLtMatmulInnerShape_t _MatmulInnerShape
+ctypedef cublasLtMatmulMatrixScale_t _MatmulMatrixScale
+
+
+###############################################################################
+# Functions
+###############################################################################
+
+cpdef intptr_t create() except? 0
+cpdef destroy(intptr_t light_handle)
+cpdef size_t get_version() except? 0
+cpdef size_t get_cudart_version() except? 0
+cpdef int get_property(int type) except? -1
+cpdef matmul(intptr_t light_handle, intptr_t compute_desc, intptr_t alpha, intptr_t a, intptr_t adesc, intptr_t b, intptr_t bdesc, intptr_t beta, intptr_t c, intptr_t cdesc, intptr_t d, intptr_t ddesc, intptr_t algo, intptr_t workspace, size_t workspace_size_in_bytes, intptr_t stream)
+cpdef matrix_transform(intptr_t light_handle, intptr_t transform_desc, intptr_t alpha, intptr_t a, intptr_t adesc, intptr_t beta, intptr_t b, intptr_t bdesc, intptr_t c, intptr_t cdesc, intptr_t stream)
+cpdef intptr_t matrix_layout_create(int type, uint64_t rows, uint64_t cols, int64_t ld) except? 0
+cpdef matrix_layout_destroy(intptr_t mat_layout)
+cpdef get_matrix_layout_attribute_dtype(int attr)
+cpdef matrix_layout_set_attribute(intptr_t mat_layout, int attr, intptr_t buf, size_t size_in_bytes)
+cpdef matrix_layout_get_attribute(intptr_t mat_layout, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef intptr_t matmul_desc_create(int compute_type, int scale_type) except? 0
+cpdef matmul_desc_destroy(intptr_t matmul_desc)
+cpdef get_matmul_desc_attribute_dtype(int attr)
+cpdef matmul_desc_set_attribute(intptr_t matmul_desc, int attr, intptr_t buf, size_t size_in_bytes)
+cpdef matmul_desc_get_attribute(intptr_t matmul_desc, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef intptr_t matrix_transform_desc_create(int scale_type) except? 0
+cpdef matrix_transform_desc_destroy(intptr_t transform_desc)
+cpdef get_matrix_transform_desc_attribute_dtype(int attr)
+cpdef matrix_transform_desc_set_attribute(intptr_t transform_desc, int attr, intptr_t buf, size_t size_in_bytes)
+cpdef matrix_transform_desc_get_attribute(intptr_t transform_desc, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef intptr_t matmul_preference_create() except? 0
+cpdef matmul_preference_destroy(intptr_t pref)
+cpdef get_matmul_preference_attribute_dtype(int attr)
+cpdef matmul_preference_set_attribute(intptr_t pref, int attr, intptr_t buf, size_t size_in_bytes)
+cpdef matmul_preference_get_attribute(intptr_t pref, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef matmul_algo_get_heuristic(intptr_t light_handle, intptr_t operation_desc, intptr_t adesc, intptr_t bdesc, intptr_t cdesc, intptr_t ddesc, intptr_t preference, int requested_algo_count, intptr_t heuristic_results_array, intptr_t return_algo_count)
+cpdef matmul_algo_init(intptr_t light_handle, int compute_type, int scale_type, int atype, int btype, int ctype, int dtype, int algo_id, intptr_t algo)
+cpdef matmul_algo_check(intptr_t light_handle, intptr_t operation_desc, intptr_t adesc, intptr_t bdesc, intptr_t cdesc, intptr_t ddesc, intptr_t algo, intptr_t result)
+cpdef get_matmul_algo_cap_attribute_dtype(int attr)
+cpdef matmul_algo_cap_get_attribute(intptr_t algo, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef get_matmul_algo_config_attribute_dtype(int attr)
+cpdef matmul_algo_config_set_attribute(intptr_t algo, int attr, intptr_t buf, size_t size_in_bytes)
+cpdef matmul_algo_config_get_attribute(intptr_t algo, int attr, intptr_t buf, size_t size_in_bytes, intptr_t size_written)
+cpdef logger_open_file(log_file)
+cpdef logger_set_level(int level)
+cpdef logger_set_mask(int mask)
+cpdef logger_force_disable()
+cpdef str get_status_name(int status)
+cpdef str get_status_string(int status)
+cpdef size_t heuristics_cache_get_capacity() except? 0
+cpdef heuristics_cache_set_capacity(size_t capacity)
+cpdef disable_cpu_instructions_set_mask(unsigned mask)
+cpdef tuple matmul_algo_get_ids(intptr_t light_handle, cublasComputeType_t compute_type, size_t scale_type, size_t atype, size_t btype, size_t ctype, size_t dtype, int requested_algo_count)

nvmath/bindings/cudss.pxd

@@ -0,0 +1,83 @@
+# Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. ALL RIGHTS RESERVED.
+#
+# SPDX-License-Identifier: Apache-2.0
+#
+# This code was automatically generated with version 0.5.0. Do not modify it directly.
+
+from libc.stdint cimport intptr_t
+
+from .cycudss cimport *
+
+
+###############################################################################
+# Types
+###############################################################################
+
+ctypedef cudssHandle_t Handle
+ctypedef cudssMatrix_t Matrix
+ctypedef cudssData_t Data
+ctypedef cudssConfig_t Config
+ctypedef cudssDistributedInterface_t DistributedInterface
+ctypedef cudssThreadingInterface_t ThreadingInterface
+ctypedef cudssDeviceMemHandler_t DeviceMemHandler
+
+ctypedef cudaStream_t Stream
+ctypedef cudaDataType DataType
+ctypedef libraryPropertyType_t LibraryPropertyType
+
+
+###############################################################################
+# Enum
+###############################################################################
+
+ctypedef cudssOpType_t _OpType
+ctypedef cudssConfigParam_t _ConfigParam
+ctypedef cudssDataParam_t _DataParam
+ctypedef cudssPhase_t _Phase
+ctypedef cudssStatus_t _Status
+ctypedef cudssMatrixType_t _MatrixType
+ctypedef cudssMatrixViewType_t _MatrixViewType
+ctypedef cudssIndexBase_t _IndexBase
+ctypedef cudssLayout_t _Layout
+ctypedef cudssAlgType_t _AlgType
+ctypedef cudssPivotType_t _PivotType
+ctypedef cudssMatrixFormat_t _MatrixFormat
+
+
+###############################################################################
+# Functions
+###############################################################################
+
+cpdef get_config_param_dtype(int attr)
+cpdef config_set(intptr_t config, int param, intptr_t value, size_t size_in_bytes)
+cpdef config_get(intptr_t config, int param, intptr_t value, size_t size_in_bytes, intptr_t size_written)
+cpdef get_data_param_dtype(int attr)
+cpdef data_set(intptr_t handle, intptr_t data, int param, intptr_t value, size_t size_in_bytes)
+cpdef data_get(intptr_t handle, intptr_t data, int param, intptr_t value, size_t size_in_bytes, intptr_t size_written)
+cpdef execute(intptr_t handle, int phase, intptr_t solver_config, intptr_t solver_data, intptr_t input_matrix, intptr_t solution, intptr_t rhs)
+cpdef set_stream(intptr_t handle, intptr_t stream)
+cpdef set_comm_layer(intptr_t handle, intptr_t comm_lib_file_name)
+cpdef set_threading_layer(intptr_t handle, thr_lib_file_name)
+cpdef intptr_t config_create() except? 0
+cpdef config_destroy(intptr_t solver_config)
+cpdef intptr_t data_create(intptr_t handle) except? 0
+cpdef data_destroy(intptr_t handle, intptr_t solver_data)
+cpdef intptr_t create() except? 0
+cpdef destroy(intptr_t handle)
+cpdef int get_property(int property_type) except? -1
+cpdef intptr_t matrix_create_dn(int64_t nrows, int64_t ncols, int64_t ld, intptr_t values, int value_type, int layout) except? 0
+cpdef intptr_t matrix_create_csr(int64_t nrows, int64_t ncols, int64_t nnz, intptr_t row_start, intptr_t row_end, intptr_t col_indices, intptr_t values, int index_type, int value_type, int mtype, int mview, int index_base) except? 0
+cpdef intptr_t matrix_create_batch_dn(int64_t batch_count, intptr_t nrows, intptr_t ncols, intptr_t ld, intptr_t values, int index_type, int value_type, int layout) except? 0
+cpdef intptr_t matrix_create_batch_csr(int64_t batch_count, intptr_t nrows, intptr_t ncols, intptr_t nnz, intptr_t row_start, intptr_t row_end, intptr_t col_indices, intptr_t values, int index_type, int value_type, int mtype, int mview, int index_base) except? 0
+cpdef matrix_destroy(intptr_t matrix)
+cpdef tuple matrix_get_dn(intptr_t matrix)
+cpdef tuple matrix_get_csr(intptr_t matrix)
+cpdef matrix_set_values(intptr_t matrix, intptr_t values)
+cpdef matrix_set_csr_pointers(intptr_t matrix, intptr_t row_offsets, intptr_t row_end, intptr_t col_indices, intptr_t values)
+cpdef matrix_get_batch_dn(intptr_t matrix, intptr_t batch_count, intptr_t nrows, intptr_t ncols, intptr_t ld, intptr_t values, intptr_t index_type, intptr_t value_type, intptr_t layout)
+cpdef matrix_get_batch_csr(intptr_t matrix, intptr_t batch_count, intptr_t nrows, intptr_t ncols, intptr_t nnz, intptr_t row_start, intptr_t row_end, intptr_t col_indices, intptr_t values, intptr_t index_type, intptr_t value_type, intptr_t mtype, intptr_t mview, intptr_t index_base)
+cpdef matrix_set_batch_values(intptr_t matrix, intptr_t values)
+cpdef matrix_set_batch_csr_pointers(intptr_t matrix, intptr_t row_offsets, intptr_t row_end, intptr_t col_indices, intptr_t values)
+cpdef int matrix_get_format(intptr_t matrix) except? -1
+cpdef get_device_mem_handler(intptr_t handle, intptr_t handler)
+cpdef set_device_mem_handler(intptr_t handle, intptr_t handler)

nvmath/bindings/cudss.pyi

@@ -0,0 +1,249 @@
+# Copyright (c) 2024-2025, NVIDIA CORPORATION & AFFILIATES. ALL RIGHTS RESERVED.
+#
+# SPDX-License-Identifier: Apache-2.0
+
+import _cython_3_0_12
+import enum
+from typing import Any, Callable, ClassVar
+
+__pyx_capi__: dict
+__test__: dict
+check_status: _cython_3_0_12.cython_function_or_method
+config_create: _cython_3_0_12.cython_function_or_method
+config_destroy: _cython_3_0_12.cython_function_or_method
+config_get: _cython_3_0_12.cython_function_or_method
+config_set: _cython_3_0_12.cython_function_or_method
+create: _cython_3_0_12.cython_function_or_method
+data_create: _cython_3_0_12.cython_function_or_method
+data_destroy: _cython_3_0_12.cython_function_or_method
+data_get: _cython_3_0_12.cython_function_or_method
+data_set: _cython_3_0_12.cython_function_or_method
+destroy: _cython_3_0_12.cython_function_or_method
+execute: _cython_3_0_12.cython_function_or_method
+get_config_param_dtype: _cython_3_0_12.cython_function_or_method
+get_data_param_dtype: _cython_3_0_12.cython_function_or_method
+get_device_mem_handler: _cython_3_0_12.cython_function_or_method
+get_property: _cython_3_0_12.cython_function_or_method
+matrix_create_batch_csr: _cython_3_0_12.cython_function_or_method
+matrix_create_batch_dn: _cython_3_0_12.cython_function_or_method
+matrix_create_csr: _cython_3_0_12.cython_function_or_method
+matrix_create_dn: _cython_3_0_12.cython_function_or_method
+matrix_destroy: _cython_3_0_12.cython_function_or_method
+matrix_get_batch_csr: _cython_3_0_12.cython_function_or_method
+matrix_get_batch_dn: _cython_3_0_12.cython_function_or_method
+matrix_get_csr: _cython_3_0_12.cython_function_or_method
+matrix_get_dn: _cython_3_0_12.cython_function_or_method
+matrix_get_format: _cython_3_0_12.cython_function_or_method
+matrix_set_batch_csr_pointers: _cython_3_0_12.cython_function_or_method
+matrix_set_batch_values: _cython_3_0_12.cython_function_or_method
+matrix_set_csr_pointers: _cython_3_0_12.cython_function_or_method
+matrix_set_values: _cython_3_0_12.cython_function_or_method
+set_comm_layer: _cython_3_0_12.cython_function_or_method
+set_device_mem_handler: _cython_3_0_12.cython_function_or_method
+set_stream: _cython_3_0_12.cython_function_or_method
+set_threading_layer: _cython_3_0_12.cython_function_or_method
+
+class AlgType(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    ALG_1: ClassVar[AlgType] = ...
+    ALG_2: ClassVar[AlgType] = ...
+    ALG_3: ClassVar[AlgType] = ...
+    ALG_DEFAULT: ClassVar[AlgType] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class ConfigParam(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    FACTORIZATION_ALG: ClassVar[ConfigParam] = ...
+    HOST_NTHREADS: ClassVar[ConfigParam] = ...
+    HYBRID_DEVICE_MEMORY_LIMIT: ClassVar[ConfigParam] = ...
+    HYBRID_EXECUTE_MODE: ClassVar[ConfigParam] = ...
+    HYBRID_MODE: ClassVar[ConfigParam] = ...
+    IR_N_STEPS: ClassVar[ConfigParam] = ...
+    IR_TOL: ClassVar[ConfigParam] = ...
+    MATCHING_TYPE: ClassVar[ConfigParam] = ...
+    MAX_LU_NNZ: ClassVar[ConfigParam] = ...
+    PIVOT_EPSILON: ClassVar[ConfigParam] = ...
+    PIVOT_EPSILON_ALG: ClassVar[ConfigParam] = ...
+    PIVOT_THRESHOLD: ClassVar[ConfigParam] = ...
+    PIVOT_TYPE: ClassVar[ConfigParam] = ...
+    REORDERING_ALG: ClassVar[ConfigParam] = ...
+    SOLVE_ALG: ClassVar[ConfigParam] = ...
+    SOLVE_MODE: ClassVar[ConfigParam] = ...
+    USE_CUDA_REGISTER_MEMORY: ClassVar[ConfigParam] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class DataParam(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    COMM: ClassVar[DataParam] = ...
+    DIAG: ClassVar[DataParam] = ...
+    HYBRID_DEVICE_MEMORY_MIN: ClassVar[DataParam] = ...
+    INERTIA: ClassVar[DataParam] = ...
+    INFO: ClassVar[DataParam] = ...
+    LU_NNZ: ClassVar[DataParam] = ...
+    MEMORY_ESTIMATES: ClassVar[DataParam] = ...
+    NPIVOTS: ClassVar[DataParam] = ...
+    PERM_COL: ClassVar[DataParam] = ...
+    PERM_REORDER_COL: ClassVar[DataParam] = ...
+    PERM_REORDER_ROW: ClassVar[DataParam] = ...
+    PERM_ROW: ClassVar[DataParam] = ...
+    USER_PERM: ClassVar[DataParam] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class IndexBase(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    ONE: ClassVar[IndexBase] = ...
+    ZERO: ClassVar[IndexBase] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class Layout(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    COL_MAJOR: ClassVar[Layout] = ...
+    ROW_MAJOR: ClassVar[Layout] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class MatrixFormat(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    BATCH: ClassVar[MatrixFormat] = ...
+    CSR: ClassVar[MatrixFormat] = ...
+    DENSE: ClassVar[MatrixFormat] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class MatrixType(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    GENERAL: ClassVar[MatrixType] = ...
+    HERMITIAN: ClassVar[MatrixType] = ...
+    HPD: ClassVar[MatrixType] = ...
+    SPD: ClassVar[MatrixType] = ...
+    SYMMETRIC: ClassVar[MatrixType] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class MatrixViewType(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    FULL: ClassVar[MatrixViewType] = ...
+    LOWER: ClassVar[MatrixViewType] = ...
+    UPPER: ClassVar[MatrixViewType] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class OpType(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    MAX: ClassVar[OpType] = ...
+    MIN: ClassVar[OpType] = ...
+    SUM: ClassVar[OpType] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class Phase(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    ANALYSIS: ClassVar[Phase] = ...
+    FACTORIZATION: ClassVar[Phase] = ...
+    REFACTORIZATION: ClassVar[Phase] = ...
+    SOLVE: ClassVar[Phase] = ...
+    SOLVE_BWD: ClassVar[Phase] = ...
+    SOLVE_DIAG: ClassVar[Phase] = ...
+    SOLVE_FWD: ClassVar[Phase] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class PivotType(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    PIVOT_COL: ClassVar[PivotType] = ...
+    PIVOT_NONE: ClassVar[PivotType] = ...
+    PIVOT_ROW: ClassVar[PivotType] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class Status(enum.IntEnum):
+    __new__: ClassVar[Callable] = ...
+    ALLOC_FAILED: ClassVar[Status] = ...
+    EXECUTION_FAILED: ClassVar[Status] = ...
+    INTERNAL_ERROR: ClassVar[Status] = ...
+    INVALID_VALUE: ClassVar[Status] = ...
+    NOT_INITIALIZED: ClassVar[Status] = ...
+    NOT_SUPPORTED: ClassVar[Status] = ...
+    SUCCESS: ClassVar[Status] = ...
+    _generate_next_value_: ClassVar[Callable] = ...
+    _member_map_: ClassVar[dict] = ...
+    _member_names_: ClassVar[list] = ...
+    _member_type_: ClassVar[type[int]] = ...
+    _unhashable_values_: ClassVar[list] = ...
+    _use_args_: ClassVar[bool] = ...
+    _value2member_map_: ClassVar[dict] = ...
+    def __format__(self, *args, **kwargs) -> str: ...
+
+class cuDSSError(Exception):
+    def __init__(self, status) -> Any: ...
+    def __reduce__(self) -> Any: ...

nvmath/bindings/cufft.pxd

@@ -0,0 +1,114 @@
+# Copyright (c) 2024-2025, NVIDIA CORPORATION & AFFILIATES. ALL RIGHTS RESERVED.
+#
+# SPDX-License-Identifier: Apache-2.0
+#
+# This code was automatically generated across versions from 11.0.3 to 12.8.0. Do not modify it directly.
+
+from libc.stdint cimport intptr_t
+
+from .cycufft cimport *
+
+
+###############################################################################
+# Types
+###############################################################################
+
+ctypedef cufftXt1dFactors Xt1dFactors
+ctypedef cufftCallbackLoadC CallbackLoadC
+ctypedef cufftCallbackLoadZ CallbackLoadZ
+ctypedef cufftCallbackLoadR CallbackLoadR
+ctypedef cufftCallbackLoadD CallbackLoadD
+ctypedef cufftJITCallbackLoadC JITCallbackLoadC
+ctypedef cufftJITCallbackLoadZ JITCallbackLoadZ
+ctypedef cufftJITCallbackLoadR JITCallbackLoadR
+ctypedef cufftJITCallbackLoadD JITCallbackLoadD
+ctypedef cufftCallbackStoreR CallbackStoreR
+ctypedef cufftJITCallbackStoreR JITCallbackStoreR
+ctypedef cufftCallbackStoreD CallbackStoreD
+ctypedef cufftJITCallbackStoreD JITCallbackStoreD
+ctypedef cufftCallbackStoreC CallbackStoreC
+ctypedef cufftJITCallbackStoreC JITCallbackStoreC
+ctypedef cufftCallbackStoreZ CallbackStoreZ
+ctypedef cufftJITCallbackStoreZ JITCallbackStoreZ
+
+ctypedef cudaStream_t Stream
+ctypedef cudaDataType DataType
+ctypedef libraryPropertyType_t LibraryPropertyType
+
+
+###############################################################################
+# Enum
+###############################################################################
+
+ctypedef libFormat_t _LibFormat
+ctypedef cufftResult _Result
+ctypedef cufftType _Type
+ctypedef cufftCompatibility _Compatibility
+ctypedef cufftXtSubFormat _XtSubFormat
+ctypedef cufftXtCopyType _XtCopyType
+ctypedef cufftXtQueryType _XtQueryType
+ctypedef cufftXtWorkAreaPolicy _XtWorkAreaPolicy
+ctypedef cufftXtCallbackType _XtCallbackType
+ctypedef cufftProperty _Property
+
+
+###############################################################################
+# Functions
+###############################################################################
+
+cpdef plan1d(intptr_t plan, int nx, int type, int batch)
+cpdef plan2d(intptr_t plan, int nx, int ny, int type)
+cpdef plan3d(intptr_t plan, int nx, int ny, int nz, int type)
+cpdef plan_many(intptr_t plan, int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch)
+cpdef size_t make_plan1d(int plan, int nx, int type, int batch) except? -1
+cpdef size_t make_plan2d(int plan, int nx, int ny, int type) except? -1
+cpdef size_t make_plan3d(int plan, int nx, int ny, int nz, int type) except? -1
+cpdef size_t make_plan_many(int plan, int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef size_t make_plan_many64(int plan, int rank, n, inembed, long long int istride, long long int idist, onembed, long long int ostride, long long int odist, int type, long long int batch) except? -1
+cpdef size_t get_size_many64(int plan, int rank, n, inembed, long long int istride, long long int idist, onembed, long long int ostride, long long int odist, int type, long long int batch) except? -1
+cpdef size_t estimate1d(int nx, int type, int batch) except? -1
+cpdef size_t estimate2d(int nx, int ny, int type) except? -1
+cpdef size_t estimate3d(int nx, int ny, int nz, int type) except? -1
+cpdef size_t estimate_many(int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef int create() except? -1
+cpdef size_t get_size1d(int handle, int nx, int type, int batch) except? -1
+cpdef size_t get_size2d(int handle, int nx, int ny, int type) except? -1
+cpdef size_t get_size3d(int handle, int nx, int ny, int nz, int type) except? -1
+cpdef size_t get_size_many(int handle, int rank, intptr_t n, intptr_t inembed, int istride, int idist, intptr_t onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef size_t get_size(int handle) except? -1
+cpdef set_work_area(int plan, intptr_t work_area)
+cpdef set_auto_allocation(int plan, int auto_allocate)
+cpdef exec_c2c(int plan, intptr_t idata, intptr_t odata, int direction)
+cpdef exec_r2c(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_c2r(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_z2z(int plan, intptr_t idata, intptr_t odata, int direction)
+cpdef exec_d2z(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_z2d(int plan, intptr_t idata, intptr_t odata)
+cpdef set_stream(int plan, intptr_t stream)
+cpdef destroy(int plan)
+cpdef int get_version() except? -1
+cpdef int get_property(int type) except? -1
+cpdef xt_set_gpus(int handle, int n_gpus, which_gpus)
+cpdef intptr_t xt_malloc(int plan, int format) except? -1
+cpdef xt_memcpy(int plan, intptr_t dst_pointer, intptr_t src_pointer, int type)
+cpdef xt_free(intptr_t descriptor)
+cpdef xt_set_work_area(int plan, intptr_t work_area)
+cpdef xt_exec_descriptor_c2c(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor_r2c(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_c2r(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_z2z(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor_d2z(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_z2d(int plan, intptr_t input, intptr_t output)
+cpdef xt_query_plan(int plan, intptr_t query_struct, int query_type)
+cpdef xt_clear_callback(int plan, int cb_type)
+cpdef xt_set_callback_shared_size(int plan, int cb_type, size_t shared_size)
+cpdef size_t xt_make_plan_many(int plan, int rank, n, inembed, long long int istride, long long int idist, int inputtype, onembed, long long int ostride, long long int odist, int outputtype, long long int batch, int executiontype) except? 0
+cpdef size_t xt_get_size_many(int plan, int rank, n, inembed, long long int istride, long long int idist, int inputtype, onembed, long long int ostride, long long int odist, int outputtype, long long int batch, int executiontype) except? 0
+cpdef xt_exec(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_set_work_area_policy(int plan, int policy, intptr_t work_size)
+cpdef xt_set_jit_callback(int plan, lto_callback_fatbin, size_t lto_callback_fatbin_size, int type, caller_info)
+cpdef xt_set_subformat_default(int plan, int subformat_forward, int subformat_inverse)
+cpdef set_plan_property_int64(int plan, int property, long long int input_value_int)
+cpdef long long int get_plan_property_int64(int plan, int property) except? -1
+cpdef reset_plan_property(int plan, int property)

nvmath/bindings/cufftMp.pxd

@@ -0,0 +1,124 @@
+# Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. ALL RIGHTS RESERVED.
+#
+# SPDX-License-Identifier: Apache-2.0
+#
+# This code was automatically generated across versions from 11.2.6 to 11.4.0. Do not modify it directly.
+
+from libc.stdint cimport intptr_t
+
+from .cycufftMp cimport *
+
+
+###############################################################################
+# Types
+###############################################################################
+
+ctypedef cufftReshapeHandle ReshapeHandle
+ctypedef cufftXt1dFactors Xt1dFactors
+ctypedef cufftCallbackLoadC CallbackLoadC
+ctypedef cufftCallbackLoadZ CallbackLoadZ
+ctypedef cufftCallbackLoadR CallbackLoadR
+ctypedef cufftCallbackLoadD CallbackLoadD
+ctypedef cufftJITCallbackLoadC JITCallbackLoadC
+ctypedef cufftJITCallbackLoadZ JITCallbackLoadZ
+ctypedef cufftJITCallbackLoadR JITCallbackLoadR
+ctypedef cufftJITCallbackLoadD JITCallbackLoadD
+ctypedef cufftCallbackStoreR CallbackStoreR
+ctypedef cufftJITCallbackStoreR JITCallbackStoreR
+ctypedef cufftCallbackStoreD CallbackStoreD
+ctypedef cufftJITCallbackStoreD JITCallbackStoreD
+ctypedef cufftCallbackStoreC CallbackStoreC
+ctypedef cufftJITCallbackStoreC JITCallbackStoreC
+ctypedef cufftCallbackStoreZ CallbackStoreZ
+ctypedef cufftJITCallbackStoreZ JITCallbackStoreZ
+
+ctypedef cudaStream_t Stream
+ctypedef cudaDataType DataType
+ctypedef libraryPropertyType_t LibraryPropertyType
+
+
+###############################################################################
+# Enum
+###############################################################################
+
+ctypedef libFormat_t _LibFormat
+ctypedef cufftResult _Result
+ctypedef cufftType _Type
+ctypedef cufftCompatibility _Compatibility
+ctypedef cufftProperty _Property
+ctypedef cufftXtSubFormat _XtSubFormat
+ctypedef cufftXtCopyType _XtCopyType
+ctypedef cufftXtQueryType _XtQueryType
+ctypedef cufftXtWorkAreaPolicy _XtWorkAreaPolicy
+ctypedef cufftXtCallbackType _XtCallbackType
+ctypedef cufftMpCommType _MpCommType
+
+
+###############################################################################
+# Functions
+###############################################################################
+
+cpdef plan1d(intptr_t plan, int nx, int type, int batch)
+cpdef plan2d(intptr_t plan, int nx, int ny, int type)
+cpdef plan3d(intptr_t plan, int nx, int ny, int nz, int type)
+cpdef plan_many(intptr_t plan, int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch)
+cpdef size_t make_plan1d(int plan, int nx, int type, int batch) except? -1
+cpdef size_t make_plan2d(int plan, int nx, int ny, int type) except? -1
+cpdef size_t make_plan3d(int plan, int nx, int ny, int nz, int type) except? -1
+cpdef size_t make_plan_many(int plan, int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef size_t make_plan_many64(int plan, int rank, n, inembed, long long int istride, long long int idist, onembed, long long int ostride, long long int odist, int type, long long int batch) except? -1
+cpdef size_t get_size_many64(int plan, int rank, n, inembed, long long int istride, long long int idist, onembed, long long int ostride, long long int odist, int type, long long int batch) except? -1
+cpdef size_t estimate1d(int nx, int type, int batch) except? -1
+cpdef size_t estimate2d(int nx, int ny, int type) except? -1
+cpdef size_t estimate3d(int nx, int ny, int nz, int type) except? -1
+cpdef size_t estimate_many(int rank, n, inembed, int istride, int idist, onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef int create() except? -1
+cpdef size_t get_size1d(int handle, int nx, int type, int batch) except? -1
+cpdef size_t get_size2d(int handle, int nx, int ny, int type) except? -1
+cpdef size_t get_size3d(int handle, int nx, int ny, int nz, int type) except? -1
+cpdef size_t get_size_many(int handle, int rank, intptr_t n, intptr_t inembed, int istride, int idist, intptr_t onembed, int ostride, int odist, int type, int batch) except? -1
+cpdef size_t get_size(int handle) except? -1
+cpdef set_work_area(int plan, intptr_t work_area)
+cpdef set_auto_allocation(int plan, int auto_allocate)
+cpdef exec_c2c(int plan, intptr_t idata, intptr_t odata, int direction)
+cpdef exec_r2c(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_c2r(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_z2z(int plan, intptr_t idata, intptr_t odata, int direction)
+cpdef exec_d2z(int plan, intptr_t idata, intptr_t odata)
+cpdef exec_z2d(int plan, intptr_t idata, intptr_t odata)
+cpdef set_stream(int plan, intptr_t stream)
+cpdef destroy(int plan)
+cpdef int get_version() except? -1
+cpdef int get_property(int type) except? -1
+cpdef set_plan_property_int64(int plan, int property, long long int input_value_int)
+cpdef long long int get_plan_property_int64(int plan, int property) except? -1
+cpdef reset_plan_property(int plan, int property)
+cpdef xt_set_gpus(int handle, int n_gpus, which_gpus)
+cpdef intptr_t xt_malloc(int plan, int format) except? -1
+cpdef xt_memcpy(int plan, intptr_t dst_pointer, intptr_t src_pointer, int type)
+cpdef xt_free(intptr_t descriptor)
+cpdef xt_set_work_area(int plan, intptr_t work_area)
+cpdef xt_exec_descriptor_c2c(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor_r2c(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_c2r(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_z2z(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor_d2z(int plan, intptr_t input, intptr_t output)
+cpdef xt_exec_descriptor_z2d(int plan, intptr_t input, intptr_t output)
+cpdef xt_query_plan(int plan, intptr_t query_struct, int query_type)
+cpdef xt_clear_callback(int plan, int cb_type)
+cpdef xt_set_callback_shared_size(int plan, int cb_type, size_t shared_size)
+cpdef size_t xt_make_plan_many(int plan, int rank, n, inembed, long long int istride, long long int idist, int inputtype, onembed, long long int ostride, long long int odist, int outputtype, long long int batch, int executiontype) except? 0
+cpdef size_t xt_get_size_many(int plan, int rank, n, inembed, long long int istride, long long int idist, int inputtype, onembed, long long int ostride, long long int odist, int outputtype, long long int batch, int executiontype) except? 0
+cpdef xt_exec(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_exec_descriptor(int plan, intptr_t input, intptr_t output, int direction)
+cpdef xt_set_work_area_policy(int plan, int policy, intptr_t work_size)
+cpdef attach_comm(int plan, int comm_type, intptr_t comm_handle)
+cpdef xt_set_distribution(int plan, int rank, lower_input, upper_input, lower_output, upper_output, strides_input, strides_output)
+cpdef xt_set_subformat_default(int plan, int subformat_forward, int subformat_inverse)
+cpdef intptr_t create_reshape() except? -1
+cpdef attach_reshape_comm(intptr_t handle, int comm_type, intptr_t comm_handle)
+cpdef size_t get_reshape_size(intptr_t handle) except? -1
+cpdef make_reshape_11_2(intptr_t handle, size_t element_size, int rank, lower_input, upper_input, lower_output, upper_output, strides_input, strides_output)
+cpdef exec_reshape_async(intptr_t handle, intptr_t data_out, intptr_t data_in, intptr_t workspace, intptr_t stream)
+cpdef destroy_reshape(intptr_t handle)
+cpdef make_reshape_11_4(intptr_t handle, size_t element_size, int rank, lower_input, upper_input, strides_input, lower_output, upper_output, strides_output, intptr_t comm_handle, int comm_type)